Compare commits

..

140 Commits

Author SHA1 Message Date
Marcel
4bcf568ed4 Merge branch 'main' of ssh://git.raddatz.cloud:222/marcel/familienarchiv
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m22s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m41s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m7s
nightly / deploy-staging (push) Successful in 2m35s
2026-06-08 16:27:41 +02:00
Marcel
ddb1ec4df8 docs(timeline): add Zeitstrahl visual specs (global Concept A, event editor)
Visual design specs for Milestone #14:
- zeitstrahl-global-concepts.html — A/B/C exploration of the global timeline
- zeitstrahl-final-spec.html — canonical Concept A (global + per-person Lebensweg)
- zeitstrahl-event-editor-spec.html — curator event editor + document quick-action

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-08 16:27:15 +02:00
d650b6c066 refactor(search): remove NLP/smart-search feature entirely (#772)
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m23s
CI / OCR Service Tests (push) Successful in 24s
CI / Backend Unit Tests (push) Successful in 3m46s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 25s
CI / Compose Bucket Idempotency (push) Successful in 1m8s
## Summary

- Removes the NLP/smart-search feature completely — the feature was too unreliable and slow; users get better results with the regular search filters
- Deletes the entire backend `search/` package (NlSearchController, NlQueryParserService, NlpClient, NlSearchRateLimiter — 14 classes + 6 test classes)
- Deletes the `nlp-service/` Python microservice (FastAPI, rapidfuzz, DB-backed person matching)
- Removes all frontend NL search components: SmartModeToggle, SmartSearchStatus, InterpretationChipRow, DisambiguationPicker, chip-types, theme-chip-removal
- Strips smart-mode logic from SearchFilterBar and documents/+page.svelte
- Removes `SMART_SEARCH_UNAVAILABLE` / `SMART_SEARCH_RATE_LIMITED` error codes from backend, frontend types, and all three i18n files (de/en/es)
- Removes `nlp-service` container and `APP_NLP_BASE_URL` from both docker-compose files
- Removes Ollama/NLP Prometheus scrape job and Grafana dashboard
- Deletes ADRs 028 (×2), 034, 035

## Test plan

- [ ] Backend compiles: `cd backend && ./mvnw compile -q` → BUILD SUCCESS
- [ ] Frontend server tests pass: `cd frontend && npm run test -- --project=server`
- [ ] No NLP/smart-search references remain in source: `grep -r "SmartSearch\|NlSearch\|nlp-service\|SMART_SEARCH" backend/src frontend/src`
- [ ] `docker compose config` validates both compose files
- [ ] Search page loads, filter bar works, no smart-mode toggle visible

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Marcel <marcel@familienarchiv>
Reviewed-on: #772
2026-06-08 10:57:00 +02:00
Marcel
e63eaadc33 docs(timeline): add Person date+precision migration as foundational issue
Replace Person birthYear/deathYear integers with birthDate/deathDate +
DatePrecision so known exact birthdays render precisely. Migration,
re-import preservation rule, and bounded blast radius captured; becomes
issue 1 the timeline's derived events depend on.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 19:26:13 +02:00
Marcel
d4a25e34d8 docs(timeline): add family timeline (Zeitstrahl) design spec
Hand-curated, year-banded vertical timeline weaving derived person
life-events, curated personal/historical events, and date-placed
letters. Includes proposed sub-issue breakdown for a milestone.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 19:18:55 +02:00
Marcel
8e63867ad8 docs(specs): UI specs for Lesereisen reader and Journey editor
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m23s
CI / OCR Service Tests (push) Successful in 24s
CI / Backend Unit Tests (push) Successful in 4m2s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 21s
CI / Compose Bucket Idempotency (push) Successful in 1m8s
nightly / deploy-staging (push) Successful in 2m44s
lesereisen-reader-spec.html — Issue #752
  LR-0 type selector on /geschichten/new
  LR-1 REISE badge on the list
  LR-2 Journey reader (ordered cards, interlude asides, no position numbers)

lesereisen-editor-spec.html — Issue #753
  LE-1 empty JourneyEditor layout
  LE-2 editor with mixed items (documents + interludes, drag handles)
  LE-3 inline note-editing state
  LE-4 mobile layout

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 19:07:34 +02:00
Marcel
6b0a06e8b1 feat(nlp-service): scaffold — models, requirements, CLAUDE.md
Task 1: Create standalone FastAPI service scaffold with models, test framework,
and documentation. Includes ParseRequest, ParseResponse Pydantic models matching
OllamaExtraction contract, plus three passing tests validating model validation.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 10:11:34 +02:00
Marcel
7c1eef710c docs(nlp): add spaCy NLP service implementation plan
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 09:52:07 +02:00
Marcel
03e22a2f26 docs(nlp): add spaCy NLP service prototype design spec
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 09:40:00 +02:00
Marcel
6878419156 merge: resolve conflicts with origin/main (#763 person name-match integration)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m31s
CI / OCR Service Tests (pull_request) Successful in 25s
CI / Backend Unit Tests (pull_request) Successful in 3m48s
CI / fail2ban Regex (pull_request) Successful in 45s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
CI / Unit & Component Tests (push) Successful in 3m20s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m48s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 23s
CI / Compose Bucket Idempotency (push) Successful in 1m8s
- Drop unused MAX_CANDIDATES constant (not referenced in service)
- Keep detached-entity safety comment in resolveTags()
- Add 3 new partial-name match tests (23a/b/c) from #763
- Use resolveByName() API in test 28 (replaces findByDisplayNameContaining)
- Add NameMatches glossary entry from #763

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:50:48 +02:00
Marcel
09b77e9b36 test(person): pin fetchPool dedup when one person matches two tokens (#763 review)
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m20s
CI / OCR Service Tests (push) Successful in 24s
CI / Backend Unit Tests (push) Successful in 3m53s
CI / fail2ban Regex (push) Successful in 44s
CI / Semgrep Security Scan (push) Successful in 21s
CI / Compose Bucket Idempotency (push) Successful in 1m5s
Assert that when the same person id is returned by two different token
fetches, the person appears exactly once in the result -- pinning
fetchPool's putIfAbsent dedup so a future refactor can't silently
double-classify a candidate.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
9d202b042b test(person): close fetch-to-classify seam for alias matches on real Postgres (#763 review)
AC#4 (maiden alias -> direct) and AC#5 (alias first name -> fetchable +
classifiable) were each split across PersonRepositoryTest (the fetch) and
PersonServiceTest (the classifier with stubs) -- nothing walked
searchByName -> resolveByName end-to-end on real Postgres. Add two tests
in the existing @DataJpaTest slice that build a real PersonService over
the autowired repositories, persist a person with a MAIDEN_NAME alias and
one with an alias firstName, and assert both classify as direct.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
8429b1e9f8 fix(search): derive disambiguation trigger aria-label from match count (#763 review)
The trigger hardcoded the multiple-people label for every count, so a
single did-you-mean picker announced "Mehrere Personen gefunden" to
screen readers while sighted users saw one name and a "Meintest du …?"
heading. Derive the trigger's accessible name from persons.length: a
single suggestion reuses the heading prop, two or more keep the
multiple-people label. Visible truncated name span unchanged.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
6959651b36 docs(search): document NameMatches and resolveByName (#763)
GLOSSARY entry for NameMatches (direct vs partial name-match strength and how
the search layer maps it); person/README adds resolveByName to the public
surface. No ADR — the matching rule is localized and justified inline.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
0ef4f4f07c feat(search): case-appropriate disambiguation picker copy (#763)
A 1-item picker now reads "Meintest du …?" (a single direct match auto-selects
and never reaches the picker), while ≥2 keeps the "Person auswählen" framing.
The prompt lives in a visible, non-truncated panel heading (the trigger span
clips at 320px), and the "(auswählen…)" cue is dropped for the 1-item case.
DisambiguationPicker takes heading + showCue props; the page derives both from
ambiguousPersons.length. New search_disambiguation_did_you_mean key in de/en/es.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
f1bb9d3a69 feat(search): map direct/partial NameMatches into resolve buckets (#763)
resolveNames now delegates to PersonService.resolveByName and maps by match
strength: 1 direct → resolved (auto-select), ≥2 direct → ambiguous, 0 direct
with partials → ambiguous suggestions, 0 candidates → folded into full-text.
A single direct match no longer forces the picker when looser substring hits
coexist. The MAX_CANDIDATES cap moved into PersonService (after classification);
the MAX_NAME_LENGTH guard, resolved-cap overflow, and sender/receiver mapping
are preserved.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
ca52145556 feat(person): add resolveByName for direct/partial name matching (#763)
Token-set containment over all of a person's name components (firstName,
lastName, alias, each PersonNameAlias first+last, title) decides direct vs
partial. Orchestrates tokenize → cap(8) → fetch pool → classify → cap(10)
after classification, with an empty-token guard and a PII-free debug log of
the outcome bucket. MAX_TOKENS is a DoS control; the after-classify cap keeps a
direct match that sorts past position 10 among partials. Read-only transaction
keeps lazy nameAliases reachable during classification (ADR-022).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
9a26bf75b0 feat(person): match alias first names in searchByName (#763)
The direct-match classifier accepts alias firstName tokens, so the fetch must
surface candidates matchable only via an alias first name. Add a.firstName to
the searchByName LIKE clause (reuses the bound :query — injection-proof). The
person_name_aliases.first_name column already exists; no migration.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
9c616f9fb8 feat(person): add name-match tokenizer for direct matching (#763)
Lowercase, split on whitespace/hyphen/apostrophe, drop empties. Applied
symmetrically to query and candidate name components so "Anna-Maria" and
"Anna Maria" tokenize alike. Foundation for resolveByName direct matching.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
0fe0ae5235 docs(search): ADR-028 fix + glossary + C4 diagram for tag resolution (#743)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
2c909f49a8 feat(search): wire theme chip removal to URL navigation in +page.svelte
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
87fd0f39bb feat(search): render removable theme chips in InterpretationChipRow
When tagsApplied is true, each resolvedTag renders as a 'Thema: Name'
chip with optional inline color style from the tag's resolved color.
Clicking × calls onRemoveChip('theme', tag.name).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
7f3ad8ce89 feat(api): add TagHint schema and extend NlQueryInterpretation with resolvedTags/tagsApplied
Manual update since Docker compose backend runs old build; regenerate with
npm run generate:api once new backend is deployed.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
aa1f6436cc feat(i18n): add search_chip_theme_prefix to de/en/es message bundles
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
b825076733 test(search): DataJpaTest for descendant-expansion via TagRepository
Verifies the recursive CTE in findDescendantIdsByName expands a parent tag
to include all child IDs, and that findByNameContainingIgnoreCase matches
both parent and child names when the fragment appears in both.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
01df815bad test(search): add 11 tag-resolution test cases to NlQueryParserServiceTest
Covers multi-tag match, no-match FTS fallback, mixed resolution, personRole
bypass, cap at 10, short-keyword skip, dedup, rawQuery suppression when all
keywords resolve, flag independence, colour propagation via resolveEffectiveColors,
and colour=null when depth constraint prevents resolution.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
dcd0e725a7 feat(search): implement keyword→tag resolution in NlQueryParserService
Keywords that substring-match the tag taxonomy become OR-union tag filters;
non-matching keywords stay as FTS text. Resolved tags surface in the
NlQueryInterpretation as TagHint objects with effective colours. The
rawQuery fallback is now guarded by hadStructuredMatch to prevent
double-apply when all keywords resolve.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
39ff63921d refactor(search): extract ChipType to chip-types.ts; audit NL fixtures
Pre-implementation step for #743: ChipType union extracted from
InterpretationChipRow and +page.svelte into shared chip-types.ts;
resolvedTags/tagsApplied neutral defaults added to test fixtures.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
5a09cd4cb4 feat(search): extend NlQueryInterpretation with resolvedTags + tagsApplied
Positional record fields added; all 3 construction sites updated with neutral
defaults; NlQueryParserService wired for TagService (4th constructor arg);
NlQueryParserServiceTest and NlSearchControllerTest synced.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
4e0ebc72c8 feat(search): add TagHint record for NL tag resolution API surface
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
0f0d89702d feat(search): add TagService.findByNameContaining for NL tag resolution
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 08:47:47 +02:00
Marcel
fb41affd4c docs(search): note vitest-browser workaround for + in path
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m22s
CI / OCR Service Tests (pull_request) Successful in 24s
CI / Backend Unit Tests (pull_request) Successful in 3m47s
CI / fail2ban Regex (pull_request) Successful in 46s
CI / Semgrep Security Scan (pull_request) Successful in 24s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
Addresses @Sara review: browser tests in this spec fail silently when
the project path contains '+' (common in git worktrees). The comment
tells developers to copy the frontend directory to a clean path.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 00:58:36 +02:00
Marcel
dc366ed403 docs(search): add detached-entity safety comment in resolveTags
Addresses @Markus review: tags fetched by findByNameContaining live outside
any transaction; Hibernate's dirty-check never fires on them. The comment
removes the ambiguity for cold readers.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-07 00:58:03 +02:00
Marcel
64b7b2315d docs(search): ADR-028 fix + glossary + C4 diagram for tag resolution (#743)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m25s
CI / OCR Service Tests (pull_request) Successful in 23s
CI / Backend Unit Tests (pull_request) Successful in 4m1s
CI / fail2ban Regex (pull_request) Successful in 45s
CI / Semgrep Security Scan (pull_request) Successful in 23s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m7s
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 23:42:23 +02:00
Marcel
2a7e133717 feat(search): wire theme chip removal to URL navigation in +page.svelte
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 23:40:33 +02:00
Marcel
5387bc9247 feat(search): render removable theme chips in InterpretationChipRow
When tagsApplied is true, each resolvedTag renders as a 'Thema: Name'
chip with optional inline color style from the tag's resolved color.
Clicking × calls onRemoveChip('theme', tag.name).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 23:33:53 +02:00
Marcel
847874abb3 feat(api): add TagHint schema and extend NlQueryInterpretation with resolvedTags/tagsApplied
Manual update since Docker compose backend runs old build; regenerate with
npm run generate:api once new backend is deployed.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 23:01:11 +02:00
Marcel
573bca4986 feat(i18n): add search_chip_theme_prefix to de/en/es message bundles
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:59:58 +02:00
Marcel
86690fdbb6 test(search): DataJpaTest for descendant-expansion via TagRepository
Verifies the recursive CTE in findDescendantIdsByName expands a parent tag
to include all child IDs, and that findByNameContainingIgnoreCase matches
both parent and child names when the fragment appears in both.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:59:07 +02:00
Marcel
6cb1025881 test(search): add 11 tag-resolution test cases to NlQueryParserServiceTest
Covers multi-tag match, no-match FTS fallback, mixed resolution, personRole
bypass, cap at 10, short-keyword skip, dedup, rawQuery suppression when all
keywords resolve, flag independence, colour propagation via resolveEffectiveColors,
and colour=null when depth constraint prevents resolution.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:57:17 +02:00
Marcel
fc557bd9ae feat(search): implement keyword→tag resolution in NlQueryParserService
Keywords that substring-match the tag taxonomy become OR-union tag filters;
non-matching keywords stay as FTS text. Resolved tags surface in the
NlQueryInterpretation as TagHint objects with effective colours. The
rawQuery fallback is now guarded by hadStructuredMatch to prevent
double-apply when all keywords resolve.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:54:33 +02:00
Marcel
e94414b81a refactor(search): extract ChipType to chip-types.ts; audit NL fixtures
Pre-implementation step for #743: ChipType union extracted from
InterpretationChipRow and +page.svelte into shared chip-types.ts;
resolvedTags/tagsApplied neutral defaults added to test fixtures.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:49:54 +02:00
Marcel
7eee688ce9 feat(search): extend NlQueryInterpretation with resolvedTags + tagsApplied
Positional record fields added; all 3 construction sites updated with neutral
defaults; NlQueryParserService wired for TagService (4th constructor arg);
NlQueryParserServiceTest and NlSearchControllerTest synced.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:37:45 +02:00
Marcel
8905135006 feat(search): add TagHint record for NL tag resolution API surface
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:35:24 +02:00
Marcel
8bd8390891 feat(search): add TagService.findByNameContaining for NL tag resolution
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 22:34:34 +02:00
Marcel
ed98729f75 docs(adr): record prod Ollama deployment + keep-alive decision (ADR-034)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m23s
CI / OCR Service Tests (pull_request) Successful in 24s
CI / Backend Unit Tests (pull_request) Successful in 3m52s
CI / fail2ban Regex (pull_request) Successful in 45s
CI / Semgrep Security Scan (pull_request) Successful in 25s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m7s
CI / Unit & Component Tests (push) Successful in 3m23s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m52s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 23s
CI / Compose Bucket Idempotency (push) Successful in 1m4s
nightly / deploy-staging (push) Successful in 2m44s
Capture the why behind deploying Ollama to prod/staging compose: the
corrected init recipe (supersedes ADR-028 §10's never-functional curl
loop), the OLLAMA_KEEP_ALIVE=-1 pin (so a future maintainer doesn't
optimize it away and reintroduce the post-idle cold-load 503), the
30->60s timeout NFR, and the memswap==mem hard-OOM trade-off.

Addresses #759 review (Markus #3, Nora #2).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:16:03 +02:00
Marcel
db87a64cc0 docs(c4): de-duplicate Ollama container in l2-containers diagram
The diagram declared Container(ollama, ...) twice — an alias collision that
renders a duplicate box. It also declared the backend->ollama relationship
twice. Keep the richer 'Ollama LLM Service' declaration and the more
specific 'NL query parsing (POST /api/generate)' relationship; drop the
duplicates.

Addresses #759 review (Markus #2).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:14:26 +02:00
Marcel
d7d6d0638c fix(infra): make dev Ollama model-init offline-safe
Mirror the prod hardening in the dev stack: guard the model pull with
`ollama list | grep -q <model>` so an already-cached model exits clean
without a registry round-trip. Keeps dev and prod on one recipe.

Addresses #759 review (Tobias #1).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:13:19 +02:00
Marcel
a2f37f85a6 fix(infra): make prod Ollama model-init offline-safe
The init command unconditionally ran `ollama pull`, which contacts the
registry to verify the manifest digest even when the model is already on
the volume. A host reboot during a registry/upstream-network blip would
then fail init non-zero, the `service_completed_successfully` gate would
never be met, and the ollama service (hence NL search) would stay down
until the registry was reachable again.

Guard the pull with `ollama list | grep -q <model>` so a cached model
exits clean without any registry round-trip.

Addresses #759 review (Tobias #1).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:12:21 +02:00
Marcel
f22a1a1cfa docs(deploy): fix prod Ollama volume name to match hyphenated compose volume
docker-compose.prod.yml declares the volume as `ollama-models` (hyphen),
so the compose-project-prefixed name is `archiv-production_ollama-models`,
not the underscored `archiv-production_ollama_models` the model-upgrade
guide documented. The documented `docker volume rm` would not have matched
the real volume.

Addresses #759 review (Tobias #2).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:09:48 +02:00
Marcel
2a0863cf3e docs(deploy): correct Ollama read timeout default to 60s
application.yaml sets app.ollama.timeout-seconds: 60 (raised from 30 to
absorb the cold model load on the first query after an Ollama restart),
but DEPLOYMENT.md still documented 30. A doc that contradicts the shipped
value is a traceability defect.

Addresses #759 review (Markus, Felix, Elicit).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 20:08:55 +02:00
Marcel
9e97687d0f fix(search): pin Ollama model in memory + raise read timeout
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m18s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m55s
CI / fail2ban Regex (pull_request) Successful in 51s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m8s
NL search recovered after deploy but went 503 again after a few minutes:
Ollama unloads the model after its default ~5 min keep-alive, so the next
query cold-loads the 4.7 GB model and exceeds the backend's 30s read
timeout (ResourceAccessException -> SMART_SEARCH_UNAVAILABLE). Warm
inference is ~18s; the cold load after idle is what timed out.

- docker-compose.{prod,yml}: set OLLAMA_KEEP_ALIVE=-1 on the ollama
  service so the model stays resident and never pays a cold-load penalty
  during normal operation (verified on staging: `ollama ps` -> UNTIL
  "Forever"; host has 47 GB free).
- application.yaml: raise app.ollama.timeout-seconds 30 -> 60 so the one
  unavoidable cold load (first query after an Ollama restart, before the
  model is pinned) completes instead of timing out.

Refs #758

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 19:27:02 +02:00
Marcel
b665e1132d fix(infra): deploy Ollama to prod/staging compose + fix broken model-init recipe
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 4m0s
CI / OCR Service Tests (pull_request) Successful in 25s
CI / Backend Unit Tests (pull_request) Successful in 3m56s
CI / fail2ban Regex (pull_request) Successful in 45s
CI / Semgrep Security Scan (pull_request) Successful in 23s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m5s
NL search returned 503 (SMART_SEARCH_UNAVAILABLE / "Intelligente Suche
nicht verfügbar") on staging because Ollama was never reachable.

Two defects, both downstream of #737:

1. Ollama was added only to the dev docker-compose.yml. Staging/prod
   deploy from the self-contained docker-compose.prod.yml, which had no
   ollama service — so the backend (defaulting to http://ollama:11434)
   hit a non-existent host (ResourceAccessException -> 503).

2. The merged model-init recipe never worked: the ollama/ollama image
   ENTRYPOINT is `ollama` (so `command: sh -c ...` ran as `ollama sh ...`
   -> "unknown command sh"), and the image ships no curl (so both the
   readiness loop and the healthcheck could never pass).

- docker-compose.prod.yml: add ollama-model-init + ollama services and
  the ollama-models volume, with the corrected recipe (entrypoint
  override to /bin/sh -c, `ollama list` for readiness and healthcheck).
- docker-compose.yml: fix the same broken entrypoint/command and the
  curl healthcheck so the dev stack actually starts Ollama.

Verified on staging end-to-end: model-init exits 0, ollama healthy,
backend reaches /api/tags, inference succeeds within the 8g limit.

Refs #758

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 19:20:22 +02:00
Marcel
87af9ab446 docs(c4): add smart-search components to l3-frontend diagram (#739 review)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m22s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m45s
CI / fail2ban Regex (pull_request) Successful in 42s
CI / Semgrep Security Scan (pull_request) Successful in 23s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
CI / Unit & Component Tests (push) Successful in 3m19s
CI / OCR Service Tests (push) Successful in 25s
CI / Backend Unit Tests (push) Successful in 3m51s
CI / fail2ban Regex (push) Successful in 48s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m6s
Markus (architect): document SearchFilterBar + the search/ components
(SmartModeToggle, InterpretationChipRow, SmartSearchStatus,
DisambiguationPicker) and the POST /api/search/nl relation.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 18:27:00 +02:00
Marcel
0058b297d8 fix(search): enlarge sub-12px text for senior legibility (#739 review)
Leonie (UX): the toggle pill (text-[7.5px]) and loading subtitle
(text-[9px]) were below the 12px floor for the 60+ audience. Bump both
to text-xs and the toggle icon to h-3.5/w-3.5. Overrides the visual
spec's tokens, which conflicted with the issue's own legibility mandate.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 18:26:24 +02:00
Marcel
230f23e37c test(search): add NL search happy-path Playwright E2E (#739)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m17s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m47s
CI / fail2ban Regex (pull_request) Successful in 46s
CI / Semgrep Security Scan (pull_request) Successful in 24s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
Mock POST /api/search/nl (delayed fixture: 2-name directional + applied
keyword), assert loading announcement → chips render → axe-clean in light
and dark → removing the keyword chip re-runs a keyword GET with the
remaining sender+receiver params. Adds a data-testid wrapper on the NL
results region for axe scoping.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:58:15 +02:00
Marcel
e604967a3f docs(search): document src/routes/search/ component directory (#739)
Add the smart-search sub-component directory to the frontend Project
Structure tree (merge blocker per #739).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:57:59 +02:00
Marcel
169e1ad9de test(search): cover smart-mode chip lifecycle hooks (#739)
SearchFilterBar drives chip-clearing via onModeToggle (mode switch) and
onSmartSearch (new query); pin that callback contract.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:54:25 +02:00
Marcel
f2f42ed415 feat(search): orchestrate NL search on the documents page (#739)
Lift smartMode to documents/+page.svelte and drive the full smart-search
lifecycle: POST /api/search/nl via csrfFetch, loading/error panels, chip
row, single-select disambiguation, and a transparent empty state. Chip
removal and disambiguation selection map the interpretation to keyword
params and re-run via GET (Option A in-page fallback). Mode toggle and
new queries reset prior interpretation.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:54:07 +02:00
Marcel
5945824b54 feat(search): wire SmartModeToggle into SearchFilterBar (#739)
Add smartMode $bindable plus onSmartSearch/onModeToggle callbacks. The
toggle pill sits in the input's right slot (decorative icon moved to the
left); smart mode disables the live oninput keyword search, adds
maxlength=500, and submits the NL query on Enter. 4 integration specs.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:47:05 +02:00
Marcel
fa41394e66 feat(search): add DisambiguationPicker single-select disclosure (#739)
Accessible disclosure: aria-expanded/aria-controls trigger, focus moves
into the option list on open, Escape and click-outside close and return
focus to the trigger, selecting a candidate emits onSelect. Single-select
(GET re-run) per the resolved #738 open decision — backend has no
multi-sender OR param. 5 vitest-browser-svelte specs.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:43:27 +02:00
Marcel
fb00c7818e feat(search): add SmartSearchStatus full-area panels (#739)
Loading panel (role=status, motion-safe spinner + pulsing subtitle) and
combined error panels: 503 (red icon + switch-to-keyword button) and
429 (amber clock icon, no action button). 5 vitest-browser-svelte specs.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:40:28 +02:00
Marcel
8ed65f8602 feat(search): add InterpretationChipRow component (#739)
Renders type-prefixed chips (Absender/Zeitraum/Stichwort), a single
directional chip for 2-name queries, gates keyword chips on
keywordsApplied, and emits onRemoveChip(type, value?). Truncating name
spans keep the 44px × button visible; chip wrappers show a focus ring.
9 vitest-browser-svelte specs (red/green).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:38:51 +02:00
Marcel
9e425c98a1 feat(search): add SmartModeToggle pill component (#739)
Toggle pill with aria-pressed, active/resting styles matching the
AND/OR operator button pattern, and mobile-expanded KI/Text labels.
4 vitest-browser-svelte specs (red/green).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:35:05 +02:00
Marcel
ddce268113 feat(search): add NL search frontend i18n keys (de/en/es)
Toggle labels, loading panel, error panels (503/429), empty-state
retry, chip type-prefixes + remove label, and disambiguation strings
for the smart search UI (#739). Formal Sie form per project standard.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 17:32:50 +02:00
4a43962c98 Merge pull request 'feat(search): NL search backend — POST /api/search/nl with Ollama integration (#738)' (#756) from worktree-feat+issue-738-nl-search-backend into main
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m17s
CI / OCR Service Tests (push) Successful in 22s
CI / Backend Unit Tests (push) Successful in 3m43s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m4s
Reviewed-on: #756
2026-06-06 16:52:43 +02:00
Marcel
9a9e1c4c40 merge(search): resolve DEPLOYMENT.md conflict — keep setup + upgrade sections
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m17s
CI / OCR Service Tests (pull_request) Successful in 23s
CI / Backend Unit Tests (pull_request) Successful in 3m45s
CI / fail2ban Regex (pull_request) Successful in 48s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m4s
Both the first-time model pull runbook (from this branch) and the model
upgrade procedure (from main) belong in DEPLOYMENT.md.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:47:49 +02:00
Marcel
62c8ce4cb2 docs(search): add NL search visual spec — toggle pill, chips, full-area states (#739)
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / OCR Service Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / fail2ban Regex (push) Has been cancelled
CI / Semgrep Security Scan (push) Has been cancelled
CI / Compose Bucket Idempotency (push) Has been cancelled
Covers the SmartModeToggle pill (inside the search input, Google AI Mode
style), InterpretationChipRow anatomy, DisambiguationPicker, and all
status/error/empty states as full-result-area panels.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:47:09 +02:00
Marcel
4c620619d4 fix(search): formal Sie form in German error strings; clean up DocumentService imports
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m19s
CI / OCR Service Tests (pull_request) Successful in 23s
CI / Backend Unit Tests (pull_request) Successful in 3m57s
CI / fail2ban Regex (pull_request) Successful in 45s
CI / Semgrep Security Scan (pull_request) Successful in 21s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m5s
- error_smart_search_unavailable/rate_limited now use "Sie" (formal) to
  match the tone of all existing German error messages
- Replace inline FQNs in DocumentService.buildPersonSpec with proper
  JoinType + Predicate imports

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:46:40 +02:00
Marcel
44baff9c9c docs(search): update CLAUDE.md, GLOSSARY, DEPLOYMENT, and C4 diagrams
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m21s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m52s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 21s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m3s
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:16:04 +02:00
Marcel
4634da9865 feat(search): add @Schema annotations and regenerate TypeScript API types
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:11:01 +02:00
Marcel
79e4a3f9db feat(search): add searchDocumentsByPersonId with Specification-based sender/receiver query
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 16:04:54 +02:00
Marcel
70e8a6e6ad feat(search): implement NlSearchController with @WebMvcTest tests (7 cases)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:58:35 +02:00
Marcel
3af1095d13 feat(search): implement NlQueryParserService with Mockito tests (23 cases)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:54:45 +02:00
Marcel
8c835e957a feat(search): implement RestClientOllamaClient with WireMock tests
Switch to wiremock-jetty12 artifact and force ee10 Jetty deps to 12.1.8
to resolve compatibility with Spring Boot 4's Jetty 12.1.8 core.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:43:49 +02:00
Marcel
fe8fcba7a7 feat(search): add NlSearchRateLimiter with Bucket4j/Caffeine
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:39:06 +02:00
Marcel
e0c80ac193 feat(search): add Ollama and rate-limit config properties
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:37:24 +02:00
Marcel
005265b5a8 feat(search): add NL search error codes and i18n strings
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:36:13 +02:00
Marcel
684c6e63de feat(search): add NL search domain records and OllamaClient interfaces
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:33:56 +02:00
Marcel
e27d52b9ee docs(c4): add L3 backend search component diagram
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:32:40 +02:00
Marcel
6f5497c7bf docs(adr): ADR-028 — NL search via Ollama
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:31:53 +02:00
Marcel
e0fac783e8 feat(person): add findByDisplayNameContaining service method
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:30:30 +02:00
Marcel
202ea85a58 build(deps): add org.wiremock:wiremock 3.9.2 as test dependency
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 15:28:55 +02:00
Marcel
7679596c70 docs(ollama): add model upgrade runbook + post-deploy smoke test to DEPLOYMENT.md
Some checks failed
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / OCR Service Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / fail2ban Regex (pull_request) Has been cancelled
CI / Semgrep Security Scan (pull_request) Has been cancelled
CI / Compose Bucket Idempotency (pull_request) Has been cancelled
CI / Unit & Component Tests (push) Successful in 3m16s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m37s
CI / fail2ban Regex (push) Successful in 47s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m4s
Addresses Elicit's and Sara's review concerns on PR #749:
- Expand §6 ollama_models section into a full model upgrade runbook (step-by-step
  docker volume rm + recreate, including production volume name prefix)
- Add re-deploy idempotency note to §3.4 (init container exits quickly when model
  already present on the volume)
- Add NL search smoke test to §3.4 (curl command distinguishing 200 from 503
  NL_SEARCH_UNAVAILABLE)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
3d5dcd1f18 docs(deployment): fix OLLAMA_API_KEY version ref and add --wait warning
Updated OLLAMA_API_KEY env vars table from 0.6.5 to 0.6.5 or 0.30.6 to
match both tested versions. Added an explicit warning in §3.4 that
docker compose up -d --wait blocks for 60–90 min on first deploy when the
model pull has not yet completed.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
52fca38f0f docs(env): correct OLLAMA_API_KEY comment — tested on 0.6.5 and 0.30.6
Both versions were tested and neither enforces the key. Comment updated to
say "0.6.5 or 0.30.6" and surface archiv-net as the sole effective control.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
662a8f3e80 fix(infra): interpolate APP_OLLAMA_BASE_URL so .env empty-value disables Ollama
Hardcoded literal overrides any .env setting — setting APP_OLLAMA_BASE_URL=
in .env had no effect on the backend container. Now uses the same pattern
as APP_OCR_TRAINING_TOKEN with a safe default.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
cbba95c3f8 docs(c4): fix Ollama container version 0.6.5 → 0.30.6 in l2-containers.puml
Diagram must match the pinned image version in docker-compose.yml.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
3536ed884c docs(adr): fix ADR-028 §12 false API-key claim, stale TBD, and §7 title
§12 stated OLLAMA_API_KEY guards against lateral movement — contradicts
§7's empirical finding that it is not enforced. Replaced with an accurate
note referencing §7. Stale pre-merge placeholder in Consequences ("Three
TBD items must be resolved") removed; all three are resolved. §7 section
title updated from "0.6.5" to "0.6.5 and 0.30.6" to match the body text.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
5a939d9222 fix(infra): escape \$\$SERVE_PID in compose command to prevent interpolation (#737)
Docker Compose interpolates $VAR in command strings — use $$ to pass a
literal $ to the shell so SERVE_PID=$! and kill $SERVE_PID work correctly.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
93e90424ab docs(adr): update ADR-028 with 0.30.6 verified findings for API key + read_only (#737)
- OLLAMA_API_KEY: non-enforcement confirmed on both 0.6.5 and 0.30.6
- read_only: true: confirmed working on both 0.6.5 and 0.30.6
- Peak RSS during pull: ~108 MiB (well under 2g limit)
- All TBD placeholders resolved

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
e8f3004c4f feat(infra): add Ollama env vars to .env.example (#737)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
9637ebbca2 feat(infra): add Ollama Docker Compose services for NL search (#737)
- ollama-model-init: one-shot init container that pulls qwen2.5:7b-instruct-q4_K_M
  into the ollama_models volume on first start
- ollama: main inference service on archiv-net (expose: only, no public port)
- ollama_models named volume for persistent model storage
- APP_OLLAMA_BASE_URL + APP_OLLAMA_API_KEY added to backend env
- Both services: cap_drop ALL, no-new-privileges, read_only+tmpfs (ADR-019 + ADR-028)
- start_period: 60s — model pre-pulled by init container

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
df10a42069 docs(deploy): document Ollama hardware requirements, env vars, and ops notes (#737)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:59:35 +02:00
Marcel
64120a30b5 docs(arch): add Ollama container to C4 level-2 container diagram (#737)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:58:49 +02:00
Marcel
25252fc709 feat(observability): add Grafana Ollama inference latency dashboard (#737)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:58:49 +02:00
Marcel
1f379a161d fix(observability): fix OCR target name + add Ollama scrape job (#737)
- prometheus.yml: ocr:8000 → ocr-service:8000 (Docker service name is
  ocr-service, not ocr — current scrape target has never resolved)
- Add Ollama scrape job on ollama:11434 /metrics

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:58:49 +02:00
Marcel
c0d034c85d docs(adr): add ADR-028 — Ollama Docker Compose service for NL search (#737)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:58:49 +02:00
Marcel
ca93cde06e docs(infra): correct server specs — Hetzner Serverbörse i7-6700 64 GB, not CX32
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m18s
CI / OCR Service Tests (push) Successful in 21s
CI / Backend Unit Tests (push) Successful in 3m46s
CI / fail2ban Regex (push) Successful in 48s
CI / Semgrep Security Scan (push) Successful in 23s
CI / Compose Bucket Idempotency (push) Successful in 1m6s
Replace all references to the CX32 VPS (8 GB RAM, Hetzner Cloud) with the
actual production server: a Hetzner Serverbörse dedicated server with an
Intel Core i7-6700 (4C/8T, 3.4 GHz) and 64 GB RAM.

Affected files:
- .claude/personas/devops.md — monthly cost line + upgrade example
- docs/infrastructure/production-compose.md — sizing section + cost table
- docs/DEPLOYMENT.md — OCR memory table + OCR_MEM_LIMIT env var description
- docs/adr/004-pdfbox-thumbnails.md — thumbnailExecutor memory ceiling note
- docs/adr/021-tmpdir-persistent-volume-staging.md — OOMKill rationale in alternatives

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06 14:51:07 +02:00
Marcel
7629e35897 docs(adr): renumber tag case-collision ADR 032 → 033 to resolve number clash (#731)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m15s
CI / OCR Service Tests (pull_request) Successful in 23s
CI / Backend Unit Tests (pull_request) Successful in 3m40s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 21s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m5s
CI / Unit & Component Tests (push) Successful in 3m13s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m40s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 21s
CI / Compose Bucket Idempotency (push) Successful in 1m7s
Both #730 (tag case-collision) and #684 (person-delete DB integrity) landed
an ADR-032 on main. Renumber the tag/case-collision one to 033 — it is
referenced only from this PR's person-domain comments and its own file, so the
move is self-contained and touches no Flyway migration. The person-delete
ADR-032 and the V71 migration comment that cites it are deliberately left
untouched (editing an applied migration would drift its Flyway checksum).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 13:52:25 +02:00
Marcel
cd741b9f57 docs(person): clarify case-collision scope at the exact-case lookups (#731)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m15s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m42s
CI / fail2ban Regex (pull_request) Successful in 46s
CI / Semgrep Security Scan (pull_request) Successful in 21s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m5s
Review noted the "never throws" claim was overstated: the exact-case Optional
lookups still surface a NonUniqueResultException on two byte-identical
same-case rows. That is a true data anomaly out of #731's scope (ambiguous =
case-insensitive) and resolves to the opaque INTERNAL_ERROR, never a wrong
row. Record that boundary at both resolution points and in ADR-032 so the gap
is not silently assumed covered.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 13:36:22 +02:00
Marcel
ddf378aaac fix(person): resolve ambiguous sender names to null on upload (#731)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m18s
CI / OCR Service Tests (pull_request) Successful in 25s
CI / Backend Unit Tests (pull_request) Successful in 3m38s
CI / fail2ban Regex (pull_request) Successful in 43s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
findByName resolved via Optional<Person>
findByFirstNameIgnoreCaseAndLastNameIgnoreCase, which threw
NonUniqueResultException once two people shared a first+last name case-
insensitively (hans müller / Hans Müller) — a 500 on the routine upload path
(DocumentService.storeDocument sender resolution).

findByName now resolves exact-case → single case-insensitive match → else
empty. The sender path deliberately diverges from the alias path: an
ambiguous name leaves the sender UNSET rather than guessing the lowest id,
because correct provenance beats a confidently-wrong pre-fill a reviewer
won't re-check. The two new name queries use explicit HQL equality so a null
first name binds as `= NULL` (no match) instead of the derived-query fold to
`first_name IS NULL`, which would widen a last-name-only row in as a sender.

Pins the opaque error path (IncorrectResultSizeDataAccessException stays
INTERNAL_ERROR with no Hibernate/SQL/row-count leak) and extends ADR-032 with
the Person section.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 13:03:04 +02:00
Marcel
20cfe41f21 fix(person): resolve case-colliding aliases without throwing (#731)
findOrCreateByAlias resolved via Optional<Person> findByAliasIgnoreCase,
which throws NonUniqueResultException once two aliases collide only by case
(müller / Müller) — a generic 500 on the importer path. Mirror the #730 tag
fix: resolve exact-case first, then the lowest-id case-insensitive sibling,
then create-when-absent (institution/group and maiden-name alias preserved).
The throwing Optional<…>IgnoreCase variant is deleted so it can't be reused.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:50:21 +02:00
Marcel
43601a3770 test(transcription): persist real persons for mention FK after V71 (#684)
All checks were successful
CI / Unit & Component Tests (push) Successful in 3m20s
CI / OCR Service Tests (push) Successful in 23s
CI / Backend Unit Tests (push) Successful in 3m39s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m6s
V71 gives transcription_block_mentioned_persons.person_id a real FK, so two
TranscriptionBlockMentionsRepositoryTest cases that inserted mention rows with
random (non-existent) person ids now violate fk_tbmp_person. Persist real
Person rows and use their ids. Caught by CI's full suite.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
6603bc5333 test(person): address PR #736 review nits
- AC-3 cascade test: assert an innocent bystander's mention row survives the
  delete, proving the cascade is scoped to the deleted person (Nora).
- Fix integration-test comment: receivers is @ManyToMany(LAZY), not an EAGER
  @ElementCollection (Sara).
- ADR-032: note the @ prefix is kept in the degraded path, stripped in live
  mentions (Leonie).
- Add trailing newline to PersonRepository.java (Felix).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
6753d115f9 fix(db): leave V56 untouched to avoid Flyway checksum drift (#684)
Editing an already-applied migration changes its Flyway checksum and would
fail validateOnMigrate against prod (where V56 is applied). Revert the V56
comment edit; V71 now records that it reverses V56's no-FK choice and points
to ADR-032 as the authoritative record, so the V56 -> V71 trail stays
discoverable without touching the applied migration. (DevOps review, PR #736.)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
73dd6c80fa docs(adr): record DB-level person-delete integrity decision (ADR-032) (#684)
Capture the reversal of V56's no-FK decision, the DB-layer-integrity
principle, and the cascade-boundary invariant (the cascade never reaches
documents rows). Numbered 032 — 028-031 are already taken on main; the
issue's '028 is next' was written before main moved.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
9ade36dd3b docs(db): annotate person-delete ON DELETE behaviour in DB diagrams (#684)
Annotate SET NULL on documents.sender_id and CASCADE on
document_receivers.person_id, and add the new
transcription_block_mentioned_persons -> persons person_id FK (CASCADE)
to both db-relationships.puml and db-orm.puml.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
378da60ae8 test(mention): lock deleted-person graceful-degradation contract (#684)
Strengthen one renderTranscriptionBody case into the AC-6 contract: a
@DisplayName with an empty mentionedPersons array (the deleted-person case
V71 produces) must render as plain readable text with no <a>, person-mention
class, data-person-id, or href. Guards against a future renderer refactor
silently reintroducing the dead-link-on-deleted-person degradation.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
6d267f2269 test(person): describe DB-cascade mechanism in delete service-path test (#684)
The deletePerson service-path guard (AC-4) is unchanged behaviourally, but its
comments described the removed reassignSenderToNull/deleteReceiverReferences
chain. Update them to the V71 ON DELETE cascade mechanism.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
ff76a3784f refactor(person): simplify mergePersons to lean on V71 cascade (#684)
Drop the explicit deleteReceiverReferences call from mergePersons — the
source's leftover receiver join rows now cascade-drop via V71's ON DELETE
CASCADE on deleteById. Remove the now-unused deleteReceiverReferences
repository method (and its repo test), and add clearAutomatically +
flushAutomatically to the remaining merge native queries so the L1 cache
cannot desync from the bulk updates. Rewrite the merge unit test with
verifyNoMoreInteractions and add an end-to-end merge regression test (AC-7).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
534665459f refactor(person): thin deletePerson to lean on V71 DB cascade (#684)
Drop the application-layer sender/receiver detach from deletePerson — the
V71 ON DELETE constraints now enforce it. Remove the now-unused
reassignSenderToNull repository method and rewrite the unit test to assert
only the existence check plus deleteById (verifyNoMoreInteractions).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
fd792f6d78 feat(person): enforce person-delete integrity at the DB layer (V71) (#684)
Add ON DELETE behaviour to the two V1 FKs into persons (documents.sender_id
-> SET NULL, document_receivers.person_id -> CASCADE) and a real FK with
ON DELETE CASCADE on the transcription_block_mentioned_persons soft reference,
cleaning up pre-existing orphan mention rows first. The cascade stays strictly
at the join/reference layer and never reaches documents rows.

Proven by new Postgres-backed PersonRepositoryTest cascade tests (AC-1/2/3/8
plus the cascade-boundary document-survival guard). Rewrites the now-stale
V56 'no FK' comment.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 12:34:46 +02:00
Marcel
bafbf609eb docs(adr): ADR-032 tag-name resolution tolerates case-collisions (#730)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m16s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m34s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
CI / Unit & Component Tests (push) Successful in 3m17s
CI / OCR Service Tests (push) Successful in 24s
CI / Backend Unit Tests (push) Successful in 3m36s
CI / fail2ban Regex (push) Successful in 46s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m6s
Records the lasting decision behind the #730 fix: exact-case-first
resolution, deterministic lowest-id case-insensitive fallback, and the
explicit refusal of a unique(lower(name)) constraint (collisions are
valid canonical nodes). Previously the rationale lived only in code
comments and the issue body. Raised as a blocker in the PR #733 review.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 11:09:10 +02:00
Marcel
2710f2e233 test(tag): close review-flagged gaps in case-collision coverage (#730)
Two adversarial gaps from PR #733 review:

- Unit: exact-case must win even when its id is NOT the lowest, proving
  exact-case short-circuits before the lowest-id tie-break (a naive
  "lowest id across all CI matches" would pick the wrong row).
- Integration: assert findAllByNameIgnoreCase folds the UPPERCASE
  "GLÜCKWÜNSCHE" — the exact string findOrCreate passes — so the umlaut
  proof matches the resolution path under test, not a lowercase probe.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 11:07:39 +02:00
Marcel
80f6468d52 refactor(tag): use orElseThrow over Optional.get in findOrCreate (#730)
The lowest-id tie-break stream is guarded non-empty, so .get() never
throws — but the project bans Optional.get(). Switch to .orElseThrow()
for the project idiom. No behaviour change.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 11:05:45 +02:00
Marcel
a58378e8f0 test(tag): pin case-colliding tag resolution on real Postgres (#730)
All checks were successful
CI / Unit & Component Tests (pull_request) Successful in 3m16s
CI / OCR Service Tests (pull_request) Successful in 21s
CI / Backend Unit Tests (pull_request) Successful in 3m35s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 21s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
Mocked TagServiceTest can't prove the two things that actually broke:
that findAllByNameIgnoreCase folds umlauts the way Postgres LOWER() does,
and that saving a document tagged with a case-colliding tag no longer
throws NonUniqueResultException. Testcontainers postgres:16-alpine:

- updateDocument on a doc tagged with the child "weihnachten" succeeds
  and keeps exactly the child tag (not the parent).
- findOrCreate("GLÜCKWÜNSCHE") resolves the Glückwünsche/glückwünsche
  umlaut pair deterministically (lowest id) without throwing — the
  regression catcher a plain-ASCII pair would miss.
- bulk-edit funnels through resolveTags → findOrCreate, guarding a
  future refactor that bypasses it.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 10:53:04 +02:00
Marcel
d000170f52 fix(tag): resolve case-colliding tag names without throwing (#730)
findOrCreate used tagRepository.findByNameIgnoreCase, which returns
Optional<Tag> and threw NonUniqueResultException whenever two tags
collided case-insensitively (a canonical parent and its same-named
lowercase child). Every document carrying such a tag became un-editable:
any save re-resolves the whole tag set by name and blew up with a 500.

Replace the throwing lookup with exact-case-first resolution: findByName
(exact) → findAllByNameIgnoreCase (lowest-id, deterministic, never
throws) → create. Delete findByNameIgnoreCase so the throwing call can't
be reintroduced. Case collisions are valid tree nodes — no migration, no
unique(lower(name)) constraint.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-06 10:49:02 +02:00
Marcel
d1ed9c022f test(stammbaum): fix #718 tab-order test for tidy-tree layout (#724)
Some checks failed
CI / Unit & Component Tests (pull_request) Successful in 3m17s
CI / OCR Service Tests (pull_request) Successful in 23s
CI / Backend Unit Tests (pull_request) Successful in 3m39s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 23s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m5s
CI / Unit & Component Tests (push) Successful in 3m19s
CI / OCR Service Tests (push) Successful in 23s
CI / fail2ban Regex (push) Has been cancelled
CI / Semgrep Security Scan (push) Has been cancelled
CI / Compose Bucket Idempotency (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
nightly / deploy-staging (push) Successful in 1m55s
The #718 keyboard-tab-order test hardcoded the visual order
['Eugenie','Walter','Clara','Hans'] on the assumption that buildLayout
sorts each generation alphabetically. #724 replaced that with the
tidy-tree layout, which orders a couple's run by structural ownership
(earliest birth year, then a deterministic id tie-break) — so Walter
(id …a1) now owns the run and Eugenie renders to his right.

Both PRs were green independently; the stale assertion only surfaced
once #718 and #724 landed together on main. Correct the expected reading
order to ['Walter','Eugenie','Clara','Hans'] and refresh the now-wrong
'alphabetical' comment. The companion self-validating test (DOM order ==
sorted by y,x) already guarded the real property, so only the hardcoded
assertion needed updating.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 18:00:59 +02:00
Marcel
1e5e8e43e8 refactor(transcribe): extract t-mark + draw-cue policy into tested helpers (#327)
Some checks failed
CI / Unit & Component Tests (push) Failing after 2m33s
CI / OCR Service Tests (push) Successful in 24s
CI / Backend Unit Tests (push) Successful in 3m42s
CI / fail2ban Regex (push) Successful in 43s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m7s
Review follow-up (Sara, fast-follow): the t no-active-region guard and the
draw-cue arm/disarm rule lived inline in the page with no direct coverage.
Extracted to pure resolveTrainingMark() (no-op when no region; recognition
enrolled flip) and canArmDraw()/shouldDisarmDraw(), each with unit tests
(10 cases total). The page now arms the draw cue only via canArmDraw and
disarms via shouldDisarmDraw, and routes t through resolveTrainingMark.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
8c198f22be polish(transcribe): review nits — kbd size, focus ring, guard, action doc (#327)
Review follow-up (Leonie, Felix, Markus): bump cheatsheet key caps to text-sm
for the 60+ audience, add a focus-visible ring to the close button, simplify
the draw-hint guard to {#if drawArmed} (the $effect already clears it outside
edit mode), and document why the transcribeShortcuts action ignores its node
and binds to window.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
6fd05e08d8 test(transcribe): prove Delete fires once via real shape + action (#327)
Review follow-up (Sara): the prior single-owner evidence was two separate
unit facts against an inert DOM stub. This renders a real AnnotationShape,
attaches the live transcribeShortcuts action, focuses the region, and presses
Delete once — asserting deleteCurrentRegion fires exactly once. A genuine
integration guard against re-introducing a double-bind.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
ab469b744c refactor(transcribe): extract region navigation into a tested pure helper (#327)
Review follow-up (Sara): j/k wrap-around and fresh-entry had no direct
coverage — the logic lived inline in the page where the action spec only
mocks the callbacks. Extracted to a pure stepRegion() with 9 unit tests
(empty list, forward/back, both wraps, fresh-entry null + unknown id,
length-1). Also replaces the inline nested ternary Felix flagged.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
f07527158c fix(transcribe): hide the "?" hint on touch-only devices (#327)
Review follow-up (Requirements Engineer, Leonie) — closes the unmet
acceptance row. The coach card's "press ?" tip rendered unconditionally, so
a touch-only tablet transcriber (no hardware keyboard) was told to press a
key they don't have. The hint is now gated behind a fine-pointer media
query ([@media(pointer:coarse)]:hidden); the cheatsheet itself only opens
via the "?" key, so it already never surfaces without a keyboard. Also bumps
the key cap from 11px to text-xs for the 60+ audience.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
9f75de0350 fix(transcribe): localise Delete key cap + annotation label, clarify Esc row (#327)
Review follow-up (Leonie, Requirements Engineer): the Delete key cap was a
hardcoded German "Entf" shown to EN/ES users — now driven by key_cap_delete
(Entf/Del/Supr). The annotation read-only aria-label was a hardcoded German
"Block anzeigen" in all locales — now annotation_view_label. Renamed the Esc
row label from "Bereich schließen" to "Panel schließen" so it no longer
collides with "Bereich" (= region) used elsewhere in the cheatsheet.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
8a9fbc6aef test(transcribe): e2e coverage for shortcuts + cheatsheet a11y (#327)
Seeds a two-block document via API (annotations.spec pattern) and drives the
keyboard: ? opens the cheatsheet, Esc closes it then a second Esc closes the
panel (Esc ladder), e toggles read/edit, and j/k walk the regions forward and
back. Adds an axe-core pass over the open dialog asserting no critical
violations and aria-modal.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
0336d07980 feat(transcribe): surface the "?" shortcut tip in the coach card (#327)
Adds a secondary keyboard hint to the existing coach footer row pointing
transcribers at the "?" cheatsheet, with a semantic <kbd>. Cross-references
the shortcuts introduced for the empty-state coach (#320).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
61256942e1 feat(transcribe): wire keyboard shortcuts into the document panel (#327)
Attaches the transcribeShortcuts action to the document page and wires every
command to existing context setters: j/k walk the sortOrder-sorted regions
and set activeAnnotationId, e toggles read/edit, n arms a draw cue (edit
only), Delete routes to the existing confirm path, ? opens the cheatsheet,
and Esc is now owned solely by the action — the inline onMount Esc listener
is removed (decision B1). Renders ShortcutCheatsheet and a draw-armed hint.

"t" toggles the document-level KURRENT_RECOGNITION training enrollment (the
only training surface that exists; there is no per-region flag yet — see
#321) and no-ops unless a region is active. Also reconciles annotation
Delete: the shape no longer self-handles the key, with onfocus syncing the
active region so the action deletes exactly once.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
6aaf8ddb9e feat(transcribe): add ShortcutCheatsheet dialog overlay (#327)
Native <dialog aria-modal> cheatsheet: showModal()/close() bridge, close
button focused on open, eight grouped <kbd> rows (nav/edit/utility), an
autosave footer line, and a reduced-motion-guarded fade. Closes on Esc,
backdrop click, and the close button; "?" while open is a no-op. Adds the
shortcut_close_panel i18n key. 8 component tests.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
1b9707c6cd feat(transcribe): add transcribeShortcuts keyboard action (#327)
Single-owner window keydown action for the Transcribe panel: j/k region
nav, e mode toggle, n draw (edit only), t training mark, Delete, ? cheat-
sheet, and the Esc precedence ladder (cheatsheet → editable no-op → close
panel). Pure input-to-callback translator with a focus guard that exempts
only "?"; removes its listener on destroy. 20 unit tests cover every key,
the panel/focus guards, the Esc matrix, and teardown.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
8353e71eed feat(transcribe): add i18n keys for shortcut cheatsheet (#327)
Adds de/en/es Paraglide keys for the keyboard-shortcut cheatsheet,
coach hint, draw-armed hint, and the discoverable annotation Delete
aria-label.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-04 17:54:24 +02:00
Marcel
0693cfddd1 fix(document): enlarge auto-title helper to 14px and assert its localized text (#726)
Some checks failed
CI / Unit & Component Tests (pull_request) Failing after 2m35s
CI / OCR Service Tests (pull_request) Successful in 22s
CI / Backend Unit Tests (pull_request) Successful in 3m33s
CI / fail2ban Regex (pull_request) Successful in 48s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
CI / Unit & Component Tests (push) Failing after 2m31s
CI / OCR Service Tests (push) Successful in 21s
CI / Backend Unit Tests (push) Successful in 3m38s
CI / fail2ban Regex (push) Successful in 44s
CI / Semgrep Security Scan (push) Successful in 22s
CI / Compose Bucket Idempotency (push) Successful in 1m6s
Bumps the title helper from text-xs (12px) to text-sm (14px) for the 60+ audience (FR-005
prefers a larger size than the field hints) and tightens the component test to assert the
actual localized string and the 14px class — addresses Leonie's and Sara's review notes.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 17:15:46 +02:00
Marcel
f656f7c1ff test(document): close review-flagged coverage gaps for auto-title sync (#726)
- save-time: precision+raw carry-over when the DTO omits them (exercises the shared skip-null
  resolvers), and a RANGE label round-trip (Sara/Elicit)
- factory: a bare Document with a null index builds "" rather than NPE-ing (Felix)
- backfill matcher: negative near-misses — ASCII hyphen vs en dash, missing separator before
  trailing text, year-with-trailing-letters, index followed by text without a separator (Sara)
- backfill integration: tighten the count assertion to exactly 1 on the clean test DB (Sara)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 17:10:50 +02:00
Marcel
7316c51d4a refactor(document): share skip-null date-field resolution between save and projection (#726)
Extract effectivePrecision/effectiveMetaDateEnd/effectiveMetaDateRaw, used by both
applyDatePrecision (the real setters) and projectedState (the title projection), so the two
can no longer drift — addresses review feedback (Markus/Felix/Sara). Writing a stored value
back when the DTO omits a field is a harmless no-op, so behaviour is unchanged (185 existing
DocumentServiceTest cases stay green). Also documents the file-replace "treat as manual" path
inline at the reassignment site.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 17:08:51 +02:00
Marcel
cf457cb96f docs(document): ADR-031 + glossary/c4/api_tests for auto-title sync (#726)
Some checks failed
CI / Unit & Component Tests (pull_request) Failing after 2m32s
CI / OCR Service Tests (pull_request) Successful in 26s
CI / Backend Unit Tests (pull_request) Successful in 3m35s
CI / fail2ban Regex (pull_request) Successful in 44s
CI / Semgrep Security Scan (pull_request) Successful in 22s
CI / Compose Bucket Idempotency (pull_request) Successful in 1m6s
ADR-031 records the shared document-package title factory, the exact-match save-time
regeneration, and the grammar-heuristic one-time backfill (with the ReDoS / no-version-spam
/ file-replace-is-manual decisions). Adds an "auto-generated title" glossary entry, extends
the document-management c4 diagram with DocumentTitleFactory / DocumentTitleBackfillMatcher
and the backfill flows, and documents POST /api/admin/backfill-titles in Admin-Auth.http as
a one-shot ADMIN call hitting port 8080 directly.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:44:56 +02:00
Marcel
83e0afb466 feat(document): explain auto-generated title under the edit title field (#726)
Adds the FR-TITLE-005 helper line under the title input in DescriptionSection, shown only
on the single-document edit form via a new showTitleHelp prop (off for the new-document and
bulk-edit forms). It is wired to the input with aria-describedby and uses text-ink-3 (WCAG AA
on bg-surface). New Paraglide key form_helper_title_autogenerated in de/en/es. Adds a
component test for the helper + aria wiring and an end-to-end pass: create an auto-titled doc,
edit its date, and see the title follow on the detail page.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:41:52 +02:00
Marcel
12db7b3596 test(document): integration-test title backfill against real Postgres (#726)
Pins backfill behaviour on postgres:16-alpine (H2 unusable — title is NOT NULL): a stale
auto-title is rewritten, the sweep is idempotent (second run touches nothing), prose is
left alone, and the mechanical rename adds no document_versions rows. Permission (401/403)
stays in the faster @WebMvcTest slice.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:32:07 +02:00
Marcel
26b45f1c78 feat(document): one-time backfill endpoint for stale auto-titles (#726)
Adds POST /api/admin/backfill-titles (ADMIN-only, synchronous) which rebuilds every
machine-generated title from the row's current state. A grammar heuristic
(DocumentTitleBackfillMatcher) decides overwritability: index matched literally via
startsWith (originalFilename is user-controlled — no regex injection / ReDoS, CWE-1333),
date-label forms derived from the same Locale.GERMAN formatters as the factory so they
cannot drift, prose left untouched, fail-closed on any surprise. Saves via the repository
directly (no recordVersion — follows backfillFileHashes), so the mechanical rename never
version-spams document_versions. Idempotent: a second run rewrites nothing. Emits one
SLF4J-parameterized scanned/updated/skipped line.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:29:57 +02:00
Marcel
e6ce00035e feat(document): regenerate auto-title on save when date/location change (#726)
updateDocument now captures the machine title from the persisted state before any
setter runs, and rebuilds it from the new state only when the submitted title still
equals that machine value — an exact comparison that relies on the edit form
round-tripping an untouched title verbatim. A hand-written or freshly-typed title is
kept; a blank submission falls back to the rebuilt auto-title (title is always present);
a file-replaced document no longer matches its import-time title and is treated as
manual. projectedState mirrors the setter asymmetry exactly (date/location overwrite
incl. null-clear; precision/end/raw skip-null from the entity).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:20:46 +02:00
Marcel
b1f77bcfb6 refactor(document): extract title composition into shared DocumentTitleFactory (#726)
Move DocumentTitleFormatter from importing into the document package and
introduce DocumentTitleFactory there as the single source of truth for the
{index} – {dateLabel} – {location} formula. DocumentImporter now consumes the
factory instead of owning the composition; the document package owns the rule,
importing depends on it (not the reverse). No behavioral change — importer
title assertions and the #666 fixture parity test stay green.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 16:15:00 +02:00
77 changed files with 9354 additions and 395 deletions

View File

@@ -154,9 +154,9 @@ Schedule monthly automated restore tests. If the restore fails, the backup is wo
```
Every alert needs: description, severity, likely cause, resolution steps, escalation path.
3. **Upgrading VPS tier before profiling**
3. **Upgrading hardware before profiling**
```
# "The app feels slow" → upgrade from CX32 to CX42
# "The app feels slow" → order more RAM / a faster CPU
# Actual cause: unindexed query scanning 100k rows
```
Profile with Grafana dashboards first. Most perceived performance issues are application bugs, not resource constraints.
@@ -404,8 +404,8 @@ Hetzner Object Storage (S3-compatible, replaces MinIO in prod)
Prometheus + Loki + Alertmanager
```
### Monthly Cost: ~23 EUR
CX32 VPS (4 vCPU, 8GB RAM): 17 EUR · Object Storage (~200GB): 5 EUR · SMTP relay: ~1 EUR
### Monthly Cost: ~6 EUR (excl. server)
Hetzner dedicated server (Serverbörse, i7-6700, 64 GB RAM): see invoice · Object Storage (~200GB): 5 EUR · SMTP relay: ~1 EUR
### Reference Documentation
- Full CI workflow, Gitea vs GitHub differences: `docs/infrastructure/ci-gitea.md`

View File

@@ -72,6 +72,25 @@ VITE_SENTRY_DSN=
# Sentry/GlitchTip auth token for source map upload at build time (optional)
SENTRY_AUTH_TOKEN=
# NL search — Ollama LLM inference
# Leave APP_OLLAMA_BASE_URL empty to disable NL search (safe default for CX32 / CI).
# Set to http://ollama:11434 to enable. Requires CX42 (16 GB RAM) to run alongside OCR.
APP_OLLAMA_BASE_URL=http://ollama:11434
# CPU limit: 4.0 is safe on both CX32 (4 vCPUs) and CX42 (8 vCPUs).
# Raise to 7.5 on CX42 for full throughput.
OLLAMA_CPU_LIMIT=4.0
# Memory limit: requires CX42 (16 GB) to run alongside OCR.
# Reduce or set APP_OLLAMA_BASE_URL= on smaller hosts.
OLLAMA_MEM_LIMIT=8g
# Ollama API key — set on the Ollama service to restrict inference API access on archiv-net.
# Generate with: openssl rand -hex 32
# NOTE: Empirically verified that OLLAMA_API_KEY is NOT enforced in Ollama 0.6.5 or 0.30.6 (ADR-028 §7).
# archiv-net network isolation is the only effective access control. Retained for forward compatibility.
OLLAMA_API_KEY=
# Production SMTP — uncomment and fill in to send real emails instead of catching them
# APP_BASE_URL=https://your-domain.example.com
# MAIL_HOST=smtp.example.com

View File

@@ -28,4 +28,18 @@ Authorization: Basic Gast_User gast
###Groups
#GET
GET http://localhost:8080/api/admin/tags
Authorization: Basic admin admin123
Authorization: Basic admin admin123
### One-time backfill: re-sync already-stale auto-titles (#726)
# RUNBOOK: a one-shot ADMIN maintenance call, NOT part of normal operation. Run it ONCE
# after deploying #726 to clean the existing backlog of stale titles (e.g. a title still
# showing "2028" after the date was corrected to "1928"). It is synchronous and idempotent
# — a second run returns {"count": 0} and writes nothing. Hit the backend DIRECTLY on
# port 8080 (NOT through the SvelteKit proxy) so the sweep can't trip the proxy timeout.
# Returns {"count": <documents rewritten>}.
POST http://localhost:8080/api/admin/backfill-titles
Authorization: Basic admin admin123
### NEGATIV-TEST: ein Nicht-Admin darf den Backfill NICHT auslösen -> 403 Forbidden
POST http://localhost:8080/api/admin/backfill-titles
Authorization: Basic Gast_User gast

View File

@@ -41,6 +41,27 @@
<type>pom</type>
<scope>import</scope>
</dependency>
<!-- Force WireMock's ee10 Jetty transitive deps to match Spring Boot's 12.1.8 core -->
<dependency>
<groupId>org.eclipse.jetty.ee10</groupId>
<artifactId>jetty-ee10-servlet</artifactId>
<version>12.1.8</version>
</dependency>
<dependency>
<groupId>org.eclipse.jetty.ee10</groupId>
<artifactId>jetty-ee10-servlets</artifactId>
<version>12.1.8</version>
</dependency>
<dependency>
<groupId>org.eclipse.jetty.ee10</groupId>
<artifactId>jetty-ee10-webapp</artifactId>
<version>12.1.8</version>
</dependency>
<dependency>
<groupId>org.eclipse.jetty</groupId>
<artifactId>jetty-ee</artifactId>
<version>12.1.8</version>
</dependency>
</dependencies>
</dependencyManagement>
<dependencies>
@@ -137,6 +158,12 @@
<artifactId>archunit-junit5</artifactId>
<version>1.3.0</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.wiremock</groupId>
<artifactId>wiremock-jetty12</artifactId>
<version>3.9.2</version>
<scope>test</scope>
</dependency>
<!-- Excel Bearbeitung (Apache POI) -->
<dependency>

View File

@@ -57,6 +57,7 @@ public interface DocumentRepository extends JpaRepository<Document, UUID>, JpaSp
@EntityGraph("Document.full")
List<Document> findByReceiversId(UUID receiverId);
// Callers access only doc.getTags() to mutate the set — receivers/sender not touched; no graph needed.
List<Document> findByTags_Id(UUID tagId);

View File

@@ -32,6 +32,8 @@ import org.springframework.data.domain.Page;
import org.springframework.data.domain.PageRequest;
import org.springframework.data.domain.Pageable;
import org.springframework.data.domain.Sort;
import jakarta.persistence.criteria.JoinType;
import jakarta.persistence.criteria.Predicate;
import org.springframework.data.jpa.domain.Specification;
import org.raddatz.familienarchiv.exception.DomainException;
import org.raddatz.familienarchiv.exception.ErrorCode;
@@ -68,6 +70,7 @@ import static org.raddatz.familienarchiv.document.DocumentSpecifications.*;
public class DocumentService {
private final DocumentRepository documentRepository;
private final DocumentTitleFactory documentTitleFactory;
private final PersonService personService;
private final FileService fileService;
private final TagService tagService;
@@ -379,8 +382,14 @@ public class DocumentService {
DocumentStatus statusBefore = doc.getStatus();
// Auto-title sync (#726): capture the machine title from the CURRENTLY-persisted state
// BEFORE any setter runs — the setters below overwrite date/location and applyDatePrecision
// skips nulls, so the old state must be read first. The submitted title is the catalog
// auto-title iff it equals this; only then does it follow date/location forward.
String autoTitleBefore = documentTitleFactory.build(doc);
// 1. Einfache Felder Update
doc.setTitle(dto.getTitle());
doc.setTitle(resolveTitle(dto.getTitle(), autoTitleBefore, doc, dto));
doc.setDocumentDate(dto.getDocumentDate());
applyDatePrecision(doc, dto);
validateDateRange(doc); // guard before any save (updateDocumentTags below persists)
@@ -424,7 +433,11 @@ public class DocumentService {
doc.setScriptType(dto.getScriptType());
}
// 4. Datei austauschen (nur wenn eine neue ausgewählt wurde)
// 4. Datei austauschen (nur wenn eine neue ausgewählt wurde).
// NB (#726): this reassigns originalFilename to the uploaded file's name. The title's index
// segment is originalFilename, so after a replace the stored title no longer matches
// build(currentState) and the row is treated as manual — neither save-time nor backfill
// rewrites it. Accepted fail-safe (ADR-031), and autoTitleBefore was already captured above.
boolean fileReplaced = newFile != null && !newFile.isEmpty();
if (fileReplaced) {
FileService.UploadResult upload = fileService.uploadFile(newFile, newFile.getOriginalFilename());
@@ -453,22 +466,68 @@ public class DocumentService {
}
/**
* Applies the three date-precision fields only when the DTO carries them.
* A null field means "not submitted" — overwriting the stored value with null
* would fabricate a precision the user never chose, the exact dishonesty #666
* exists to prevent. A row with a genuinely-unknown precision must keep it when
* an unrelated edit (e.g. a location typo) is saved.
* Decides the title to persist on an edit (#726). The submitted title is the catalog
* auto-title only when it equals {@code autoBefore} (built from the stored state) — an exact
* comparison with no heuristic, relying on the edit form round-tripping the stored title
* verbatim when untouched. A machine title is rebuilt from the new state so a corrected
* date/location flows into it; a hand-written or freshly-typed title is kept verbatim. A blank
* submission is never persisted (title is always present) — it falls back to the rebuilt
* auto-title, which always carries at least the index.
*/
private String resolveTitle(String submitted, String autoBefore, Document doc, DocumentUpdateDTO dto) {
if (submitted == null || submitted.isBlank()) {
return documentTitleFactory.build(projectedState(doc, dto));
}
if (!Objects.equals(submitted, autoBefore)) {
return submitted;
}
return documentTitleFactory.build(projectedState(doc, dto));
}
/**
* The document state the regenerated title is built from. It is composed from the SAME
* resolvers the real setters use — {@code documentDate}/{@code location} overwritten from the
* DTO (a null value clears the field), precision/end/raw resolved skip-null via
* {@link #effectivePrecision}/{@link #effectiveMetaDateEnd}/{@link #effectiveMetaDateRaw} — so
* the projection cannot drift from {@link #updateDocument}. The index ({@code originalFilename})
* is never touched by a metadata edit.
*/
private Document projectedState(Document doc, DocumentUpdateDTO dto) {
return Document.builder()
.originalFilename(doc.getOriginalFilename())
.documentDate(dto.getDocumentDate())
.location(dto.getLocation())
.metaDatePrecision(effectivePrecision(doc, dto))
.metaDateEnd(effectiveMetaDateEnd(doc, dto))
.metaDateRaw(effectiveMetaDateRaw(doc, dto))
.build();
}
/**
* Applies the three date-precision fields skip-null: a null DTO field means "not submitted",
* so the stored value is kept rather than overwritten with null — which would fabricate a
* precision the user never chose, the exact dishonesty #666 exists to prevent. Expressed via
* the shared {@code effective*} resolvers so {@link #projectedState} stays lock-step (writing
* the stored value back when the DTO omits a field is a harmless no-op).
*/
private void applyDatePrecision(Document doc, DocumentUpdateDTO dto) {
if (dto.getMetaDatePrecision() != null) {
doc.setMetaDatePrecision(dto.getMetaDatePrecision());
}
if (dto.getMetaDateEnd() != null) {
doc.setMetaDateEnd(dto.getMetaDateEnd());
}
if (dto.getMetaDateRaw() != null) {
doc.setMetaDateRaw(dto.getMetaDateRaw());
}
doc.setMetaDatePrecision(effectivePrecision(doc, dto));
doc.setMetaDateEnd(effectiveMetaDateEnd(doc, dto));
doc.setMetaDateRaw(effectiveMetaDateRaw(doc, dto));
}
// Skip-null date-field resolution shared by applyDatePrecision (the real setters) and
// projectedState (the title projection) — the single rule keeps them from diverging (#726).
private static DatePrecision effectivePrecision(Document doc, DocumentUpdateDTO dto) {
return dto.getMetaDatePrecision() != null ? dto.getMetaDatePrecision() : doc.getMetaDatePrecision();
}
private static LocalDate effectiveMetaDateEnd(Document doc, DocumentUpdateDTO dto) {
return dto.getMetaDateEnd() != null ? dto.getMetaDateEnd() : doc.getMetaDateEnd();
}
private static String effectiveMetaDateRaw(Document doc, DocumentUpdateDTO dto) {
return dto.getMetaDateRaw() != null ? dto.getMetaDateRaw() : doc.getMetaDateRaw();
}
/**
@@ -976,6 +1035,28 @@ public class DocumentService {
return documentRepository.findByReceiversId(receiverId);
}
public DocumentSearchResult searchDocumentsByPersonId(UUID personId, LocalDate from, LocalDate to, Pageable pageable) {
Person person = personService.getById(personId);
Specification<Document> spec = buildPersonSpec(person, from, to);
Page<Document> page = documentRepository.findAll(spec, pageable);
List<DocumentListItem> items = enrichItems(page.getContent(), null);
return DocumentSearchResult.paged(items, pageable, page.getTotalElements());
}
private Specification<Document> buildPersonSpec(Person person, LocalDate from, LocalDate to) {
return (root, query, cb) -> {
if (query != null) query.distinct(true);
var receiversJoin = root.join("receivers", JoinType.LEFT);
var senderPredicate = cb.equal(root.get("sender"), person);
var receiverPredicate = cb.equal(receiversJoin, person);
var personPredicate = cb.or(senderPredicate, receiverPredicate);
var predicates = new ArrayList<>(List.of(personPredicate));
if (from != null) predicates.add(cb.greaterThanOrEqualTo(root.get("documentDate"), from));
if (to != null) predicates.add(cb.lessThanOrEqualTo(root.get("documentDate"), to));
return cb.and(predicates.toArray(new Predicate[0]));
};
}
public long getIncompleteCount() {
return documentRepository.countByMetadataCompleteFalse();
}
@@ -1010,6 +1091,43 @@ public class DocumentService {
tagService.delete(tagId);
}
/**
* One-time cleanup of already-stale auto-titles (#726, FR-003). For every document whose
* stored title passes the {@link DocumentTitleBackfillMatcher} overwrite heuristic, rebuilds
* the title from the row's current state and persists it only when it actually changed.
* Idempotent: a second run rebuilds the same value and saves nothing. Hand-written prose is
* left untouched.
*
* <p>Saves via {@code documentRepository.save} directly — it must NOT route through
* {@link #updateDocument} (which versions every write), following the {@link #backfillFileHashes}
* precedent: a mechanical rename must not snapshot the whole corpus into {@code document_versions}.
*
* @return the number of documents whose title was rewritten
*/
@Transactional
public int backfillTitles() {
List<Document> docs = documentRepository.findAll();
int updated = 0;
int skipped = 0;
for (Document doc : docs) {
if (!DocumentTitleBackfillMatcher.isOverwritable(
doc.getTitle(), doc.getOriginalFilename(), doc.getLocation())) {
skipped++;
continue;
}
String rebuilt = documentTitleFactory.build(doc);
if (rebuilt.equals(doc.getTitle())) {
skipped++; // already correct — keep idempotent, no write
continue;
}
doc.setTitle(rebuilt);
documentRepository.save(doc); // direct save, no recordVersion (mechanical rename)
updated++;
}
log.info("Title backfill complete: scanned={} updated={} skipped={}", docs.size(), updated, skipped);
return updated;
}
@Transactional
public int backfillFileHashes() {
List<Document> docs = documentRepository.findByFileHashIsNullAndFilePathIsNotNull();

View File

@@ -0,0 +1,101 @@
package org.raddatz.familienarchiv.document;
import java.time.LocalDate;
import java.time.format.DateTimeFormatter;
import java.util.LinkedHashSet;
import java.util.Locale;
import java.util.Set;
import java.util.regex.Pattern;
import java.util.stream.Collectors;
/**
* Heuristic overwrite test for the one-time title backfill (#726, FR-004): decides whether a
* STORED title is a machine-generated auto-title (and so may be rebuilt from the row's current
* state) versus hand-written prose (left untouched). Used ONLY by the backfill — save-time
* regeneration uses an exact old-vs-new comparison instead, with no heuristic.
*
* <p>A stored title is overwritable iff, after stripping the literal {@code index} prefix:
* <ol>
* <li>it is exactly {@code {index}}, or</li>
* <li>{@code {index} {dateLabel}} with an optional trailing {@code {location}} segment
* (any location — a present, valid date label is itself strong evidence of a machine
* title), or</li>
* <li>{@code {index} {location}} where the segment equals the document's current location
* (no date label, so the segment must match the known location to be distinguished from
* prose).</li>
* </ol>
*
* <p>Security: the {@code index} is compared <em>literally</em> via {@link String#startsWith}
* (never compiled into a regex) because {@code originalFilename} is user-controlled and may carry
* regex metacharacters — an unquoted pattern would be a ReDoS / regex-injection vector
* (CWE-1333 / CWE-625). The date-label sub-patterns use only bounded, non-nested quantifiers over
* short tokens, so there is no catastrophic backtracking. Fail-closed: any null/blank index or
* structural surprise returns {@code false}.
*/
final class DocumentTitleBackfillMatcher {
private static final String SEPARATOR = " ";
// German month tokens derived from the SAME Locale.GERMAN formatters DocumentTitleFormatter
// uses, so the matcher's accepted spellings cannot drift from what the factory emits (full
// names "Januar"…"Dezember"; abbreviations "Jan."…"Dez." — note May/June/July/März carry no
// period). Pattern.quote each so a "." in an abbreviation is literal, never a wildcard.
private static final String FULL_MONTH = monthAlternation("MMMM");
private static final String ABBR_MONTH = monthAlternation("MMM");
private static final String SEASON = "(?:Frühling|Sommer|Herbst|Winter)";
private static final String YEAR = "\\d{1,4}";
private static final String DAY_NUM = "\\d{1,2}";
// One complete date label, anchored, optionally followed by a free-form trailing location
// segment. Only bounded/non-nested quantifiers over short tokens plus a single trailing
// ".+" → linear, no catastrophic backtracking (FR-004 ReDoS guard).
private static final Pattern DATE_LABEL_WITH_OPTIONAL_LOCATION = Pattern.compile(
"^(?:" + String.join("|",
YEAR, // 1916
"ca\\. " + YEAR, // ca. 1920
FULL_MONTH + " " + YEAR, // Juni 1916
DAY_NUM + "\\. " + FULL_MONTH + " " + YEAR, // 24. Dezember 1943
SEASON + " " + YEAR, // Sommer 1916
"Datum unbekannt",
DAY_NUM + "\\." + DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR, // 10.11. Jan. 1917
DAY_NUM + "\\. " + ABBR_MONTH + " " + DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR, // 30. Jan. 2. Feb. 1917
DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR + " " + DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR, // 30. Dez. 1916 2. Jan. 1917
DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR, // 10. Jan. 1917 (range end == start)
"ab " + DAY_NUM + "\\. " + ABBR_MONTH + " " + YEAR) // ab 10. Jan. 1917
+ ")(?: .+)?$");
private DocumentTitleBackfillMatcher() {
}
static boolean isOverwritable(String title, String index, String location) {
if (title == null || index == null || index.isBlank()) {
return false; // fail closed
}
if (!title.startsWith(index)) {
return false; // index is matched LITERALLY, never as a regex
}
String tail = title.substring(index.length());
if (tail.isEmpty()) {
return true; // exactly {index}
}
if (!tail.startsWith(SEPARATOR)) {
return false;
}
String body = tail.substring(SEPARATOR.length());
if (DATE_LABEL_WITH_OPTIONAL_LOCATION.matcher(body).matches()) {
return true; // {dateLabel} (+ optional trailing location)
}
// No date label: the lone segment must equal the document's current location to be
// distinguished from hand-written prose.
return location != null && !location.isBlank() && body.equals(location);
}
private static String monthAlternation(String pattern) {
DateTimeFormatter formatter = DateTimeFormatter.ofPattern(pattern, Locale.GERMAN);
Set<String> tokens = new LinkedHashSet<>();
for (int month = 1; month <= 12; month++) {
tokens.add(formatter.format(LocalDate.of(2000, month, 15)));
}
return tokens.stream().map(Pattern::quote).collect(Collectors.joining("|", "(?:", ")"));
}
}

View File

@@ -0,0 +1,39 @@
package org.raddatz.familienarchiv.document;
import org.springframework.stereotype.Component;
/**
* Single source of truth for the auto-generated document title
* {@code {index} {dateLabel} {location}}.
*
* <p>The {@code document} package owns this formula; {@code importing} consumes it
* (see ADR for issue #726). The leading {@code index} is the document's
* {@code originalFilename}; the date label is the honest German label produced by
* {@link DocumentTitleFormatter} (the Java half of the #666 date-label split); the
* trailing location is the {@code meta_location} verbatim, omitted when blank.
*/
@Component
public class DocumentTitleFactory {
static final String SEPARATOR = " ";
/**
* Composes the auto-title from the document's current state. The date segment is
* dropped for UNKNOWN precision or a null date (the honest "no date" case); the
* location segment is dropped when blank.
*/
public String build(Document doc) {
// originalFilename is NOT NULL in production; guard only so a synthetic/partial entity
// never trips StringBuilder(null) with an opaque NPE.
StringBuilder title = new StringBuilder(doc.getOriginalFilename() == null ? "" : doc.getOriginalFilename());
if (doc.getDocumentDate() != null && doc.getMetaDatePrecision() != DatePrecision.UNKNOWN) {
title.append(SEPARATOR).append(DocumentTitleFormatter.formatTitleDate(
doc.getDocumentDate(), doc.getMetaDatePrecision(),
doc.getMetaDateEnd(), doc.getMetaDateRaw()));
}
if (doc.getLocation() != null && !doc.getLocation().isBlank()) {
title.append(SEPARATOR).append(doc.getLocation());
}
return title.toString();
}
}

View File

@@ -1,6 +1,4 @@
package org.raddatz.familienarchiv.importing;
import org.raddatz.familienarchiv.document.DatePrecision;
package org.raddatz.familienarchiv.document;
import java.time.LocalDate;
import java.time.format.DateTimeFormatter;

View File

@@ -78,4 +78,8 @@ public class DomainException extends RuntimeException {
public static DomainException tooManyRequests(ErrorCode code, String message, long retryAfterSeconds) {
return new DomainException(code, HttpStatus.TOO_MANY_REQUESTS, message, retryAfterSeconds);
}
public static DomainException serviceUnavailable(ErrorCode code, String message) {
return new DomainException(code, HttpStatus.SERVICE_UNAVAILABLE, message);
}
}

View File

@@ -5,6 +5,7 @@ import lombok.extern.slf4j.Slf4j;
import org.raddatz.familienarchiv.document.DatePrecision;
import org.raddatz.familienarchiv.document.Document;
import org.raddatz.familienarchiv.document.DocumentService;
import org.raddatz.familienarchiv.document.DocumentTitleFactory;
import org.raddatz.familienarchiv.document.DocumentStatus;
import org.raddatz.familienarchiv.document.ThumbnailAsyncRunner;
import org.raddatz.familienarchiv.exception.DomainException;
@@ -74,6 +75,7 @@ public class DocumentImporter {
Pattern.compile("[A-Za-z\\u00C0-\\u00D6\\u00D8-\\u00F6\\u00F8-\\u00FF]{1,4}-+\\d+x?");
private final DocumentService documentService;
private final DocumentTitleFactory documentTitleFactory;
private final PersonService personService;
private final TagService tagService;
private final S3Client s3Client;
@@ -181,7 +183,7 @@ public class DocumentImporter {
applyAttribution(doc, row);
applyDates(doc, row);
applyAuthoritativeAssociations(doc, row);
applyFileMetadata(doc, s3Key, contentType, status, index);
applyFileMetadata(doc, s3Key, contentType, status);
applyComputedFlags(doc);
return doc;
}
@@ -217,14 +219,15 @@ public class DocumentImporter {
attachTag(doc, row.get("tags"));
}
// S3 key, content type, status, and the index-derived title.
// S3 key, content type, status, and the index-derived title. The title formula lives in
// the document package's DocumentTitleFactory (single source of truth, #726); by this point
// applyDates has populated the date/location and originalFilename carries the index.
private void applyFileMetadata(Document doc, String s3Key, String contentType,
DocumentStatus status, String index) {
DocumentStatus status) {
doc.setStatus(status);
doc.setFilePath(s3Key);
doc.setContentType(contentType);
doc.setTitle(buildTitle(index, doc.getDocumentDate(), doc.getMetaDatePrecision(),
doc.getMetaDateEnd(), doc.getMetaDateRaw(), doc.getLocation()));
doc.setTitle(documentTitleFactory.build(doc));
}
// metadataComplete: a document counts as fully described if any of the three "who/when"
@@ -235,20 +238,6 @@ public class DocumentImporter {
|| !doc.getReceivers().isEmpty());
}
// The title carries the date at the HONEST precision (never a fabricated day) via the
// shared DocumentTitleFormatter, plus the location — kept under 20 lines by delegating.
private static String buildTitle(String index, LocalDate date, DatePrecision precision,
LocalDate end, String raw, String location) {
StringBuilder title = new StringBuilder(index);
if (date != null && precision != DatePrecision.UNKNOWN) {
title.append(" ").append(DocumentTitleFormatter.formatTitleDate(date, precision, end, raw));
}
if (location != null && !location.isBlank()) {
title.append(" ").append(location);
}
return title.toString();
}
// ─── attribution routing — register-first, always retain raw ─────────────────────
private Person resolveSender(String slug, String rawName) {

View File

@@ -0,0 +1,13 @@
package org.raddatz.familienarchiv.person;
import java.util.List;
/**
* Result of {@link PersonService#resolveByName(String)}: candidate persons split by name-match
* strength. {@code direct} = every query token is a whole-token match across the person's name
* components (alias/maiden-name aware); {@code partial} = matched the substring fetch but is not
* direct. The vocabulary is deliberately name-match strength ({@code direct}/{@code partial}), not
* the search layer's resolved/ambiguous buckets — the caller maps these into its own outcome.
*/
public record NameMatches(List<Person> direct, List<Person> partial) {
}

View File

@@ -19,7 +19,8 @@ public interface PersonRepository extends JpaRepository<Person, UUID> {
"LOWER(CONCAT(COALESCE(p.firstName, ''),' ',p.lastName)) LIKE LOWER(CONCAT('%', :query, '%')) OR " +
"LOWER(CONCAT(p.lastName, ' ', COALESCE(p.firstName, ''))) LIKE LOWER(CONCAT('%', :query, '%')) OR " +
"LOWER(p.alias) LIKE LOWER(CONCAT('%', :query, '%')) OR " +
"LOWER(a.lastName) LIKE LOWER(CONCAT('%', :query, '%')) " +
"LOWER(a.lastName) LIKE LOWER(CONCAT('%', :query, '%')) OR " +
"LOWER(a.firstName) LIKE LOWER(CONCAT('%', :query, '%')) " +
"ORDER BY p.lastName ASC, p.firstName ASC")
List<Person> searchByName(@Param("query") String query);
@@ -29,14 +30,36 @@ public interface PersonRepository extends JpaRepository<Person, UUID> {
// Stammbaum-Knoten: alle Personen mit family_member = true.
List<Person> findByFamilyMemberTrueOrderByLastNameAscFirstNameAsc();
// Lookup by full alias string, used during ODS mass import
Optional<Person> findByAliasIgnoreCase(String alias);
// Exact-case alias lookup — the first resolution step in findOrCreateByAlias.
// Case-colliding aliases across persons (müller / Müller) are valid human labels, NOT
// duplicates: source_ref is the stable identity (ADR-025/033), alias is editable. Do NOT
// add a unique(lower(alias)) constraint — see ADR-033.
Optional<Person> findByAlias(String alias);
// Plural case-insensitive alias lookup — the fallback step. Returns ALL case-folding
// siblings so the service can pick a deterministic one (lowest id) instead of letting a
// derived Optional<…>IgnoreCase throw NonUniqueResultException. See ADR-033.
List<Person> findAllByAliasIgnoreCase(String alias);
// Lookup by the normalizer person_id, used for idempotent canonical re-import (Phase 3).
Optional<Person> findBySourceRef(String sourceRef);
// Exact first+last name match, used for filename-based sender lookup
Optional<Person> findByFirstNameIgnoreCaseAndLastNameIgnoreCase(String firstName, String lastName);
// Exact-case first+last name match — the first step of filename-based sender resolution.
// Explicit `=` (HQL, not a derived query) so a null firstName binds as `first_name = NULL`
// — never a match — instead of the derived-query fold to `first_name IS NULL`, which would
// pull a last-name-only row in as a sender (a provenance defect). See ADR-033.
@Query("SELECT p FROM Person p WHERE p.firstName = :firstName AND p.lastName = :lastName")
Optional<Person> findByFirstNameAndLastName(@Param("firstName") String firstName,
@Param("lastName") String lastName);
// Plural case-insensitive first+last name match — lets findByName bail to empty on 2+ matches
// instead of letting a derived Optional<…>IgnoreCase throw NonUniqueResultException. Same
// null fail-closed guarantee as above: LOWER(:firstName) is NULL for a null arg, so a null
// first name resolves to no match (not first_name IS NULL widening). See ADR-033.
@Query("SELECT p FROM Person p WHERE LOWER(p.firstName) = LOWER(:firstName) "
+ "AND LOWER(p.lastName) = LOWER(:lastName)")
List<Person> findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase(@Param("firstName") String firstName,
@Param("lastName") String lastName);
// --- PersonSummaryDTO with document count ---
@@ -189,18 +212,15 @@ public interface PersonRepository extends JpaRepository<Person, UUID> {
List<Person> findCorrespondentsWithFilter(@Param("personId") UUID personId, @Param("q") String q);
// --- Merge helpers (native SQL to bypass JPA entity layer) ---
// clearAutomatically + flushAutomatically keep the L1 cache from desyncing: these bulk
// updates run beneath Hibernate, and mergePersons follows them with a deleteById whose
// ON DELETE CASCADE (V71) also fires beneath the session.
@Modifying
@Modifying(clearAutomatically = true, flushAutomatically = true)
@Query(value = "UPDATE documents SET sender_id = :target WHERE sender_id = :source", nativeQuery = true)
void reassignSender(@Param("source") UUID source, @Param("target") UUID target);
// Used by deletePerson: detach a deleted person from documents they sent, so the hard
// delete cannot orphan a documents.sender_id FK (the column is nullable).
@Modifying
@Query(value = "UPDATE documents SET sender_id = NULL WHERE sender_id = :source", nativeQuery = true)
void reassignSenderToNull(@Param("source") UUID source);
@Modifying
@Modifying(clearAutomatically = true, flushAutomatically = true)
@Query(value = """
INSERT INTO document_receivers (document_id, person_id)
SELECT document_id, :target FROM document_receivers
@@ -210,8 +230,4 @@ public interface PersonRepository extends JpaRepository<Person, UUID> {
)
""", nativeQuery = true)
void insertMissingReceiverReference(@Param("source") UUID source, @Param("target") UUID target);
@Modifying
@Query(value = "DELETE FROM document_receivers WHERE person_id = :source", nativeQuery = true)
void deleteReceiverReferences(@Param("source") UUID source);
}
}

View File

@@ -1,8 +1,15 @@
package org.raddatz.familienarchiv.person;
import java.util.ArrayList;
import java.util.Comparator;
import java.util.LinkedHashMap;
import java.util.LinkedHashSet;
import java.util.List;
import java.util.Locale;
import java.util.Optional;
import java.util.Set;
import java.util.UUID;
import java.util.stream.Collectors;
import org.springframework.lang.Nullable;
@@ -23,11 +30,20 @@ import org.springframework.transaction.annotation.Transactional;
import org.springframework.web.server.ResponseStatusException;
import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
@Service
@RequiredArgsConstructor
@Slf4j
public class PersonService {
// Co-located with the fetch loop that owns them (issue #763). MAX_TOKENS caps the number of
// unindexed leading-wildcard LIKE scans per name — a DoS control, not just perf. MAX_CANDIDATES
// bounds each result bucket and is applied AFTER classification so a direct match that sorts
// past position 10 among partials is never discarded.
private static final int MAX_TOKENS = 8;
private static final int MAX_CANDIDATES = 10;
private final PersonRepository personRepository;
private final PersonNameAliasRepository aliasRepository;
@@ -68,15 +84,13 @@ public class PersonService {
}
/**
* Hard-deletes a person used by triage. Detaches the person from any documents they
* sent (nulls sender_id) and from any received-document references first, so the delete
* cannot orphan an FK and fail with a 500.
* Hard-deletes a person used by triage. Referential integrity is enforced by the database
* (V71's {@code ON DELETE} constraints: sender_id {@code SET NULL}, receiver and @-mention
* rows {@code CASCADE}), so the service stays thin — it only verifies existence then deletes.
*/
@Transactional
public void deletePerson(UUID id) {
getById(id);
personRepository.reassignSenderToNull(id);
personRepository.deleteReceiverReferences(id);
personRepository.deleteById(id);
}
@@ -100,6 +114,96 @@ public class PersonService {
return personRepository.findAllById(ids);
}
public List<Person> findByDisplayNameContaining(String fragment) {
return personRepository.searchByName(fragment);
}
// Name-match tokenizer (issue #763): lowercase, split on whitespace/hyphen/apostrophe,
// drop empties. Applied symmetrically to the query and to every candidate name component so
// that "Anna-Maria" and "Anna Maria" tokenize alike. Order-preserving for deterministic tests.
static Set<String> tokenize(String raw) {
if (raw == null || raw.isBlank()) {
return Set.of();
}
LinkedHashSet<String> tokens = new LinkedHashSet<>();
for (String part : raw.toLowerCase(Locale.ROOT).split("[\\s\\-']+")) {
if (!part.isEmpty()) {
tokens.add(part);
}
}
return tokens;
}
/**
* Resolves an extracted person name into {@link NameMatches} by name-match strength.
* Orchestrates tokenize → cap → fetch pool → classify → cap-after-classify. Read-only
* transaction keeps the Hibernate session open so each candidate's lazy {@code nameAliases}
* are reachable during classification (see ADR-022).
*/
@Transactional(readOnly = true)
public NameMatches resolveByName(String name) {
Set<String> queryTokens = capTokens(tokenize(name));
if (queryTokens.isEmpty()) {
log.debug("resolveByName outcome=no-match tokens=0");
return new NameMatches(List.of(), List.of());
}
return classify(fetchPool(queryTokens), queryTokens);
}
private Set<String> capTokens(Set<String> tokens) {
return tokens.stream().limit(MAX_TOKENS).collect(Collectors.toCollection(LinkedHashSet::new));
}
private List<Person> fetchPool(Set<String> queryTokens) {
LinkedHashMap<UUID, Person> pool = new LinkedHashMap<>();
for (String token : queryTokens) {
for (Person candidate : findByDisplayNameContaining(token)) {
pool.putIfAbsent(candidate.getId(), candidate);
}
}
return new ArrayList<>(pool.values());
}
private NameMatches classify(List<Person> pool, Set<String> queryTokens) {
List<Person> direct = new ArrayList<>();
List<Person> partial = new ArrayList<>();
for (Person candidate : pool) {
if (personTokens(candidate).containsAll(queryTokens)) {
direct.add(candidate);
} else {
partial.add(candidate);
}
}
List<Person> cappedDirect = cap(direct);
List<Person> cappedPartial = cap(partial);
log.debug("resolveByName outcome={} tokens={}", outcome(cappedDirect, cappedPartial), queryTokens.size());
return new NameMatches(cappedDirect, cappedPartial);
}
private static Set<String> personTokens(Person person) {
Set<String> tokens = new LinkedHashSet<>();
tokens.addAll(tokenize(person.getFirstName()));
tokens.addAll(tokenize(person.getLastName()));
tokens.addAll(tokenize(person.getAlias()));
tokens.addAll(tokenize(person.getTitle()));
for (PersonNameAlias alias : person.getNameAliases()) {
tokens.addAll(tokenize(alias.getFirstName()));
tokens.addAll(tokenize(alias.getLastName()));
}
return tokens;
}
private static List<Person> cap(List<Person> people) {
return people.size() > MAX_CANDIDATES ? people.subList(0, MAX_CANDIDATES) : people;
}
private static String outcome(List<Person> direct, List<Person> partial) {
if (direct.size() == 1) return "direct=1";
if (direct.size() >= 2) return "direct>=2";
if (!partial.isEmpty()) return "partial-only";
return "no-match";
}
public List<Person> findAllFamilyMembers() {
return personRepository.findByFamilyMemberTrueOrderByLastNameAscFirstNameAsc();
}
@@ -112,7 +216,19 @@ public class PersonService {
}
public Optional<Person> findByName(String firstName, String lastName) {
return personRepository.findByFirstNameIgnoreCaseAndLastNameIgnoreCase(firstName, lastName);
// Same scope as findOrCreateByAlias (#731): a case-collision resolves without throwing;
// two byte-identical same-case persons are an out-of-scope data anomaly the exact
// Optional below would surface as the opaque INTERNAL_ERROR, not a wrong sender.
Optional<Person> exact = personRepository.findByFirstNameAndLastName(firstName, lastName);
if (exact.isPresent()) return exact;
List<Person> caseInsensitive =
personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase(firstName, lastName);
// Deliberate divergence from findOrCreateByAlias: an ambiguous filename leaves the sender
// UNSET rather than picking the lowest id. The archive's value is correct provenance — a
// confidently-wrong pre-filled "Hans Müller" is worse than an empty field, because a
// reviewer won't re-check a pre-filled value. Do NOT "consistency-clean" this into the
// lowest-id fallback. See ADR-033.
return caseInsensitive.size() == 1 ? Optional.of(caseInsensitive.get(0)) : Optional.empty();
}
/** Lookup by the normalizer person_id — used by the canonical importer for register-first matching. */
@@ -127,32 +243,45 @@ public class PersonService {
PersonType type = PersonTypeClassifier.classify(alias);
if (type == PersonType.SKIP) return null;
return personRepository.findByAliasIgnoreCase(alias).orElseGet(() -> {
if (type == PersonType.INSTITUTION || type == PersonType.GROUP) {
return personRepository.save(Person.builder()
.alias(alias)
.lastName(alias)
.personType(type)
.build());
}
// Aliases differing only by case (müller / Müller) are valid distinct persons, not
// duplicates, so a CASE-COLLISION must not throw: exact-case first, then the lowest-id
// case-insensitive sibling, then create. Mirrors the tag path — see ADR-033.
// Scope (#731): "ambiguous" means case-insensitive. Two BYTE-IDENTICAL same-case aliases
// are a true data anomaly out of scope here; the exact Optional below would surface that
// as the opaque INTERNAL_ERROR (never a wrong row), not silently pick one.
Optional<Person> exact = personRepository.findByAlias(alias);
if (exact.isPresent()) return exact.get(); // exact-case wins
List<Person> caseInsensitive = personRepository.findAllByAliasIgnoreCase(alias);
if (!caseInsensitive.isEmpty()) {
return caseInsensitive.stream().min(Comparator.comparing(Person::getId)).orElseThrow(); // deterministic tie-break — list is non-empty, never throws
}
PersonNameParser.SplitName split = PersonNameParser.split(alias);
Person person = personRepository.save(Person.builder()
// Create-when-absent: institution/group keep the full label in lastName; a person name
// is split and a maiden name (geb. …) becomes a MAIDEN_NAME alias.
if (type == PersonType.INSTITUTION || type == PersonType.GROUP) {
return personRepository.save(Person.builder()
.alias(alias)
.firstName(split.firstName())
.lastName(split.lastName())
.lastName(alias)
.personType(type)
.build());
if (split.maidenName() != null) {
int nextSortOrder = aliasRepository.findMaxSortOrder(person.getId()) + 1;
aliasRepository.save(PersonNameAlias.builder()
.person(person)
.lastName(split.maidenName())
.type(PersonNameAliasType.MAIDEN_NAME)
.sortOrder(nextSortOrder)
.build());
}
return person;
});
}
PersonNameParser.SplitName split = PersonNameParser.split(alias);
Person person = personRepository.save(Person.builder()
.alias(alias)
.firstName(split.firstName())
.lastName(split.lastName())
.build());
if (split.maidenName() != null) {
int nextSortOrder = aliasRepository.findMaxSortOrder(person.getId()) + 1;
aliasRepository.save(PersonNameAlias.builder()
.person(person)
.lastName(split.maidenName())
.type(PersonNameAliasType.MAIDEN_NAME)
.sortOrder(nextSortOrder)
.build());
}
return person;
}
/**
@@ -295,6 +424,12 @@ public class PersonService {
return personRepository.save(person);
}
/**
* Merges the source person into the target, then deletes the source. Sender references move
* to the target; receiver references the target lacks are inserted. The source's leftover
* receiver join rows are not deleted explicitly — they cascade-drop via V71's
* {@code ON DELETE CASCADE} on {@code document_receivers.person_id} when the source is deleted.
*/
@Transactional
public void mergePersons(UUID sourceId, UUID targetId) {
if (sourceId.equals(targetId)) {
@@ -311,9 +446,7 @@ public class PersonService {
// Add target as receiver where source is receiver but target is not yet
personRepository.insertMissingReceiverReference(sourceId, targetId);
// Remove all remaining source receiver references (duplicates already handled)
personRepository.deleteReceiverReferences(sourceId);
// Source's remaining receiver rows cascade-drop via V71's ON DELETE CASCADE.
personRepository.deleteById(sourceId);
}

View File

@@ -20,8 +20,9 @@ Features: person CRUD, name alias management, person merge (deduplication), fami
| `getById(UUID)` | document, geschichte, ocr | Fetch one person by ID |
| `getAllById(List<UUID>)` | document | Bulk fetch for sender/receiver resolution |
| `findAll(String q)` | document, dashboard | List all persons |
| `findByName(String firstName, String lastName)` | document | Typeahead search |
| `findOrCreateByAlias(String rawName)` | importing | Idempotent create during mass import; type classification happens internally |
| `findByName(String firstName, String lastName)` | document | Filename-based **sender resolution** in `storeDocument`: exact-case match → single case-insensitive match → else **empty** (ambiguous names leave the sender unset; a null first name never matches). See ADR-033. |
| `resolveByName(String name)` | search | NL-search name resolution returning `NameMatches` (direct vs partial). Token/word-boundary, alias-aware matching so a single direct match auto-selects even when looser substring hits coexist ("Clara Cram" vs "Clara Cramer"). See #763. |
| `findOrCreateByAlias(String rawName)` | importing | Idempotent create during mass import; type classification happens internally. Resolves exact-case → lowest-id case-insensitive sibling → create — never throws on case-colliding aliases. See ADR-033. |
| `findAllFamilyMembers()` | dashboard | Family member list for stats |
| `findCorrespondents()` | document | Correspondent list for conversation filter |
| `count()` | dashboard | Total person count for stats |

View File

@@ -20,7 +20,14 @@ public interface TagRepository extends JpaRepository<Tag, UUID> {
}
Optional<Tag> findByNameIgnoreCase(String name);
// Tag-name resolution (see TagService.findOrCreate). Names that collide case-insensitively across
// the canonical tree are VALID — a parent and its same-named lowercase child (e.g. "Geburt" /
// "Geburt/geburt") are distinct nodes with their own source_ref and document attachments. So
// resolution must be exact-case first, then a non-throwing list for the case-insensitive fallback.
// Do NOT add a unique(lower(name)) constraint — it would reject these legitimate rows. See #730.
Optional<Tag> findByName(String name);
List<Tag> findAllByNameIgnoreCase(String name);
// Lookup by the canonical tag_path, used for idempotent canonical re-import (Phase 3).
Optional<Tag> findBySourceRef(String sourceRef);

View File

@@ -2,6 +2,7 @@ package org.raddatz.familienarchiv.tag;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Comparator;
import java.util.HashMap;
import java.util.HashSet;
import java.util.LinkedHashMap;
@@ -45,6 +46,10 @@ public class TagService {
return enrichWithRelatives(matched);
}
public List<Tag> findByNameContaining(String fragment) {
return tagRepository.findByNameContainingIgnoreCase(fragment);
}
public Tag getById(UUID id) {
return tagRepository.findById(id)
.orElseThrow(() -> DomainException.notFound(ErrorCode.TAG_NOT_FOUND, "Tag not found: " + id));
@@ -55,10 +60,21 @@ public class TagService {
return tagRepository.findBySourceRef(sourceRef);
}
/**
* Resolves a tag name to a single tag, creating one when absent. Never throws on case-insensitive
* collisions: names that differ only by case are valid distinct nodes in the canonical tree (a
* parent and its same-named lowercase child), so resolution prefers an exact-case match, then
* falls back to the lowest-id case-insensitive match, then creates. See #730.
*/
public Tag findOrCreate(String name) {
String cleanName = name.trim();
return tagRepository.findByNameIgnoreCase(cleanName)
.orElseGet(() -> tagRepository.save(Tag.builder().name(cleanName).build()));
Optional<Tag> exact = tagRepository.findByName(cleanName);
if (exact.isPresent()) return exact.get(); // exact-case wins (edit round-trip replays the stored name)
List<Tag> caseInsensitive = tagRepository.findAllByNameIgnoreCase(cleanName);
if (!caseInsensitive.isEmpty()) {
return caseInsensitive.stream().min(Comparator.comparing(Tag::getId)).orElseThrow(); // deterministic tie-break by id — list is non-empty, never throws
}
return tagRepository.save(Tag.builder().name(cleanName).build()); // create-when-absent (orphan tag: null sourceRef/parentId)
}
/**

View File

@@ -51,6 +51,12 @@ public class AdminController {
return ResponseEntity.ok(new BackfillResult(count));
}
@PostMapping("/backfill-titles")
public ResponseEntity<BackfillResult> backfillTitles() {
int count = documentService.backfillTitles();
return ResponseEntity.ok(new BackfillResult(count));
}
@PostMapping("/generate-thumbnails")
public ResponseEntity<ThumbnailBackfillService.BackfillStatus> generateThumbnails() {
thumbnailBackfillService.runBackfillAsync();

View File

@@ -11,3 +11,4 @@ springdoc:
swagger-ui:
enabled: true
path: /swagger-ui.html

View File

@@ -0,0 +1,53 @@
-- Move person-delete referential integrity from application code into the database (#684).
--
-- Before this migration, PersonService.deletePerson nulled documents.sender_id and removed
-- document_receivers rows in Java before deleting the person, because the two V1 FKs into
-- persons had no ON DELETE behaviour. Any other delete path (a future endpoint, a manual
-- psql, a batch job) could still orphan rows or 500. This migration makes the database the
-- single source of truth so a person delete is safe from every path.
--
-- Cascade boundary: the cascade stays STRICTLY at the join/reference layer and NEVER reaches
-- documents rows — a cascade into documents would destroy historical letters. sender_id is
-- SET NULL (documents.senderText preserves the raw textual attribution); the receiver join
-- row and the @-mention sidecar row are dropped.
--
-- No NOT VALID + VALIDATE two-step: these tables are small (thousands of rows → sub-second
-- ACCESS EXCLUSIVE lock). Do NOT copy this drop-and-recreate pattern onto a large table.
--
-- Not audit-logged: a DB ON DELETE cascade runs below AuditService — a known, accepted trade.
-- The person-delete action itself is still logged at the service layer.
-- documents.sender_id → ON DELETE SET NULL (deleted sender clears the link; the document survives).
ALTER TABLE public.documents
DROP CONSTRAINT fkl5xhww7es3b4um01vmly4y18m,
ADD CONSTRAINT fkl5xhww7es3b4um01vmly4y18m
FOREIGN KEY (sender_id) REFERENCES public.persons(id) ON DELETE SET NULL;
-- document_receivers.person_id → ON DELETE CASCADE (drop the join row), the symmetric
-- completion of V14, which added the same to the document_id side of this table.
ALTER TABLE public.document_receivers
DROP CONSTRAINT fkcg7r68qvosqricx1betgrlt7s,
ADD CONSTRAINT fkcg7r68qvosqricx1betgrlt7s
FOREIGN KEY (person_id) REFERENCES public.persons(id) ON DELETE CASCADE;
-- Soft reference fix: transcription_block_mentioned_persons.person_id was a UUID with no FK
-- (V56), so deleting a person left dangling mention rows. Give it a real FK with CASCADE.
-- This reverses V56's deliberate "no FK on person_id" choice — that comment is now historical
-- but is intentionally left untouched, because editing an already-applied migration changes its
-- Flyway checksum and would fail validateOnMigrate in prod. ADR-032 is the authoritative record.
-- Clean up pre-existing orphans first — production likely holds dangling rows because the old
-- deletePerson never cleaned mention rows, and the ADD CONSTRAINT validation scan fails on them.
-- A DO block with RAISE NOTICE surfaces the purge count: Flyway runs each statement via JDBC
-- and discards a trailing SELECT's result set, so a "SELECT count(*)" would log nothing.
DO $$
DECLARE removed int;
BEGIN
DELETE FROM transcription_block_mentioned_persons m
WHERE NOT EXISTS (SELECT 1 FROM persons p WHERE p.id = m.person_id);
GET DIAGNOSTICS removed = ROW_COUNT;
RAISE NOTICE 'V71 orphaned_mention_rows_removed=%', removed;
END $$;
ALTER TABLE public.transcription_block_mentioned_persons
ADD CONSTRAINT fk_tbmp_person
FOREIGN KEY (person_id) REFERENCES public.persons(id) ON DELETE CASCADE;

View File

@@ -624,4 +624,88 @@ class DocumentRepositoryTest {
.reviewed(reviewed)
.build();
}
// ─── searchDocumentsByPersonId (via Specification) ───────────────────────
private Page<Document> searchByPerson(Person person, LocalDate from, LocalDate to) {
Specification<Document> spec = (root, query, cb) -> {
if (query != null) query.distinct(true);
var receiversJoin = root.join("receivers", jakarta.persistence.criteria.JoinType.LEFT);
var personPredicate = cb.or(
cb.equal(root.get("sender"), person),
cb.equal(receiversJoin, person));
var predicates = new java.util.ArrayList<>(java.util.List.of(personPredicate));
if (from != null) predicates.add(cb.greaterThanOrEqualTo(root.get("documentDate"), from));
if (to != null) predicates.add(cb.lessThanOrEqualTo(root.get("documentDate"), to));
return cb.and(predicates.toArray(new jakarta.persistence.criteria.Predicate[0]));
};
return documentRepository.findAll(spec, PageRequest.of(0, 10));
}
@Test
void searchByPersonSpec_returnsDocument_whenPersonIsSender() {
Person person = personRepository.save(Person.builder().lastName("Raddatz").build());
Document doc = documentRepository.save(Document.builder()
.title("Senderbrief").originalFilename("sender.pdf")
.status(DocumentStatus.UPLOADED).sender(person).build());
Page<Document> result = searchByPerson(person, null, null);
assertThat(result.getContent()).extracting(Document::getId).containsExactly(doc.getId());
}
@Test
void searchByPersonSpec_returnsDocument_whenPersonIsReceiver() {
Person person = personRepository.save(Person.builder().lastName("Raddatz").build());
Document doc = documentRepository.save(Document.builder()
.title("Empfängerbrief").originalFilename("receiver.pdf")
.status(DocumentStatus.UPLOADED)
.receivers(new java.util.HashSet<>(List.of(person))).build());
Page<Document> result = searchByPerson(person, null, null);
assertThat(result.getContent()).extracting(Document::getId).containsExactly(doc.getId());
}
@Test
void searchByPersonSpec_returnsDocumentOnce_whenPersonIsBothSenderAndReceiver() {
Person person = personRepository.save(Person.builder().lastName("Raddatz").build());
Document doc = documentRepository.save(Document.builder()
.title("SenderEmpfänger").originalFilename("both.pdf")
.status(DocumentStatus.UPLOADED).sender(person)
.receivers(new java.util.HashSet<>(List.of(person))).build());
Page<Document> result = searchByPerson(person, null, null);
assertThat(result.getContent()).hasSize(1);
assertThat(result.getContent().get(0).getId()).isEqualTo(doc.getId());
}
@Test
void searchByPersonSpec_excludesDocuments_outsideDateRange() {
Person person = personRepository.save(Person.builder().lastName("Raddatz").build());
Document inside = documentRepository.save(Document.builder()
.title("Innen").originalFilename("inside.pdf").status(DocumentStatus.UPLOADED)
.sender(person).documentDate(LocalDate.of(1918, 6, 15)).build());
documentRepository.save(Document.builder()
.title("Außen").originalFilename("outside.pdf").status(DocumentStatus.UPLOADED)
.sender(person).documentDate(LocalDate.of(1920, 1, 1)).build());
Page<Document> result = searchByPerson(person, LocalDate.of(1914, 1, 1), LocalDate.of(1918, 12, 31));
assertThat(result.getContent()).extracting(Document::getId).containsExactly(inside.getId());
}
@Test
void searchByPersonSpec_returnsEmpty_whenNoMatchingDocuments() {
Person person = personRepository.save(Person.builder().lastName("Raddatz").build());
Person other = personRepository.save(Person.builder().lastName("Braun").build());
documentRepository.save(Document.builder()
.title("Fremder Brief").originalFilename("other.pdf")
.status(DocumentStatus.UPLOADED).sender(other).build());
Page<Document> result = searchByPerson(person, null, null);
assertThat(result.getContent()).isEmpty();
}
}

View File

@@ -5,6 +5,7 @@ import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.ArgumentCaptor;
import org.mockito.InjectMocks;
import org.mockito.Mock;
import org.mockito.Spy;
import org.mockito.junit.jupiter.MockitoExtension;
import org.raddatz.familienarchiv.audit.AuditKind;
import org.raddatz.familienarchiv.audit.AuditLogQueryService;
@@ -74,6 +75,9 @@ class DocumentServiceTest {
@Mock AuditLogQueryService auditLogQueryService;
@Mock TranscriptionBlockQueryService transcriptionBlockQueryService;
@Mock ThumbnailAsyncRunner thumbnailAsyncRunner;
// Real factory (pure, dependency-free) so save-time title-regeneration tests exercise the
// shared composition rather than a stub — the #726 single source of truth.
@Spy DocumentTitleFactory documentTitleFactory = new DocumentTitleFactory();
@InjectMocks DocumentService documentService;
// ─── deleteDocument ───────────────────────────────────────────────────────
@@ -228,6 +232,216 @@ class DocumentServiceTest {
assertThat(doc.getMetaDateRaw()).isEqualTo("Juni 1916");
}
// ─── updateDocument save-time auto-title regeneration (#726) ──────────────
//
// Exact old-vs-new comparison: the title is the catalog auto-title iff the submitted
// title equals what the factory builds from the CURRENTLY-persisted state. The edit form
// round-trips the stored title verbatim when untouched, so an equal submission means the
// user did not type over it. makeStored() seeds index/date/precision/location and sets the
// stored title to the matching auto-title, mirroring a freshly-imported row.
private Document makeStored(String index, LocalDate date, DatePrecision precision, String location) {
Document doc = Document.builder()
.id(UUID.randomUUID())
.originalFilename(index)
.documentDate(date)
.metaDatePrecision(precision)
.location(location)
.receivers(new HashSet<>())
.tags(new HashSet<>())
.build();
doc.setTitle(documentTitleFactory.build(doc));
return doc;
}
/** A DTO that round-trips the stored auto-title untouched, with new date/precision/location. */
private static DocumentUpdateDTO editDto(String submittedTitle, LocalDate date,
DatePrecision precision, String location) {
DocumentUpdateDTO dto = new DocumentUpdateDTO();
dto.setTitle(submittedTitle);
dto.setDocumentDate(date);
dto.setMetaDatePrecision(precision);
dto.setLocation(location);
return dto;
}
private Document runUpdate(Document stored, DocumentUpdateDTO dto) throws Exception {
when(documentRepository.findById(stored.getId())).thenReturn(Optional.of(stored));
when(documentRepository.save(any())).thenReturn(stored);
documentService.updateDocument(stored.getId(), dto, null, null);
return stored;
}
@Test
void updateDocument_regeneratesAutoTitle_whenDateChanges() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(2028, 1, 1), DatePrecision.YEAR, "Berlin");
// title untouched ("C-0029 2028 Berlin"), date corrected to 1928
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 1928 Berlin");
}
@Test
void updateDocument_keepsHandWrittenTitle_whenDateChanges() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
stored.setTitle("C-0029 Brief an Mutter"); // hand-written, ≠ auto-title
DocumentUpdateDTO dto = editDto("C-0029 Brief an Mutter", LocalDate.of(1930, 1, 1), DatePrecision.YEAR, null);
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 Brief an Mutter");
}
@Test
void updateDocument_freshlyTypedTitleWins_overRegeneration() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(2028, 1, 1), DatePrecision.YEAR, "Berlin");
// user changed the date AND typed a new title in the same save
DocumentUpdateDTO dto = editDto("Geburtsanzeige", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("Geburtsanzeige");
}
@Test
void updateDocument_regeneratesWithNewDateAndLocation() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(2028, 1, 1), DatePrecision.YEAR, "Berlin");
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "München");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 1928 München");
}
@Test
void updateDocument_dropsTrailingLocationSegment_whenLocationCleared() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
// location cleared (null), title untouched
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 1928");
}
@Test
void updateDocument_regeneratedTitle_doesNotContainOldDate() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(2028, 1, 1), DatePrecision.YEAR, "Berlin");
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).doesNotContain("2028");
}
@Test
void updateDocument_relabelsOnPrecisionChange_yearToDay() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
// stored auto-title "C-0029 1928"; set a full day at DAY precision
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 15), DatePrecision.DAY, null);
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 15. Januar 1928");
}
@Test
void updateDocument_populatesTitle_whenDateAddedToUnknownRow() throws Exception {
Document stored = makeStored("C-0029", null, DatePrecision.UNKNOWN, null);
// stored auto-title is just "C-0029"; add a 1928 YEAR date
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 1928");
}
@Test
void updateDocument_roundTripsSeasonLabel() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1943, 4, 1), DatePrecision.SEASON, null);
stored.setMetaDateRaw("Frühling 1943");
stored.setTitle(documentTitleFactory.build(stored)); // "C-0029 Frühling 1943"
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1943, 4, 1), DatePrecision.SEASON, null);
dto.setMetaDateRaw("Frühling 1943");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 Frühling 1943");
}
@Test
void updateDocument_carriesStoredPrecisionAndRaw_whenDtoOmitsThem() throws Exception {
// Only the year changes; precision/end/raw are omitted from the DTO, so projectedState
// must carry them from the entity (exercises the skip-null effective* resolvers).
Document stored = makeStored("C-0029", LocalDate.of(1943, 4, 1), DatePrecision.SEASON, null);
stored.setMetaDateRaw("Frühling 1943");
stored.setTitle(documentTitleFactory.build(stored)); // "C-0029 Frühling 1943"
DocumentUpdateDTO dto = editDto(stored.getTitle(), LocalDate.of(1944, 4, 1), null, null);
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 Frühling 1944");
}
@Test
void updateDocument_roundTripsRangeLabel_atSaveTime() throws Exception {
Document stored = Document.builder()
.id(UUID.randomUUID())
.originalFilename("C-0029")
.documentDate(LocalDate.of(1917, 1, 10))
.metaDatePrecision(DatePrecision.RANGE)
.metaDateEnd(LocalDate.of(1917, 1, 11))
.receivers(new HashSet<>())
.tags(new HashSet<>())
.build();
stored.setTitle(documentTitleFactory.build(stored)); // "C-0029 10.11. Jan. 1917"
DocumentUpdateDTO dto = new DocumentUpdateDTO();
dto.setTitle(stored.getTitle());
dto.setDocumentDate(LocalDate.of(1918, 1, 10));
dto.setMetaDatePrecision(DatePrecision.RANGE);
dto.setMetaDateEnd(LocalDate.of(1918, 1, 11));
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 10.11. Jan. 1918");
}
@Test
void updateDocument_doesNotRegenerateToBlank_whenSubmittedTitleEmpty() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
DocumentUpdateDTO dto = editDto("", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isNotBlank();
}
@Test
void updateDocument_treatsFileReplacedDoc_asManual() throws Exception {
// originalFilename was reassigned by an earlier file-replace, so the stored title (built
// at import from the old index) no longer matches build(currentState) → treated as manual.
Document stored = makeStored("scan_2024.pdf", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
stored.setTitle("C-0029 1928 Berlin"); // legacy import title, ≠ build("scan_2024.pdf"…)
DocumentUpdateDTO dto = editDto("C-0029 1928 Berlin", LocalDate.of(1930, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo("C-0029 1928 Berlin");
}
@Test
void updateDocument_idempotent_whenNothingChanges() throws Exception {
Document stored = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
String before = stored.getTitle();
DocumentUpdateDTO dto = editDto(before, LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
runUpdate(stored, dto);
assertThat(stored.getTitle()).isEqualTo(before);
}
// ─── updateDocument date-range validation (#678) ──────────────────────────
/** Builds a stored doc ready for an updateDocument call (collections initialised). */
@@ -481,6 +695,59 @@ class DocumentServiceTest {
verify(documentVersionService).recordVersion(any(Document.class));
}
// ─── backfillTitles — one-time stale-title cleanup (#726, FR-003) ─────────
@Test
void backfillTitles_rewritesStaleAutoTitle_andCountsIt() {
Document stale = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
stale.setTitle("C-0029 2028 Berlin"); // stale stored title (date typo never fixed)
when(documentRepository.findAll()).thenReturn(List.of(stale));
when(documentRepository.save(any())).thenReturn(stale);
int count = documentService.backfillTitles();
assertThat(count).isEqualTo(1);
assertThat(stale.getTitle()).isEqualTo("C-0029 1928 Berlin");
verify(documentRepository).save(stale);
}
@Test
void backfillTitles_skipsProse() {
Document prose = makeStored("C-0030", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
prose.setTitle("C-0030 Brief an Mutter");
when(documentRepository.findAll()).thenReturn(List.of(prose));
int count = documentService.backfillTitles();
assertThat(count).isZero();
assertThat(prose.getTitle()).isEqualTo("C-0030 Brief an Mutter");
verify(documentRepository, never()).save(any());
}
@Test
void backfillTitles_isIdempotent_forAlreadyCorrectTitle() {
Document fresh = makeStored("C-0031", LocalDate.of(1940, 1, 1), DatePrecision.YEAR, null);
// title already equals build(current state) → nothing to do
when(documentRepository.findAll()).thenReturn(List.of(fresh));
int count = documentService.backfillTitles();
assertThat(count).isZero();
verify(documentRepository, never()).save(any());
}
@Test
void backfillTitles_neverRecordsVersions() {
Document stale = makeStored("C-0029", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
stale.setTitle("C-0029 2028 Berlin");
when(documentRepository.findAll()).thenReturn(List.of(stale));
when(documentRepository.save(any())).thenReturn(stale);
documentService.backfillTitles();
verify(documentVersionService, never()).recordVersion(any());
}
// ─── thumbnail dispatch ───────────────────────────────────────────────────
@Test

View File

@@ -0,0 +1,90 @@
package org.raddatz.familienarchiv.document;
import org.junit.jupiter.api.Test;
import org.raddatz.familienarchiv.PostgresContainerConfig;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.test.context.SpringBootTest;
import org.springframework.context.annotation.Import;
import org.springframework.test.context.ActiveProfiles;
import org.springframework.test.context.bean.override.mockito.MockitoBean;
import org.springframework.transaction.annotation.Transactional;
import software.amazon.awssdk.services.s3.S3Client;
import java.time.LocalDate;
import java.util.UUID;
import static org.assertj.core.api.Assertions.assertThat;
/**
* End-to-end backfill against a real Postgres (#726, FR-003). H2 is unusable here — the
* {@code title} column is NOT NULL and the title-sync semantics depend on that — so this pins the
* behaviour on {@code postgres:16-alpine}: a stale auto-title is rewritten, the sweep is
* idempotent, prose is left alone, and the mechanical rename writes no {@code document_versions}
* rows. Permission enforcement (401/403) is covered faster by the {@code @WebMvcTest} slice in
* {@code AdminControllerTest}.
*/
@SpringBootTest(webEnvironment = SpringBootTest.WebEnvironment.NONE)
@ActiveProfiles("test")
@Import(PostgresContainerConfig.class)
@Transactional
class DocumentTitleBackfillIntegrationTest {
@MockitoBean S3Client s3Client;
@Autowired DocumentService documentService;
@Autowired DocumentRepository documentRepository;
@Autowired DocumentVersionRepository documentVersionRepository;
private Document persist(String index, String title, LocalDate date, DatePrecision precision, String location) {
return documentRepository.save(Document.builder()
.originalFilename(index)
.title(title)
.documentDate(date)
.metaDatePrecision(precision)
.location(location)
.status(DocumentStatus.PLACEHOLDER)
.build());
}
@Test
void backfill_rewritesStaleAutoTitle() {
Document stale = persist("C-0029", "C-0029 2028 Berlin",
LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
int count = documentService.backfillTitles();
assertThat(count).isEqualTo(1); // exactly the one stale row seeded (clean test DB)
assertThat(documentRepository.findById(stale.getId()).orElseThrow().getTitle())
.isEqualTo("C-0029 1928 Berlin");
}
@Test
void backfill_isIdempotent_secondRunChangesNothing() {
persist("C-0029", "C-0029 2028 Berlin", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
documentService.backfillTitles();
int secondRun = documentService.backfillTitles();
assertThat(secondRun).isZero();
}
@Test
void backfill_skipsProse() {
Document prose = persist("C-0030", "C-0030 Brief an Mutter",
LocalDate.of(1928, 1, 1), DatePrecision.YEAR, null);
documentService.backfillTitles();
assertThat(documentRepository.findById(prose.getId()).orElseThrow().getTitle())
.isEqualTo("C-0030 Brief an Mutter");
}
@Test
void backfill_addsNoDocumentVersionRows() {
persist("C-0029", "C-0029 2028 Berlin", LocalDate.of(1928, 1, 1), DatePrecision.YEAR, "Berlin");
long versionsBefore = documentVersionRepository.count();
documentService.backfillTitles();
assertThat(documentVersionRepository.count()).isEqualTo(versionsBefore);
}
}

View File

@@ -0,0 +1,175 @@
package org.raddatz.familienarchiv.document;
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.Timeout;
import java.util.concurrent.TimeUnit;
import static org.assertj.core.api.Assertions.assertThat;
/**
* The backfill overwrite heuristic (FR-004) in isolation — every emittable date-label form is
* recognised, prose is left alone, and a regex-metacharacter index is matched literally without
* hanging. The exact label spellings mirror {@code docs/date-label-fixtures.json}.
*/
class DocumentTitleBackfillMatcherTest {
private static boolean overwritable(String title, String location) {
return DocumentTitleBackfillMatcher.isOverwritable(title, "C-0029", location);
}
// ─── each date-label form (index + form) is overwritable ──────────────────
@Test
void year_form() {
assertThat(overwritable("C-0029 1916", null)).isTrue();
}
@Test
void approx_form() {
assertThat(overwritable("C-0029 ca. 1920", null)).isTrue();
}
@Test
void month_form() {
assertThat(overwritable("C-0029 Juni 1916", null)).isTrue();
}
@Test
void day_form() {
assertThat(overwritable("C-0029 24. Dezember 1943", null)).isTrue();
}
@Test
void season_form() {
assertThat(overwritable("C-0029 Sommer 1916", null)).isTrue();
}
@Test
void unknown_label_form() {
assertThat(overwritable("C-0029 Datum unbekannt", null)).isTrue();
}
@Test
void range_same_month_form() {
assertThat(overwritable("C-0029 10.11. Jan. 1917", null)).isTrue();
}
@Test
void range_cross_month_form() {
assertThat(overwritable("C-0029 30. Jan. 2. Feb. 1917", null)).isTrue();
}
@Test
void range_cross_year_form() {
assertThat(overwritable("C-0029 30. Dez. 1916 2. Jan. 1917", null)).isTrue();
}
@Test
void range_single_day_form() {
assertThat(overwritable("C-0029 10. Jan. 1917", null)).isTrue();
}
@Test
void range_open_form() {
assertThat(overwritable("C-0029 ab 10. Jan. 1917", null)).isTrue();
}
// ─── date label + trailing location (any location) ────────────────────────
@Test
void date_form_with_trailing_location() {
assertThat(overwritable("C-0029 1916 Berlin", null)).isTrue();
}
@Test
void range_with_internal_separator_plus_trailing_location() {
// The range label itself contains " "; the trailing " Berlin" must still be peeled.
assertThat(overwritable("C-0029 30. Jan. 2. Feb. 1917 Berlin", null)).isTrue();
}
// ─── index-only and index+location cases ──────────────────────────────────
@Test
void exactly_index() {
assertThat(overwritable("C-0029", null)).isTrue();
}
@Test
void index_plus_location_equal_to_current() {
assertThat(overwritable("C-0029 Berlin", "Berlin")).isTrue();
}
// ─── prose is left untouched ──────────────────────────────────────────────
@Test
void prose_segment_not_matching_location_is_skipped() {
assertThat(overwritable("C-0029 Brief an Mutter", "Berlin")).isFalse();
}
@Test
void location_only_segment_is_skipped_when_no_current_location() {
// No date label, and the doc has no location to compare against → cannot prove machine.
assertThat(overwritable("C-0029 Berlin", null)).isFalse();
}
@Test
void title_not_starting_with_index_is_skipped() {
assertThat(overwritable("Ganz anderer Titel", null)).isFalse();
}
// ─── near-miss: shapes that look almost machine-built but are not ──────────
@Test
void ascii_hyphen_instead_of_en_dash_separator_is_skipped() {
// The separator is " " (en dash); a plain " - " is not the machine separator.
assertThat(overwritable("C-0029 - 1916", null)).isFalse();
}
@Test
void date_label_without_separator_before_trailing_text_is_skipped() {
// "1916 Berlin" is not a date label and is not joined by " "; prose, not machine.
assertThat(overwritable("C-0029 1916 Berlin", null)).isFalse();
}
@Test
void year_with_trailing_letters_is_not_a_year_label() {
assertThat(overwritable("C-0029 1916er Brief", null)).isFalse();
}
@Test
void index_immediately_followed_by_text_without_separator_is_skipped() {
assertThat(overwritable("C-0029x 1916", null)).isFalse();
}
// ─── fail-closed guards ───────────────────────────────────────────────────
@Test
void null_title_is_not_overwritable() {
assertThat(DocumentTitleBackfillMatcher.isOverwritable(null, "C-0029", null)).isFalse();
}
@Test
void null_index_is_not_overwritable() {
assertThat(DocumentTitleBackfillMatcher.isOverwritable("C-0029 1916", null, null)).isFalse();
}
@Test
void blank_index_is_not_overwritable() {
assertThat(DocumentTitleBackfillMatcher.isOverwritable(" 1916", " ", null)).isFalse();
}
// ─── ReDoS / regex-metacharacter index is matched literally and terminates ─
@Test
@Timeout(value = 5, unit = TimeUnit.SECONDS)
void index_with_regex_metacharacters_is_matched_literally_and_terminates() {
String hostileIndex = "C-0029(.*).pdf";
// Literal prefix → matches; trailing date label → overwritable. Must not hang.
assertThat(DocumentTitleBackfillMatcher.isOverwritable(
hostileIndex + " 1916", hostileIndex, null)).isTrue();
// A title that does NOT start with the literal hostile index is skipped, also fast.
assertThat(DocumentTitleBackfillMatcher.isOverwritable(
"C-0029 1916", hostileIndex, null)).isFalse();
}
}

View File

@@ -0,0 +1,89 @@
package org.raddatz.familienarchiv.document;
import org.junit.jupiter.api.Test;
import java.time.LocalDate;
import static org.assertj.core.api.Assertions.assertThat;
/**
* The auto-title composition {@code {index} {dateLabel} {location}} in isolation.
* The honest date-label forms themselves are pinned by {@link DocumentTitleFormatterTest}
* against the shared #666 fixture; here we assert only how the factory composes the
* three segments and which segments it omits.
*/
class DocumentTitleFactoryTest {
private final DocumentTitleFactory factory = new DocumentTitleFactory();
private static Document.DocumentBuilder doc(String index) {
return Document.builder()
.originalFilename(index)
.metaDatePrecision(DatePrecision.UNKNOWN);
}
@Test
void index_only_when_no_date_and_no_location() {
assertThat(factory.build(doc("C-0029").build())).isEqualTo("C-0029");
}
@Test
void index_and_year_date() {
Document d = doc("C-0029")
.documentDate(LocalDate.of(1928, 1, 15))
.metaDatePrecision(DatePrecision.YEAR)
.build();
assertThat(factory.build(d)).isEqualTo("C-0029 1928");
}
@Test
void index_date_and_location() {
Document d = doc("C-0029")
.documentDate(LocalDate.of(1928, 1, 15))
.metaDatePrecision(DatePrecision.YEAR)
.location("Berlin")
.build();
assertThat(factory.build(d)).isEqualTo("C-0029 1928 Berlin");
}
@Test
void location_without_date_attaches_directly_to_index() {
Document d = doc("C-0029").location("Berlin").build();
assertThat(factory.build(d)).isEqualTo("C-0029 Berlin");
}
@Test
void unknown_precision_omits_the_date_segment() {
Document d = doc("C-0029")
.documentDate(LocalDate.of(1928, 1, 15))
.metaDatePrecision(DatePrecision.UNKNOWN)
.build();
assertThat(factory.build(d)).isEqualTo("C-0029");
}
@Test
void blank_location_is_omitted() {
Document d = doc("C-0029")
.documentDate(LocalDate.of(1928, 1, 15))
.metaDatePrecision(DatePrecision.YEAR)
.location(" ")
.build();
assertThat(factory.build(d)).isEqualTo("C-0029 1928");
}
@Test
void bare_document_with_null_index_builds_empty_string_not_npe() {
// originalFilename is NOT NULL in production; the guard keeps a synthetic/partial entity
// from tripping StringBuilder(null) with an opaque NPE.
assertThat(factory.build(Document.builder().build())).isEqualTo("");
}
@Test
void day_precision_renders_the_full_german_label() {
Document d = doc("C-0029")
.documentDate(LocalDate.of(1928, 1, 15))
.metaDatePrecision(DatePrecision.DAY)
.build();
assertThat(factory.build(d)).isEqualTo("C-0029 15. Januar 1928");
}
}

View File

@@ -1,10 +1,9 @@
package org.raddatz.familienarchiv.importing;
package org.raddatz.familienarchiv.document;
import com.fasterxml.jackson.databind.JsonNode;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.junit.jupiter.api.DynamicTest;
import org.junit.jupiter.api.TestFactory;
import org.raddatz.familienarchiv.document.DatePrecision;
import java.nio.file.Files;
import java.nio.file.Path;

View File

@@ -0,0 +1,123 @@
package org.raddatz.familienarchiv.document;
import org.junit.jupiter.api.Test;
import org.raddatz.familienarchiv.PostgresContainerConfig;
import org.raddatz.familienarchiv.tag.Tag;
import org.raddatz.familienarchiv.tag.TagRepository;
import org.raddatz.familienarchiv.tag.TagService;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.test.context.SpringBootTest;
import org.springframework.context.annotation.Import;
import org.springframework.test.context.ActiveProfiles;
import org.springframework.test.context.bean.override.mockito.MockitoBean;
import org.springframework.transaction.annotation.Transactional;
import software.amazon.awssdk.services.s3.S3Client;
import java.time.LocalDate;
import java.util.Comparator;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
import java.util.UUID;
import static org.assertj.core.api.Assertions.assertThat;
import static org.assertj.core.api.Assertions.assertThatCode;
/**
* #730 — tag-name resolution against a real Postgres. A mocked repo can't prove the two things that
* actually break: that {@code findAllByNameIgnoreCase} folds case the way Postgres {@code LOWER()}
* does (critical for umlauts like {@code ü}), and that saving a document tagged with a case-colliding
* tag no longer throws {@code NonUniqueResultException}. H2 folds case differently, so this pins the
* behaviour on {@code postgres:16-alpine}. The four-branch resolution logic itself is covered faster
* by the mocked {@code TagServiceTest}.
*/
@SpringBootTest(webEnvironment = SpringBootTest.WebEnvironment.NONE)
@ActiveProfiles("test")
@Import(PostgresContainerConfig.class)
@Transactional
class TagCaseCollisionIntegrationTest {
@MockitoBean S3Client s3Client;
@Autowired DocumentService documentService;
@Autowired DocumentRepository documentRepository;
@Autowired TagRepository tagRepository;
@Autowired TagService tagService;
private Tag persistTag(String name, String sourceRef, UUID parentId) {
return tagRepository.save(Tag.builder().name(name).sourceRef(sourceRef).parentId(parentId).build());
}
private Document persistDocTaggedWith(Tag tag) {
return documentRepository.save(Document.builder()
.originalFilename("C-7301")
.title("Weihnachtsbrief")
.documentDate(LocalDate.of(1928, 1, 1))
.metaDatePrecision(DatePrecision.YEAR)
.status(DocumentStatus.UPLOADED)
.tags(new HashSet<>(Set.of(tag)))
.build());
}
@Test
void updateDocument_succeedsAndKeepsExactChildTag_whenTaggedWithCaseCollidingChild() throws Exception {
Tag parent = persistTag("Weihnachten", "Weihnachten", null);
Tag child = persistTag("weihnachten", "Weihnachten/weihnachten", parent.getId());
Document doc = persistDocTaggedWith(child);
DocumentUpdateDTO dto = new DocumentUpdateDTO();
dto.setTitle("Weihnachtsbrief");
dto.setDocumentDate(LocalDate.of(1930, 1, 1)); // change the date — the field that 500'd on staging
dto.setMetaDatePrecision(DatePrecision.YEAR);
dto.setTags("weihnachten"); // the edit form round-trips the stored child name
assertThatCode(() -> documentService.updateDocument(doc.getId(), dto, null, null))
.doesNotThrowAnyException();
Set<Tag> tags = documentRepository.findById(doc.getId()).orElseThrow().getTags();
assertThat(tags).hasSize(1);
assertThat(tags.iterator().next().getId()).isEqualTo(child.getId()); // child kept, not the parent
}
@Test
void findOrCreate_resolvesUmlautCollisionDeterministically_withoutThrow() {
// The regression catcher: a plain-ASCII pair would stay green even if Postgres folded ü wrongly.
Tag parent = persistTag("Glückwünsche", "Glückwünsche", null);
Tag child = persistTag("glückwünsche", "Glückwünsche/glückwünsche", parent.getId());
// Proof that real Postgres LOWER() folds the umlaut so both rows match case-insensitively.
// Query with the UPPERCASE form findOrCreate actually passes — folding LOWER('GLÜCKWÜNSCHE')
// against LOWER(name) is the exact step under test; a lowercase probe wouldn't exercise it.
assertThat(tagRepository.findAllByNameIgnoreCase("GLÜCKWÜNSCHE")).hasSize(2);
// No exact-case "GLÜCKWÜNSCHE" row exists → resolution falls through to the case-insensitive
// branch with two candidates and must pick the lowest id deterministically, never throwing.
UUID expected = List.of(parent, child).stream().min(Comparator.comparing(Tag::getId)).orElseThrow().getId();
Tag first = tagService.findOrCreate("GLÜCKWÜNSCHE");
Tag second = tagService.findOrCreate("GLÜCKWÜNSCHE");
assertThat(first.getId()).isEqualTo(expected);
assertThat(second.getId()).isEqualTo(first.getId());
}
@Test
void bulkEdit_resolvesCaseCollidingTagThroughFindOrCreate_withoutThrow() {
// Bulk-edit shares resolveTags → findOrCreate; this guards a future refactor that bypasses it.
Tag parent = persistTag("Weihnachten", "Weihnachten", null);
Tag child = persistTag("weihnachten", "Weihnachten/weihnachten", parent.getId());
Document doc = documentRepository.save(Document.builder()
.originalFilename("C-7302")
.title("Brief")
.status(DocumentStatus.UPLOADED)
.build());
DocumentBulkEditDTO dto = new DocumentBulkEditDTO();
dto.setTagNames(List.of("weihnachten"));
assertThatCode(() -> documentService.applyBulkEditToDocument(doc.getId(), dto, null))
.doesNotThrowAnyException();
Set<Tag> tags = documentRepository.findById(doc.getId()).orElseThrow().getTags();
assertThat(tags).hasSize(1);
assertThat(tags.iterator().next().getId()).isEqualTo(child.getId());
}
}

View File

@@ -12,6 +12,8 @@ import org.raddatz.familienarchiv.document.annotation.DocumentAnnotation;
import org.raddatz.familienarchiv.document.DocumentStatus;
import org.raddatz.familienarchiv.document.transcription.PersonMention;
import org.raddatz.familienarchiv.document.transcription.TranscriptionBlock;
import org.raddatz.familienarchiv.person.Person;
import org.raddatz.familienarchiv.person.PersonRepository;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.jdbc.test.autoconfigure.AutoConfigureTestDatabase;
import org.springframework.boot.data.jpa.test.autoconfigure.DataJpaTest;
@@ -30,6 +32,7 @@ class TranscriptionBlockMentionsRepositoryTest {
@Autowired TranscriptionBlockRepository blockRepository;
@Autowired DocumentRepository documentRepository;
@Autowired AnnotationRepository annotationRepository;
@Autowired PersonRepository personRepository;
@Autowired EntityManager em;
private UUID documentId;
@@ -55,8 +58,9 @@ class TranscriptionBlockMentionsRepositoryTest {
@Test
void mentionedPersons_roundTripsTwoEntries() {
UUID auguste = UUID.randomUUID();
UUID hermann = UUID.randomUUID();
// person_id is a real FK since V71 — the mentioned persons must exist.
UUID auguste = personRepository.save(Person.builder().firstName("Auguste").lastName("Raddatz").build()).getId();
UUID hermann = personRepository.save(Person.builder().firstName("Hermann").lastName("Müller").build()).getId();
TranscriptionBlock saved = blockRepository.saveAndFlush(TranscriptionBlock.builder()
.annotationId(annotationId)
@@ -97,8 +101,9 @@ class TranscriptionBlockMentionsRepositoryTest {
@Test
void findByPersonIdWithMentionsFetched_returnsOnlyBlocksReferencingPerson_withMentionsLoaded() {
UUID augusteId = UUID.randomUUID();
UUID hermannId = UUID.randomUUID();
// person_id is a real FK since V71 — the mentioned persons must exist.
UUID augusteId = personRepository.save(Person.builder().firstName("Auguste").lastName("Raddatz").build()).getId();
UUID hermannId = personRepository.save(Person.builder().firstName("Hermann").lastName("Müller").build()).getId();
blockRepository.saveAndFlush(TranscriptionBlock.builder()
.annotationId(annotationId).documentId(documentId)

View File

@@ -12,6 +12,7 @@ import org.mockito.MockedStatic;
import org.mockito.junit.jupiter.MockitoExtension;
import org.slf4j.LoggerFactory;
import org.springframework.dao.DataIntegrityViolationException;
import org.springframework.dao.IncorrectResultSizeDataAccessException;
import org.springframework.http.ResponseEntity;
import static org.assertj.core.api.Assertions.assertThat;
@@ -37,6 +38,30 @@ class GlobalExceptionHandlerTest {
}
}
@Test
void handleGeneric_incorrectResultSize_staysOpaque_noHibernateOrRowCountLeak() {
// #731: before the fix, a case-colliding alias/name made Hibernate throw
// NonUniqueResultException → IncorrectResultSizeDataAccessException, which has no
// dedicated handler and falls through to handleGeneric. The fix removes the throw, but
// this pins the handler: a stray one must stay opaque — no Hibernate class name, no SQL,
// no "2 results were returned" row count reaching the client (CWE-209).
IncorrectResultSizeDataAccessException ex = new IncorrectResultSizeDataAccessException(
"query did not return a unique result: 2 results were returned", 1, 2);
try (MockedStatic<Sentry> sentryMock = mockStatic(Sentry.class)) {
ResponseEntity<GlobalExceptionHandler.ErrorResponse> response = handler.handleGeneric(ex);
assertThat(response.getStatusCode().value()).isEqualTo(500);
assertThat(response.getBody()).isNotNull();
assertThat(response.getBody().code()).isEqualTo(ErrorCode.INTERNAL_ERROR);
assertThat(response.getBody().message())
.isEqualTo("An unexpected error occurred")
.doesNotContain("results were returned")
.doesNotContain("NonUnique")
.doesNotContain("IncorrectResultSize");
}
}
@Test
void handleDataIntegrityViolation_returns400_withoutLeakingConstraint_orSentry() {
// A DataIntegrityViolationException carries the constraint name + SQL in its message;

View File

@@ -11,6 +11,7 @@ import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
import org.raddatz.familienarchiv.document.Document;
import org.raddatz.familienarchiv.document.DocumentService;
import org.raddatz.familienarchiv.document.DocumentTitleFactory;
import org.raddatz.familienarchiv.document.DocumentStatus;
import org.raddatz.familienarchiv.document.ThumbnailAsyncRunner;
import org.raddatz.familienarchiv.person.Person;
@@ -59,8 +60,10 @@ class DocumentImporterTest {
// override this stub locally (load_skipsFile_whenMagicByteCheckThrowsIoException).
lenient().when(fileStreamOpener.open(any(File.class)))
.thenAnswer(inv -> new java.io.FileInputStream(inv.getArgument(0, File.class)));
importer = new DocumentImporter(documentService, personService, tagService, s3Client,
thumbnailAsyncRunner, fileStreamOpener);
// Real factory (pure, dependency-free) so the title-content assertions below exercise
// the shared composition rather than a stub — the #726 single source of truth.
importer = new DocumentImporter(documentService, new DocumentTitleFactory(), personService,
tagService, s3Client, thumbnailAsyncRunner, fileStreamOpener);
ReflectionTestUtils.setField(importer, "bucketName", "test-bucket");
}

View File

@@ -21,6 +21,7 @@ import jakarta.persistence.PersistenceContext;
import java.util.List;
import java.util.Optional;
import java.util.Set;
import java.util.UUID;
import static org.assertj.core.api.Assertions.assertThat;
@@ -120,37 +121,60 @@ class PersonRepositoryTest {
.containsExactly("Anna", "Clara");
}
// ─── findByAliasIgnoreCase ────────────────────────────────────────────────
// ─── findByAlias (exact) / findAllByAliasIgnoreCase (case-folding siblings) ───
@Test
void findByAliasIgnoreCase_returnsMatchingPerson() {
void findByAlias_returnsExactCaseMatchOnly() {
personRepository.save(Person.builder()
.firstName("Karl").lastName("Brandt").alias("Opa Karl").build());
Optional<Person> found = personRepository.findByAliasIgnoreCase("opa karl");
assertThat(found).isPresent();
assertThat(found.get().getFirstName()).isEqualTo("Karl");
assertThat(personRepository.findByAlias("Opa Karl")).isPresent();
assertThat(personRepository.findByAlias("opa karl")).isEmpty(); // exact-case: a folded form does NOT match
}
@Test
void findByAliasIgnoreCase_returnsEmpty_whenAliasDoesNotMatch() {
Optional<Person> found = personRepository.findByAliasIgnoreCase("nobody");
assertThat(found).isEmpty();
void findAllByAliasIgnoreCase_returnsEmpty_whenAliasDoesNotMatch() {
assertThat(personRepository.findAllByAliasIgnoreCase("nobody")).isEmpty();
}
// ─── findByFirstNameIgnoreCaseAndLastNameIgnoreCase ───────────────────────
@Test
void findAllByAliasIgnoreCase_foldsUmlautCase_inRealPostgres() {
// Proves Postgres LOWER() folds ü the same way for both rows — a plain-ASCII probe would
// stay green even if umlaut folding regressed. Both case-colliding aliases must match.
personRepository.save(Person.builder().lastName("Müller").alias("Müller").build());
personRepository.save(Person.builder().lastName("müller").alias("müller").build());
assertThat(personRepository.findAllByAliasIgnoreCase("MÜLLER")).hasSize(2);
}
// ─── findByFirstNameAndLastName (exact) / findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase ───
@Test
void findByFirstNameIgnoreCaseAndLastNameIgnoreCase_returnsMatch() {
void findByFirstNameAndLastName_returnsExactCaseMatchOnly() {
personRepository.save(Person.builder().firstName("Maria").lastName("Raddatz").build());
Optional<Person> found = personRepository.findByFirstNameIgnoreCaseAndLastNameIgnoreCase(
"maria", "raddatz");
assertThat(personRepository.findByFirstNameAndLastName("Maria", "Raddatz")).isPresent();
assertThat(personRepository.findByFirstNameAndLastName("maria", "raddatz")).isEmpty(); // exact-case only
}
assertThat(found).isPresent();
assertThat(found.get().getFirstName()).isEqualTo("Maria");
@Test
void findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase_foldsUmlautCase_inRealPostgres() {
personRepository.save(Person.builder().firstName("Hans").lastName("Müller").build());
personRepository.save(Person.builder().firstName("hans").lastName("müller").build());
assertThat(personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase("HANS", "MÜLLER"))
.hasSize(2);
}
@Test
void findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase_nullFirstName_foldsToNoMatch() {
// Fail-closed: a last-name-only filename (null first name) must NOT widen to first_name IS
// NULL and pull in the institution/last-name-only row as a "sender". Proven on real
// Postgres because a mocked unit test cannot catch the IS NULL vs `= NULL` semantics.
personRepository.save(Person.builder().lastName("Müller").build()); // first_name NULL
assertThat(personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase(null, "Müller"))
.isEmpty();
}
// ─── findCorrespondents ───────────────────────────────────────────────────
@@ -366,30 +390,6 @@ class PersonRepositoryTest {
assertThat(result).hasSize(1);
}
// ─── deleteReceiverReferences ─────────────────────────────────────────────
@Test
void deleteReceiverReferences_removesPersonFromAllDocumentReceivers() {
Person toDelete = personRepository.save(Person.builder().firstName("Weg").lastName("Person").build());
Person sender = personRepository.save(Person.builder().firstName("Send").lastName("Er").build());
Document doc1 = documentRepository.save(Document.builder()
.title("Brief 1").originalFilename("b1.pdf")
.status(DocumentStatus.UPLOADED)
.sender(sender).receivers(Set.of(toDelete)).build());
Document doc2 = documentRepository.save(Document.builder()
.title("Brief 2").originalFilename("b2.pdf")
.status(DocumentStatus.UPLOADED)
.sender(sender).receivers(Set.of(toDelete)).build());
personRepository.deleteReceiverReferences(toDelete.getId());
entityManager.flush();
entityManager.clear();
assertThat(documentRepository.findById(doc1.getId()).orElseThrow().getReceivers()).isEmpty();
assertThat(documentRepository.findById(doc2.getId()).orElseThrow().getReceivers()).isEmpty();
}
// ─── searchByName with aliases ───────────────────────────────────────────
@Test
@@ -428,6 +428,67 @@ class PersonRepositoryTest {
assertThat(results).hasSize(1);
}
@Test
void searchByName_findsByAliasFirstName() {
Person clara = personRepository.save(Person.builder().firstName("Clara").lastName("Cram").build());
aliasRepository.save(PersonNameAlias.builder()
.person(clara).firstName("Wilhelmina").lastName("de Gruyter")
.type(PersonNameAliasType.BIRTH).sortOrder(0).build());
List<Person> results = personRepository.searchByName("Wilhelmina");
assertThat(results).hasSize(1);
assertThat(results.get(0).getLastName()).isEqualTo("Cram");
}
@Test
void searchByName_ordersByLastNameThenFirstName() {
personRepository.save(Person.builder().firstName("Clara").lastName("Cram").build());
personRepository.save(Person.builder().firstName("Anna").lastName("Cram").build());
personRepository.save(Person.builder().firstName("Bernd").lastName("Cram").build());
List<Person> results = personRepository.searchByName("Cram");
assertThat(results).extracting(Person::getFirstName)
.containsExactly("Anna", "Bernd", "Clara");
}
// ─── resolveByName fetch→classify, end-to-end on real Postgres (#763 review) ───
// The classifier unit tests in PersonServiceTest stub searchByName, so they never prove the
// fetch query actually finds an alias-only match and feeds it into classification. These walk
// the whole searchByName → resolveByName path over the real Postgres slice, closing AC#4/#5.
@Test
void resolveByName_maidenAlias_classifiesAsDirect_endToEnd() {
PersonService personService = new PersonService(personRepository, aliasRepository);
Person clara = personRepository.save(Person.builder().firstName("Clara").lastName("Müller").build());
aliasRepository.save(PersonNameAlias.builder()
.person(clara).lastName("Cram").type(PersonNameAliasType.MAIDEN_NAME).sortOrder(0).build());
// Detach so resolveByName re-fetches with its lazy nameAliases loaded from the DB —
// the fresh-session behaviour the @Transactional(readOnly=true) path has in production.
entityManager.flush();
entityManager.clear();
NameMatches matches = personService.resolveByName("Clara Cram");
assertThat(matches.direct()).extracting(Person::getId).containsExactly(clara.getId());
}
@Test
void resolveByName_aliasFirstName_classifiesAsDirect_endToEnd() {
PersonService personService = new PersonService(personRepository, aliasRepository);
Person clara = personRepository.save(Person.builder().firstName("Clara").lastName("Cram").build());
aliasRepository.save(PersonNameAlias.builder()
.person(clara).firstName("Wilhelmina").lastName("de Gruyter")
.type(PersonNameAliasType.BIRTH).sortOrder(0).build());
entityManager.flush();
entityManager.clear();
NameMatches matches = personService.resolveByName("Wilhelmina");
assertThat(matches.direct()).extracting(Person::getId).containsExactly(clara.getId());
}
// ─── searchWithDocumentCount with aliases ────────────────────────────────
@Test
@@ -707,4 +768,146 @@ class PersonRepositoryTest {
assertThat(found).isPresent();
assertThat(found.get().getGeneration()).isNull();
}
// ─── #684: ON DELETE integrity enforced at the database layer ──────────────
// A raw deleteById (bypassing PersonService) must keep referential integrity:
// documents.sender_id → SET NULL, document_receivers.person_id → CASCADE, and the
// transcription_block_mentioned_persons soft reference → CASCADE. These run against
// real Postgres because the FK ON DELETE behaviour never fires on H2.
@Test
void deleteById_personSenderOfAReceiverOfB_nullsSender_dropsReceiverRow_bothDocumentsSurvive() {
Person target = personRepository.save(Person.builder().firstName("Weg").lastName("Person").build());
Person bystander = personRepository.save(Person.builder().firstName("Bleibt").lastName("Hier").build());
Document sent = documentRepository.save(Document.builder()
.title("Gesendet").originalFilename("sent.pdf")
.status(DocumentStatus.UPLOADED).sender(target).build());
Document received = documentRepository.save(Document.builder()
.title("Empfangen").originalFilename("received.pdf")
.status(DocumentStatus.UPLOADED).sender(bystander)
.receivers(Set.of(target)).build());
entityManager.flush();
entityManager.clear();
personRepository.deleteById(target.getId());
entityManager.flush();
entityManager.clear();
assertThat(personRepository.findById(target.getId())).isEmpty();
Document reloadedSent = documentRepository.findById(sent.getId()).orElseThrow();
assertThat(reloadedSent.getSender()).isNull(); // AC-1: SET NULL
Document reloadedReceived = documentRepository.findById(received.getId()).orElseThrow();
assertThat(reloadedReceived.getReceivers())
.noneMatch(p -> p.getId().equals(target.getId())); // AC-2: CASCADE drops the join row
// Cascade-boundary guard (Nora, non-negotiable): the cascade stops at the join/reference
// layer — both documents themselves survive. Guards against a future migration turning
// documents.sender_id SET NULL into CASCADE and destroying historical letters.
assertThat(documentRepository.findById(sent.getId())).isPresent();
assertThat(documentRepository.findById(received.getId())).isPresent();
}
@Test
void deleteById_receiverWithCoReceiver_dropsOnlyDeletedPersonsJoinRow() {
Person target = personRepository.save(Person.builder().firstName("Weg").lastName("Person").build());
Person coReceiver = personRepository.save(Person.builder().firstName("Mit").lastName("Empfänger").build());
Person sender = personRepository.save(Person.builder().firstName("Send").lastName("Er").build());
Document doc = documentRepository.save(Document.builder()
.title("Brief").originalFilename("brief.pdf")
.status(DocumentStatus.UPLOADED).sender(sender)
.receivers(Set.of(target, coReceiver)).build());
entityManager.flush();
entityManager.clear();
personRepository.deleteById(target.getId());
entityManager.flush();
entityManager.clear();
Document reloaded = documentRepository.findById(doc.getId()).orElseThrow();
assertThat(reloaded.getReceivers()).extracting(Person::getId)
.containsExactly(coReceiver.getId()); // co-receiver untouched
}
@Test
void deleteById_personIsSenderAndReceiverOfSameDocument_documentSurvives_senderNull_receiverDropped() {
// AC-8: the trickier same-document interaction the cross-document cases don't exercise.
Person target = personRepository.save(Person.builder().firstName("Beides").lastName("Person").build());
Person coReceiver = personRepository.save(Person.builder().firstName("Mit").lastName("Empfänger").build());
Document doc = documentRepository.save(Document.builder()
.title("Selbstbrief").originalFilename("self.pdf")
.status(DocumentStatus.UPLOADED).sender(target)
.receivers(Set.of(target, coReceiver)).build());
entityManager.flush();
entityManager.clear();
personRepository.deleteById(target.getId());
entityManager.flush();
entityManager.clear();
Document reloaded = documentRepository.findById(doc.getId()).orElseThrow();
assertThat(reloaded.getSender()).isNull();
assertThat(reloaded.getReceivers()).extracting(Person::getId)
.containsExactly(coReceiver.getId());
}
@Test
void deleteById_mentionedPerson_dropsMentionRow_blockTextSurvives() {
// AC-3: the @-mention sidecar is a CASCADE soft reference, but the literal "@Name" lives
// in transcription_blocks.text and must stay visible as plain text after the person goes.
Person mentioned = personRepository.save(Person.builder().firstName("Auguste").lastName("Raddatz").build());
Person survivor = personRepository.save(Person.builder().firstName("Clara").lastName("Cram").build());
Document doc = documentRepository.save(Document.builder()
.title("Brief").originalFilename("brief.pdf")
.status(DocumentStatus.UPLOADED).build());
entityManager.flush();
UUID annotationId = UUID.randomUUID();
UUID blockId = UUID.randomUUID();
entityManager.createNativeQuery(
"INSERT INTO document_annotations (id, document_id, page_number, x, y, width, height, color) "
+ "VALUES (?1, ?2, 1, 0.1, 0.2, 0.3, 0.1, '#fff')")
.setParameter(1, annotationId).setParameter(2, doc.getId()).executeUpdate();
entityManager.createNativeQuery(
"INSERT INTO transcription_blocks (id, annotation_id, document_id, text) VALUES (?1, ?2, ?3, ?4)")
.setParameter(1, blockId).setParameter(2, annotationId).setParameter(3, doc.getId())
.setParameter(4, "Brief an @Auguste Raddatz und @Clara Cram").executeUpdate();
// Two mention rows on the same block: the deleted person and an innocent bystander.
entityManager.createNativeQuery(
"INSERT INTO transcription_block_mentioned_persons (block_id, person_id, display_name) "
+ "VALUES (?1, ?2, ?3)")
.setParameter(1, blockId).setParameter(2, mentioned.getId())
.setParameter(3, "Auguste Raddatz").executeUpdate();
entityManager.createNativeQuery(
"INSERT INTO transcription_block_mentioned_persons (block_id, person_id, display_name) "
+ "VALUES (?1, ?2, ?3)")
.setParameter(1, blockId).setParameter(2, survivor.getId())
.setParameter(3, "Clara Cram").executeUpdate();
entityManager.flush();
entityManager.clear();
personRepository.deleteById(mentioned.getId());
entityManager.flush();
entityManager.clear();
Number mentionRows = (Number) entityManager.createNativeQuery(
"SELECT count(*) FROM transcription_block_mentioned_persons WHERE person_id = ?1")
.setParameter(1, mentioned.getId()).getSingleResult();
assertThat(mentionRows.longValue()).isZero();
// The cascade is scoped to the deleted person — the bystander's mention row is untouched.
Number survivorRows = (Number) entityManager.createNativeQuery(
"SELECT count(*) FROM transcription_block_mentioned_persons WHERE person_id = ?1")
.setParameter(1, survivor.getId()).getSingleResult();
assertThat(survivorRows.longValue()).isEqualTo(1);
String text = (String) entityManager.createNativeQuery(
"SELECT text FROM transcription_blocks WHERE id = ?1")
.setParameter(1, blockId).getSingleResult();
assertThat(text).isEqualTo("Brief an @Auguste Raddatz und @Clara Cram");
}
}

View File

@@ -4,6 +4,7 @@ import org.junit.jupiter.api.Test;
import org.raddatz.familienarchiv.PostgresContainerConfig;
import org.raddatz.familienarchiv.document.Document;
import org.raddatz.familienarchiv.document.DocumentRepository;
import org.raddatz.familienarchiv.document.DocumentService;
import org.raddatz.familienarchiv.document.DocumentStatus;
import org.raddatz.familienarchiv.person.Person;
import org.raddatz.familienarchiv.person.PersonType;
@@ -16,10 +17,13 @@ import org.springframework.test.context.bean.override.mockito.MockitoBean;
import org.springframework.transaction.annotation.Transactional;
import software.amazon.awssdk.services.s3.S3Client;
import org.springframework.mock.web.MockMultipartFile;
import jakarta.persistence.EntityManager;
import jakarta.persistence.PersistenceContext;
import java.util.Set;
import java.util.UUID;
import static org.assertj.core.api.Assertions.assertThat;
@@ -33,6 +37,7 @@ class PersonServiceIntegrationTest {
@Autowired PersonService personService;
@Autowired PersonRepository personRepository;
@Autowired DocumentRepository documentRepository;
@Autowired DocumentService documentService;
@PersistenceContext EntityManager entityManager;
@@ -75,6 +80,93 @@ class PersonServiceIntegrationTest {
assertThat(result.getLastName()).isEqualTo("Cram");
}
// ─── #731: case-colliding alias resolution against real Postgres ───────────
// The umlaut pair is mandatory — only the real DB proves Postgres LOWER() folds ü; a
// plain-ASCII test would stay green while umlaut aliases regressed.
@Test
void findOrCreateByAlias_resolvesUmlautAliasCollision_toLowestId_withoutThrow() {
Person muller = personRepository.save(Person.builder().lastName("Müller").alias("Müller").build());
Person mullerLower = personRepository.save(Person.builder().lastName("müller").alias("müller").build());
UUID expected = muller.getId().compareTo(mullerLower.getId()) <= 0 ? muller.getId() : mullerLower.getId();
// No exact-case "MÜLLER" row → falls through to the case-insensitive branch with two
// candidates and must pick the lowest id, never throwing NonUniqueResultException.
Person resolved = personService.findOrCreateByAlias("MÜLLER");
assertThat(resolved.getId()).isEqualTo(expected);
}
@Test
void findOrCreateByAlias_umlautAliasCollision_isDeterministicAcrossCalls() {
personRepository.save(Person.builder().lastName("Müller").alias("Müller").build());
personRepository.save(Person.builder().lastName("müller").alias("müller").build());
Person first = personService.findOrCreateByAlias("MÜLLER");
Person second = personService.findOrCreateByAlias("MÜLLER");
assertThat(second.getId()).isEqualTo(first.getId());
}
// ─── #731: filename-based sender resolution against real Postgres ──────────
@Test
void storeDocument_resolvesSender_whenFilenameNameIsUnique() throws Exception {
Person hans = personRepository.save(Person.builder().firstName("Hans").lastName("Müller").build());
Document doc = uploadNamed("1965-03-12_Müller_Hans.pdf").document();
assertThat(doc.getSender()).isNotNull();
assertThat(doc.getSender().getId()).isEqualTo(hans.getId());
}
@Test
void storeDocument_resolvesSender_onSingleCaseInsensitiveMatch() throws Exception {
Person hans = personRepository.save(Person.builder().firstName("Hans").lastName("Müller").build());
// Filename folds to "hans müller"; the only stored person is "Hans Müller".
Document doc = uploadNamed("1965-03-12_müller_hans.pdf").document();
assertThat(doc.getSender()).isNotNull();
assertThat(doc.getSender().getId()).isEqualTo(hans.getId());
}
@Test
void storeDocument_leavesSenderUnset_whenFilenameNameIsAmbiguous() throws Exception {
// Two persons collide case-insensitively; the filename casing ("HANS"/"MÜLLER") matches
// neither exactly → no exact-case winner → bail to null (never an arbitrary guess), no 500.
personRepository.save(Person.builder().firstName("Hans").lastName("Müller").build());
personRepository.save(Person.builder().firstName("hans").lastName("müller").build());
Document doc = uploadNamed("1965-03-12_MÜLLER_HANS.pdf").document();
assertThat(doc.getSender()).isNull();
}
@Test
void storeDocument_leavesSenderUnset_whenFilenameHasNoFirstName() throws Exception {
// A last-name-only filename never resolves to a sender (the parser yields no parsed name).
personRepository.save(Person.builder().lastName("Müller").build());
Document doc = uploadNamed("1965-03-12_Müller.pdf").document();
assertThat(doc.getSender()).isNull();
}
@Test
void findByName_nullFirstName_resolvesToEmpty_inRealPostgres() {
// Fail-closed against the real DB: a null first name must NOT widen to first_name IS NULL
// and pick up the last-name-only row.
personRepository.save(Person.builder().lastName("Müller").build()); // first_name NULL
assertThat(personService.findByName(null, "Müller")).isEmpty();
}
private DocumentService.StoreResult uploadNamed(String filename) throws Exception {
MockMultipartFile file = new MockMultipartFile("file", filename, "application/pdf", new byte[]{1, 2, 3});
return documentService.storeDocument(file, null);
}
// ─── #667: confirm round-trip + reader-default semantics ──────────────────
@Test
@@ -180,9 +272,9 @@ class PersonServiceIntegrationTest {
@Test
void deletePerson_detachesSentAndReceivedReferences_beforeDelete_noOrphan() {
// A person referenced as BOTH a document sender and a document receiver must delete
// cleanly: deletePerson nulls the sender_id FK and removes the receiver join row first
// (reassignSenderToNull → deleteReceiverReferences → deleteById), so no FK orphan and
// the documents themselves survive.
// cleanly via the service path: deletePerson just calls deleteById, and V71's ON DELETE
// constraints null the sender_id FK and drop the receiver join row, so there is no FK
// orphan and the documents themselves survive.
Person target = personRepository.save(Person.builder()
.firstName("Weg").lastName("Person").provisional(true).build());
Person bystander = personRepository.save(Person.builder()
@@ -196,16 +288,16 @@ class PersonServiceIntegrationTest {
.status(DocumentStatus.UPLOADED).sender(bystander)
.receivers(new java.util.HashSet<>(Set.of(target))).build());
// Persist the fixture and detach everything so the native @Modifying deletes operate on
// the database directly without the persistence context holding stale references that
// would re-flush a now-deleted person as a transient association.
// Persist the fixture and detach everything so the delete operates on the database
// directly without the persistence context holding stale references.
entityManager.flush();
entityManager.clear();
personService.deletePerson(target.getId());
// Native @Modifying queries bypass the persistence context — clear it so the asserting
// reads observe the post-delete database state, not stale managed entities.
// The ON DELETE cascade fires beneath Hibernate — flush the delete and clear the L1
// cache so the asserting reads observe the post-delete database state, not stale
// managed entities still holding the dropped sender/receiver associations.
entityManager.flush();
entityManager.clear();
@@ -220,4 +312,38 @@ class PersonServiceIntegrationTest {
// The other person and the documents themselves survive the delete.
assertThat(personRepository.findById(bystander.getId())).isPresent();
}
@Test
void mergePersons_targetInheritsReferences_sourceJoinRowCascadeDrops_noFkError() {
// AC-7: merging a source who is sender of A and receiver of B into a target leaves the
// target as sender of A and receiver of B, drops the source's leftover receiver row via
// V71's ON DELETE CASCADE (no explicit delete, no FK error), and co-receivers are intact.
Person source = personRepository.save(Person.builder().firstName("Anna").lastName("Alt").build());
Person target = personRepository.save(Person.builder().firstName("Anna").lastName("Neu").build());
Person coReceiver = personRepository.save(Person.builder().firstName("Mit").lastName("Empfänger").build());
Person sender = personRepository.save(Person.builder().firstName("Send").lastName("Er").build());
Document docA = documentRepository.save(Document.builder()
.title("Von Anna").originalFilename("a.pdf")
.status(DocumentStatus.UPLOADED).sender(source).build());
Document docB = documentRepository.save(Document.builder()
.title("An Anna").originalFilename("b.pdf")
.status(DocumentStatus.UPLOADED).sender(sender)
.receivers(new java.util.HashSet<>(Set.of(source, coReceiver))).build());
entityManager.flush();
entityManager.clear();
personService.mergePersons(source.getId(), target.getId());
entityManager.flush();
entityManager.clear();
assertThat(personRepository.findById(source.getId())).isEmpty();
Document reloadedA = documentRepository.findById(docA.getId()).orElseThrow();
assertThat(reloadedA.getSender().getId()).isEqualTo(target.getId());
Document reloadedB = documentRepository.findById(docB.getId()).orElseThrow();
assertThat(reloadedB.getReceivers()).extracting(Person::getId)
.containsExactlyInAnyOrder(target.getId(), coReceiver.getId());
}
}

View File

@@ -27,6 +27,7 @@ import static org.mockito.ArgumentMatchers.any;
import static org.mockito.ArgumentMatchers.argThat;
import static org.mockito.Mockito.never;
import static org.mockito.Mockito.verify;
import static org.mockito.Mockito.verifyNoMoreInteractions;
import static org.mockito.Mockito.when;
@ExtendWith(MockitoExtension.class)
@@ -147,9 +148,11 @@ class PersonServiceTest {
personService.deletePerson(id);
verify(personRepository).reassignSenderToNull(id);
verify(personRepository).deleteReceiverReferences(id);
// Integrity is enforced by V71's ON DELETE constraints — the service only checks
// existence then deletes; it no longer detaches sender/receiver references itself.
verify(personRepository).findById(id);
verify(personRepository).deleteById(id);
verifyNoMoreInteractions(personRepository);
}
@Test
@@ -372,14 +375,57 @@ class PersonServiceTest {
// ─── findOrCreateByAlias ─────────────────────────────────────────────────
@Test
void findOrCreateByAlias_returnsExisting_whenAliasFound() {
String alias = "Walter de Gruyter";
Person existing = Person.builder().id(UUID.randomUUID()).alias(alias).build();
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.of(existing));
void findOrCreateByAlias_returnsExactCaseMatch_overCaseInsensitiveSibling() {
String alias = "müller";
Person exact = Person.builder().id(UUID.randomUUID()).alias("müller").build();
when(personRepository.findByAlias(alias)).thenReturn(Optional.of(exact));
Person result = personService.findOrCreateByAlias(alias);
assertThat(result).isEqualTo(existing);
assertThat(result).isEqualTo(exact);
verify(personRepository, never()).findAllByAliasIgnoreCase(any());
verify(personRepository, never()).save(any());
}
@Test
void findOrCreateByAlias_returnsExactCaseMatch_evenWhenMultipleSiblingsCollide() {
String alias = "Müller";
Person exact = Person.builder().id(UUID.randomUUID()).alias("Müller").build();
when(personRepository.findByAlias(alias)).thenReturn(Optional.of(exact));
Person result = personService.findOrCreateByAlias(alias);
assertThat(result).isEqualTo(exact);
// exact-case short-circuits — the case-insensitive siblings are never consulted.
verify(personRepository, never()).findAllByAliasIgnoreCase(any());
}
@Test
void findOrCreateByAlias_usesSingleCaseInsensitiveMatch_whenNoExactCase() {
String alias = "müller";
Person only = Person.builder().id(UUID.randomUUID()).alias("Müller").build();
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of(only));
Person result = personService.findOrCreateByAlias(alias);
assertThat(result).isEqualTo(only);
verify(personRepository, never()).save(any());
}
@Test
void findOrCreateByAlias_returnsLowestIdDeterministically_whenMultipleCaseInsensitiveMatches() {
String alias = "müller";
Person lower = Person.builder().id(UUID.fromString("00000000-0000-0000-0000-000000000001")).alias("Müller").build();
Person higher = Person.builder().id(UUID.fromString("00000000-0000-0000-0000-000000000002")).alias("müller").build();
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of(higher, lower)); // unordered
Person first = personService.findOrCreateByAlias(alias);
Person second = personService.findOrCreateByAlias(alias);
assertThat(first.getId()).isEqualTo(lower.getId()); // lowest id wins
assertThat(second.getId()).isEqualTo(first.getId()); // same result every call — never throws
verify(personRepository, never()).save(any());
}
@@ -387,7 +433,8 @@ class PersonServiceTest {
void findOrCreateByAlias_createsNew_whenAliasNotFound() {
String alias = "Clara Cram";
Person saved = Person.builder().id(UUID.randomUUID()).alias(alias).firstName("Clara").lastName("Cram").build();
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.empty());
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of());
when(personRepository.save(any())).thenReturn(saved);
Person result = personService.findOrCreateByAlias(alias);
@@ -400,7 +447,8 @@ class PersonServiceTest {
void findOrCreateByAlias_createsMaidenNameAlias_whenGebPresent() {
String alias = "Clara Cram geb. de Gruyter";
Person saved = Person.builder().id(UUID.randomUUID()).alias(alias).firstName("Clara").lastName("Cram").build();
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.empty());
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of());
when(personRepository.save(any())).thenReturn(saved);
when(aliasRepository.findMaxSortOrder(saved.getId())).thenReturn(0);
when(aliasRepository.save(any())).thenAnswer(inv -> inv.getArgument(0));
@@ -422,7 +470,8 @@ class PersonServiceTest {
@Test
void findOrCreateByAlias_setsInstitutionType_withFullNameInLastName() {
String alias = "Arthur Collignon GmbH";
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.empty());
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of());
when(personRepository.save(any())).thenAnswer(inv -> {
Person p = inv.getArgument(0);
p.setId(UUID.randomUUID());
@@ -439,7 +488,8 @@ class PersonServiceTest {
@Test
void findOrCreateByAlias_setsGroupType_withFullNameInLastName() {
String alias = "Geschwister de Gruyter";
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.empty());
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of());
when(personRepository.save(any())).thenAnswer(inv -> {
Person p = inv.getArgument(0);
p.setId(UUID.randomUUID());
@@ -457,7 +507,8 @@ class PersonServiceTest {
void findOrCreateByAlias_noAlias_whenNoGeb() {
String alias = "Clara Cram";
Person saved = Person.builder().id(UUID.randomUUID()).alias(alias).firstName("Clara").lastName("Cram").build();
when(personRepository.findByAliasIgnoreCase(alias)).thenReturn(Optional.empty());
when(personRepository.findByAlias(alias)).thenReturn(Optional.empty());
when(personRepository.findAllByAliasIgnoreCase(alias)).thenReturn(List.of());
when(personRepository.save(any())).thenReturn(saved);
personService.findOrCreateByAlias(alias);
@@ -469,11 +520,54 @@ class PersonServiceTest {
void findOrCreateByAlias_trimsInput() {
String alias = " Clara Cram ";
Person saved = Person.builder().id(UUID.randomUUID()).alias("Clara Cram").build();
when(personRepository.findByAliasIgnoreCase("Clara Cram")).thenReturn(Optional.of(saved));
when(personRepository.findByAlias("Clara Cram")).thenReturn(Optional.of(saved));
personService.findOrCreateByAlias(alias);
verify(personRepository).findByAliasIgnoreCase("Clara Cram");
verify(personRepository).findByAlias("Clara Cram");
}
// ─── findByName (filename-based sender resolution) ────────────────────────
@Test
void findByName_returnsExactCaseMatch_overCaseInsensitiveSibling() {
Person exact = Person.builder().id(UUID.randomUUID()).firstName("Hans").lastName("Müller").build();
when(personRepository.findByFirstNameAndLastName("Hans", "Müller")).thenReturn(Optional.of(exact));
assertThat(personService.findByName("Hans", "Müller")).contains(exact);
verify(personRepository, never()).findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase(any(), any());
}
@Test
void findByName_usesSingleCaseInsensitiveMatch_whenNoExactCase() {
Person only = Person.builder().id(UUID.randomUUID()).firstName("Hans").lastName("Müller").build();
when(personRepository.findByFirstNameAndLastName("hans", "müller")).thenReturn(Optional.empty());
when(personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase("hans", "müller"))
.thenReturn(List.of(only));
assertThat(personService.findByName("hans", "müller")).contains(only);
}
@Test
void findByName_bailsToEmpty_whenTwoOrMoreCaseInsensitiveMatches() {
Person a = Person.builder().id(UUID.randomUUID()).firstName("Hans").lastName("Müller").build();
Person b = Person.builder().id(UUID.randomUUID()).firstName("hans").lastName("müller").build();
when(personRepository.findByFirstNameAndLastName("hans", "müller")).thenReturn(Optional.empty());
when(personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase("hans", "müller"))
.thenReturn(List.of(a, b));
// Ambiguous sender → unset, never an arbitrary guess (provenance correctness over a
// confidently-wrong pre-fill). This is the deliberate divergence from the alias path.
assertThat(personService.findByName("hans", "müller")).isEmpty();
}
@Test
void findByName_returnsEmpty_whenFirstNameNullFoldsToNoMatch() {
when(personRepository.findByFirstNameAndLastName(null, "Müller")).thenReturn(Optional.empty());
when(personRepository.findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase(null, "Müller"))
.thenReturn(List.of());
assertThat(personService.findByName(null, "Müller")).isEmpty();
}
// ─── updatePerson (notes) ────────────────────────────────────────────────
@@ -700,10 +794,14 @@ class PersonServiceTest {
personService.mergePersons(sourceId, targetId);
verify(personRepository).findById(sourceId);
verify(personRepository).findById(targetId);
verify(personRepository).reassignSender(sourceId, targetId);
verify(personRepository).insertMissingReceiverReference(sourceId, targetId);
verify(personRepository).deleteReceiverReferences(sourceId);
verify(personRepository).deleteById(sourceId);
// The source's leftover receiver rows cascade-drop via V71's ON DELETE CASCADE on
// deleteById — merge no longer deletes them explicitly.
verifyNoMoreInteractions(personRepository);
}
// ─── getAliases ─────────────────────────────────────────────────────────
@@ -800,4 +898,165 @@ class PersonServiceTest {
.extracting(e -> ((DomainException) e).getStatus().value())
.isEqualTo(403);
}
@Test
void findByDisplayNameContaining_delegatesToSearchByName() {
Person walter = Person.builder().id(UUID.randomUUID()).firstName("Walter").lastName("Müller").build();
when(personRepository.searchByName("Walter")).thenReturn(List.of(walter));
List<Person> result = personService.findByDisplayNameContaining("Walter");
assertThat(result).containsExactly(walter);
verify(personRepository).searchByName("Walter");
}
// ─── tokenize (name-match contract) ───────────────────────────────────────
@Test
void tokenize_hyphenatedName_splitsOnHyphen() {
assertThat(PersonService.tokenize("Anna-Maria")).containsExactly("anna", "maria");
}
@Test
void tokenize_apostropheName_splitsOnApostrophe() {
assertThat(PersonService.tokenize("D'Angelo")).containsExactly("d", "angelo");
}
@Test
void tokenize_umlautName_lowercasesToSingleToken() {
assertThat(PersonService.tokenize("Müller")).containsExactly("müller");
}
@Test
void tokenize_doubleSpace_dropsEmptyTokens() {
assertThat(PersonService.tokenize("Clara Cram")).containsExactly("clara", "cram");
}
@Test
void tokenize_allWhitespace_returnsEmpty() {
assertThat(PersonService.tokenize(" ")).isEmpty();
}
@Test
void tokenize_null_returnsEmpty() {
assertThat(PersonService.tokenize(null)).isEmpty();
}
// ─── resolveByName (direct / partial classification) ──────────────────────
@Test
void resolveByName_singleDirectMatch_classifiesAsDirect() {
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cram").build();
when(personRepository.searchByName("clara")).thenReturn(List.of(clara));
when(personRepository.searchByName("cram")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.direct()).containsExactly(clara);
}
@Test
void resolveByName_maidenAliasToken_classifiesAsDirect() {
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Müller")
.nameAliases(List.of(PersonNameAlias.builder().lastName("Cram")
.type(PersonNameAliasType.MAIDEN_NAME).build()))
.build();
when(personRepository.searchByName("clara")).thenReturn(List.of(clara));
when(personRepository.searchByName("cram")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.direct()).containsExactly(clara);
}
@Test
void resolveByName_aliasFirstNameToken_isFetchedAndClassified() {
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cram")
.nameAliases(List.of(PersonNameAlias.builder().firstName("Wilhelmina").lastName("de Gruyter")
.type(PersonNameAliasType.BIRTH).build()))
.build();
when(personRepository.searchByName("wilhelmina")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Wilhelmina");
assertThat(result.direct()).containsExactly(clara);
}
@Test
void resolveByName_middleName_stillDirect() {
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara Maria").lastName("Cram").build();
when(personRepository.searchByName("clara")).thenReturn(List.of(clara));
when(personRepository.searchByName("cram")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.direct()).containsExactly(clara);
}
@Test
void resolveByName_reorderedTokens_stillDirect() {
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cram").build();
when(personRepository.searchByName("cram")).thenReturn(List.of(clara));
when(personRepository.searchByName("clara")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Cram Clara");
assertThat(result.direct()).containsExactly(clara);
}
@Test
void resolveByName_cramVsCramer_classifiesAsPartial() {
Person cramer = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cramer").build();
when(personRepository.searchByName("clara")).thenReturn(List.of(cramer));
when(personRepository.searchByName("cram")).thenReturn(List.of(cramer));
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.partial()).containsExactly(cramer);
}
@Test
void resolveByName_emptyAfterTokenizing_returnsNoCandidates() {
NameMatches result = personService.resolveByName(" - ");
assertThat(result.direct()).isEmpty();
verify(personRepository, never()).searchByName(any());
}
@Test
void resolveByName_directSortsBeyondCap_stillReturnedAsDirect() {
List<Person> pool = new java.util.ArrayList<>();
for (int i = 0; i < 10; i++) {
pool.add(Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cramer").build());
}
Person direct = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cram").build();
pool.add(direct);
when(personRepository.searchByName("clara")).thenReturn(pool);
when(personRepository.searchByName("cram")).thenReturn(pool);
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.direct()).containsExactly(direct);
}
@Test
void resolveByName_over8Tokens_issuesAtMost8Fetches() {
personService.resolveByName("a b c d e f g h i j");
verify(personRepository, org.mockito.Mockito.atMost(8)).searchByName(any());
}
@Test
void resolveByName_samePersonFromTwoTokens_appearsOnce() {
// Both token fetches return the same person id — fetchPool's putIfAbsent must dedup so the
// candidate is classified once, not twice.
Person clara = Person.builder().id(UUID.randomUUID()).firstName("Clara").lastName("Cram").build();
when(personRepository.searchByName("clara")).thenReturn(List.of(clara));
when(personRepository.searchByName("cram")).thenReturn(List.of(clara));
NameMatches result = personService.resolveByName("Clara Cram");
assertThat(result.direct()).hasSize(1);
assertThat(result.partial()).isEmpty();
}
}

View File

@@ -53,20 +53,68 @@ class TagServiceTest {
// ─── findOrCreate ─────────────────────────────────────────────────────────
@Test
void findOrCreate_returnsExisting_whenNameFound() {
Tag existing = Tag.builder().id(UUID.randomUUID()).name("Familie").build();
when(tagRepository.findByNameIgnoreCase("Familie")).thenReturn(Optional.of(existing));
void findOrCreate_exactCaseWins_overCaseInsensitiveSibling() {
// "Geburt" (parent) and "geburt" (child) both exist; the edit round-trip replays the stored
// name "geburt", which must bind to the exact-case row, not the parent.
Tag exact = Tag.builder().id(UUID.randomUUID()).name("geburt").build();
when(tagRepository.findByName("geburt")).thenReturn(Optional.of(exact));
Tag result = tagService.findOrCreate("Familie");
Tag result = tagService.findOrCreate("geburt");
assertThat(result).isEqualTo(existing);
assertThat(result).isEqualTo(exact);
verify(tagRepository, never()).save(any());
}
@Test
void findOrCreate_createsNew_whenNameNotFound() {
void findOrCreate_exactCaseWins_evenWhenItsIdIsNotTheLowest() {
// Adversarial guard: exact-case must short-circuit BEFORE the lowest-id rule. Here the exact row
// has the higher id, so a naive "always pick lowest id across all CI matches" would pick wrong.
Tag exactHigherId = Tag.builder().id(UUID.fromString("00000000-0000-0000-0000-000000000009")).name("geburt").build();
when(tagRepository.findByName("geburt")).thenReturn(Optional.of(exactHigherId));
Tag result = tagService.findOrCreate("geburt");
assertThat(result).isEqualTo(exactHigherId);
verify(tagRepository, never()).findAllByNameIgnoreCase(any()); // exact-case wins without consulting the CI list
verify(tagRepository, never()).save(any());
}
@Test
void findOrCreate_usesSingleCaseInsensitiveMatch_whenNoExactCase() {
// Stored name is "Weihnachten"; a save replays "weihnachten" (no exact-case row) → bind to the
// single case-insensitive match rather than creating a duplicate.
Tag stored = Tag.builder().id(UUID.randomUUID()).name("Weihnachten").build();
when(tagRepository.findByName("weihnachten")).thenReturn(Optional.empty());
when(tagRepository.findAllByNameIgnoreCase("weihnachten")).thenReturn(List.of(stored));
Tag result = tagService.findOrCreate("weihnachten");
assertThat(result).isEqualTo(stored);
verify(tagRepository, never()).save(any());
}
@Test
void findOrCreate_returnsLowestIdDeterministically_whenMultipleCaseInsensitiveMatches() {
// Two rows collide case-insensitively and neither equals the query exactly. Resolution must be
// deterministic (lowest id) and never throw — proven by calling twice and getting the same id.
Tag lowerId = Tag.builder().id(UUID.fromString("00000000-0000-0000-0000-000000000001")).name("Reisepläne").build();
Tag higherId = Tag.builder().id(UUID.fromString("00000000-0000-0000-0000-000000000002")).name("reisepläne").build();
when(tagRepository.findByName("REISEPLÄNE")).thenReturn(Optional.empty());
when(tagRepository.findAllByNameIgnoreCase("REISEPLÄNE")).thenReturn(List.of(higherId, lowerId));
Tag first = tagService.findOrCreate("REISEPLÄNE");
Tag second = tagService.findOrCreate("REISEPLÄNE");
assertThat(first.getId()).isEqualTo(lowerId.getId());
assertThat(second.getId()).isEqualTo(first.getId());
verify(tagRepository, never()).save(any());
}
@Test
void findOrCreate_createsOrphanTag_whenNameAbsent() {
Tag saved = Tag.builder().id(UUID.randomUUID()).name("Krieg").build();
when(tagRepository.findByNameIgnoreCase("Krieg")).thenReturn(Optional.empty());
when(tagRepository.findByName("Krieg")).thenReturn(Optional.empty());
when(tagRepository.findAllByNameIgnoreCase("Krieg")).thenReturn(List.of());
when(tagRepository.save(any())).thenReturn(saved);
Tag result = tagService.findOrCreate("Krieg");
@@ -76,13 +124,15 @@ class TagServiceTest {
}
@Test
void findOrCreate_trimsWhitespaceBeforeLookup() {
Tag existing = Tag.builder().id(UUID.randomUUID()).name("Urlaub").build();
when(tagRepository.findByNameIgnoreCase("Urlaub")).thenReturn(Optional.of(existing));
void findOrCreate_trimsWhitespace_thenLandsOnCaseInsensitiveChild() {
Tag child = Tag.builder().id(UUID.randomUUID()).name("weihnachten").build();
when(tagRepository.findByName("weihnachten")).thenReturn(Optional.empty());
when(tagRepository.findAllByNameIgnoreCase("weihnachten")).thenReturn(List.of(child));
tagService.findOrCreate(" Urlaub ");
Tag result = tagService.findOrCreate(" weihnachten ");
verify(tagRepository).findByNameIgnoreCase("Urlaub");
assertThat(result).isEqualTo(child);
verify(tagRepository).findByName("weihnachten");
}
// ─── update ───────────────────────────────────────────────────────────────
@@ -616,4 +666,17 @@ class TagServiceTest {
// verify findAllById was called at least twice: once for extras, once inside resolveEffectiveColors
verify(tagRepository, atLeastOnce()).findAllById(any());
}
// ─── findByNameContaining ─────────────────────────────────────────────────
@Test
void findByNameContaining_delegatesToRepository() {
Tag krieg = Tag.builder().id(UUID.randomUUID()).name("Krieg").build();
when(tagRepository.findByNameContainingIgnoreCase("krieg")).thenReturn(List.of(krieg));
List<Tag> result = tagService.findByNameContaining("krieg");
assertThat(result).containsExactly(krieg);
verify(tagRepository).findByNameContainingIgnoreCase("krieg");
}
}

View File

@@ -132,6 +132,31 @@ class AdminControllerTest {
.andExpect(jsonPath("$.count").value(3));
}
// ─── POST /api/admin/backfill-titles (#726) ────────────────────────────────
@Test
void backfillTitles_returns401_whenUnauthenticated() throws Exception {
mockMvc.perform(post("/api/admin/backfill-titles").with(csrf()))
.andExpect(status().isUnauthorized());
}
@Test
@WithMockUser(roles = "USER")
void backfillTitles_returns403_whenNotAdmin() throws Exception {
mockMvc.perform(post("/api/admin/backfill-titles").with(csrf()))
.andExpect(status().isForbidden());
}
@Test
@WithMockUser(authorities = "ADMIN")
void backfillTitles_returns200_withCount_whenAdmin() throws Exception {
when(documentService.backfillTitles()).thenReturn(7);
mockMvc.perform(post("/api/admin/backfill-titles").with(csrf()))
.andExpect(status().isOk())
.andExpect(jsonPath("$.count").value(7));
}
// ─── POST /api/admin/generate-thumbnails ───────────────────────────────────
@Test

View File

@@ -52,11 +52,12 @@ The OCR service requires significant RAM for model loading. The dev compose sets
| Production target | RAM | Recommended OCR limit | Notes |
|---|---|---|---|
| Hetzner CX42 | 16 GB | 12 GB | Recommended for OCR-enabled production |
| Hetzner CX32 | 8 GB | 6 GB | Accept reduced batch sizes and slower throughput |
| Hetzner CX22 | 4 GB | — | Disable the OCR service (`profiles: [ocr]`); run OCR on demand only |
| Current server (Hetzner Serverbörse, i7-6700) | 64 GB | 12 GB | Default `mem_limit: 12g` works comfortably |
| ≥ 16 GB RAM | 16+ GB | 12 GB | Default works |
| 8 GB RAM | 8 GB | 6 GB | Set `OCR_MEM_LIMIT=6g`; accept reduced batch sizes |
| 4 GB RAM | 4 GB | — | Disable OCR service (`profiles: [ocr]`); run OCR on demand only |
A CX32 cannot honour the default `mem_limit: 12g` — set the `OCR_MEM_LIMIT=6g` env var (in `.env.production` / `.env.staging`, or as a Gitea secret consumed by the workflow) before deploying on a CX32. The prod compose interpolates this var with a 12g default.
On servers with less than 16 GB RAM the default `mem_limit: 12g` cannot be honoured — set the `OCR_MEM_LIMIT` env var (in `.env.production` / `.env.staging`, or as a Gitea secret consumed by the workflow). The prod compose interpolates this var with a 12g default.
### Dev vs production differences
@@ -140,7 +141,7 @@ All vars are set in `.env` at the repo root (copy from `.env.example`). The back
| `ALLOWED_PDF_HOSTS` | SSRF protection — comma-separated list of allowed PDF source hosts. **Do not widen to `*`** | `minio,localhost,127.0.0.1` | YES | — |
| `KRAKEN_MODEL_PATH` | Directory containing Kraken HTR models (populated by `download-kraken-models.sh`) | `/app/models/` | — | — |
| `BLLA_MODEL_PATH` | Kraken baseline layout analysis model path | `/app/models/blla.mlmodel` | — | — |
| `OCR_MEM_LIMIT` | Container memory cap for ocr-service in `docker-compose.prod.yml`. Set to `6g` on CX32 hosts; leave unset on CX42+ to use the 12g default | `12g` (prod compose default) | — | — |
| `OCR_MEM_LIMIT` | Container memory cap for ocr-service in `docker-compose.prod.yml`. Set to `6g` on servers with 8 GB RAM; leave unset (12g default) on servers with ≥ 16 GB RAM | `12g` (prod compose default) | — | — |
| `XDG_CACHE_HOME` | XDG cache base dir — redirects Matplotlib and other XDG-aware libraries away from the read-only `HOME` (`/home/ocr`) to the writable cache volume | `/app/cache` | — | — |
| `TORCH_HOME` | PyTorch model cache — redirects `~/.cache/torch` to the writable models volume | `/app/models/torch` | — | — |
@@ -264,6 +265,7 @@ git.raddatz.cloud A <server IP>
### 3.4 First deploy
```bash
# 1. Trigger nightly.yml manually (Repo → Actions → nightly → "Run workflow")
# Expected: docker compose up -d --wait succeeds for archiv-staging, then

View File

@@ -45,6 +45,9 @@ _See also [TranscriptionBlock](#transcriptionblock-transcriptionblock)._
**raw attribution** (`Document.senderText`, `Document.receiverText`, `Document.metaDateRaw`) — the original spreadsheet cell text for a document's sender, receiver, and date, preserved verbatim even after a `Person` or normalized date is linked. It keeps provenance intact and enables an "as written in the original" view.
**auto-generated title** (`DocumentTitleFactory`) — a `Document` title composed by the formula `{index} {dateLabel} {location}` (index = `originalFilename`; date label honest at the row's precision; location omitted when blank). On edit, an unchanged auto-title follows a corrected date/location forward (exact old-vs-new match in `DocumentService.updateDocument`); a hand-written title is kept verbatim. `POST /api/admin/backfill-titles` rewrites already-stale ones in one sweep using a grammar heuristic (`DocumentTitleBackfillMatcher`).
_Not to be confused with a hand-written title_ — only a title that still equals what the factory builds is treated as machine-generated and rewritten; prose is left untouched.
**DocumentVersion** (`DocumentVersion`) — an append-only snapshot of a `Document`'s metadata at a point in time. Append-only by convention; no consumer-facing create or update endpoint exists. The entity uses Lombok `@Data` (which generates setters), so immutability is enforced by application convention, not at the Java level.
**Tag** (`Tag`) — a hierarchical category that can be applied to `Document`s. Tags are self-referencing via a `parent_id` foreign key, forming a tree structure.
@@ -162,6 +165,8 @@ _See also [Chronik](#chronik-internal)._
**Domain** — a Tier-1 bounded context with its own entities, controller, service, repository, and DTOs. Backend domains: `document`, `person`, `tag`, `user`, `geschichte`, `notification`, `ocr`, `audit`, `dashboard`. Frontend domains mirror this structure under `src/lib/`.
**NameMatches** — the Person-domain result of `PersonService.resolveByName(name)`: candidate persons split by name-match strength into `direct` and `partial`. A match is **direct** when every query token is a whole-token match (order-independent, alias/maiden-name aware) across all of a person's name components (`firstName`, `lastName`, `alias`, each `PersonNameAlias` first+last, `title`); a **partial** matched the substring fetch but is not direct (e.g. "Cram" → "Clara Cramer").
---
## Infrastructure Terms

View File

@@ -35,7 +35,7 @@ Render thumbnails in-process in Spring Boot using **Apache PDFBox 3.0.4** (alrea
**Harder:**
- PDFBox is a parser attack surface. Mitigated by a 30-second watchdog timeout in `ThumbnailAsyncRunner` and by the fire-and-forget contract (failures never break upload).
- Memory ceiling: the `thumbnailExecutor` is capped at 2 threads on the CX32 (8 GB). A busy backfill alongside OCR can approach the 3 GB heap — acceptable but not comfortable. Streaming via `FileService.downloadFileStream` keeps this bounded for PDFs up to 50 MB.
- Memory ceiling: the `thumbnailExecutor` is capped at 2 threads on memory-constrained hosts. A busy backfill alongside OCR can approach the 3 GB heap on an 8 GB server — acceptable but not comfortable. The current production server (64 GB) has ample headroom. Streaming via `FileService.downloadFileStream` keeps this bounded for PDFs up to 50 MB.
### Operational caveats (intentional)

View File

@@ -62,7 +62,7 @@ The `/tmp` tmpfs remains at 512 MB and continues to serve training-ZIP extractio
## Alternatives considered
**Approach B — Enlarge `/tmp` to 4 GB**
One-line change. Discarded because: (1) 4 GB tmpfs counts against the cgroup `mem_limit`; on CX32 hosts with `OCR_MEM_LIMIT=6g` the combined Surya resident set + tmpfs would trigger OOMKill on cold start; (2) staging GB-scale model files through RAM is using the wrong storage tier; (3) any future model larger than 4 GB requires another bump.
One-line change. Discarded because: (1) 4 GB tmpfs counts against the cgroup `mem_limit`; on servers with `OCR_MEM_LIMIT=6g` the combined Surya resident set + tmpfs would trigger OOMKill on cold start; (2) staging GB-scale model files through RAM is using the wrong storage tier; (3) any future model larger than 4 GB requires another bump.
**Approach C — Both TMPDIR redirect and enlarged /tmp**
Belt-and-suspenders: Approach A + 1 GB tmpfs. Discarded in favour of the cleaner Approach A. The defence-in-depth benefit does not outweigh the extra compose churn; the 512 MB cap on `/tmp` is intentional.

View File

@@ -0,0 +1,112 @@
# ADR-031 — The document title is a shared `document`-package factory, re-synced by an exact match on save and a grammar heuristic on a one-time backfill
**Date:** 2026-06-04
**Status:** Accepted
**Issue:** #726 (auto-sync document titles with date/location: save-time + one-time backfill)
**Milestone:**
---
## Context
A document title was a string built **once**, at import time, by a private
`DocumentImporter.buildTitle()` composing `{index} {dateLabel} {location}` (index =
`originalFilename`, date label honest at the row's precision via `DocumentTitleFormatter`,
location verbatim). Nothing rebuilt it afterwards. When an archivist later corrected a date
or location in the edit form, the title kept its stale value (e.g. it still read `2028`
after the date was fixed to `1928`), because the edit form round-trips the stored title
verbatim and `updateDocument` simply re-persisted it.
Two distinct problems live here:
1. **Going forward**, an edit to date/location must flow into a title that was machine-built
— but must never overwrite a title a human wrote.
2. **The existing backlog** of already-stale titles must be cleaned once. For these rows the
pre-edit state is gone, so there is no exact value to compare against.
The composition formula also existed only inside `importing`, which is the wrong owner: a
title is a `document` concern, and three call sites (import, save-time, backfill) must share
one rule or they will drift.
## Decision
### 1. One formula, owned by the `document` package (`DocumentTitleFactory`)
Extract the composition into `DocumentTitleFactory` (a `@Component` in the `document`
package) with `build(Document)`. `DocumentImporter` (package `importing`) now consumes it.
`DocumentTitleFormatter` moves into `document` alongside the factory (it stays
package-private; `importing` reaches the formula only through the factory). The direction is
deliberate: `document` owns the rule, `importing` depends on it — not the reverse. The
German date *label* remains the deliberate Java/TS dual implementation pinned by
`docs/date-label-fixtures.json` (#666); this ADR touches the **composition** only and does
not collapse the frontend `formatDocumentDate`.
### 2. Save-time regeneration is an EXACT match, not a heuristic
In `DocumentService.updateDocument` only (bulk edit is out of scope), capture
`autoTitleBefore = titleFactory.build(doc)` from the **currently-persisted** state *before*
any setter runs. Then:
- if the **submitted** title equals `autoTitleBefore`, it was the machine value → rebuild
from the new state;
- otherwise keep the submitted title verbatim (hand-written or freshly typed).
This is an exact old-vs-new comparison — no false positives, no false negatives — relying on
the edit form round-tripping an untouched title verbatim. `projectedState` mirrors the
existing setter asymmetry exactly: `documentDate`/`location` overwrite unconditionally (a
null clears them), while precision/end/raw are taken from the DTO only when non-null and
otherwise kept from the entity. A blank submission is never persisted (the title is always
present) — it falls back to the rebuilt auto-title, which always carries at least the index.
### 3. The one-time backlog cleanup is a grammar heuristic, behind an ADMIN endpoint
`POST /api/admin/backfill-titles` (synchronous, under `AdminController`'s class-level
`@RequirePermission(Permission.ADMIN)`) sweeps every document and, for each whose stored
title passes the overwrite test, rebuilds it via the factory. Because the pre-edit state is
gone, the test (`DocumentTitleBackfillMatcher`, used **only** here) is a grammar heuristic:
after stripping the **literal** index prefix, the remainder must be exactly the index, a
known date-label form (+ an optional trailing location), or a lone segment equal to the
document's current location. Prose is left untouched; anything malformed fails closed.
The backfill saves via `documentRepository.save` directly and **never** routes through
`updateDocument` — following the `backfillFileHashes` precedent — so a mechanical rename does
not snapshot the whole corpus into `document_versions`. It is idempotent (a second run
rewrites nothing) and logs one SLF4J-parameterized `scanned/updated/skipped` line; the
response is `BackfillResult(count)`.
### 4. Edit-form feedback (FR-005)
A localized helper line (de/en/es) under the title input explains that the title is built
from date/place and that a hand-edit is preserved, wired via `aria-describedby` and shown
only on the single-document edit form. A live preview was considered and declined.
## Consequences
- The three call sites can never diverge — there is exactly one formula
(`NFR-MAINT-001`). Save-time cost is a string build + compare; the backfill is one
synchronous transactional sweep over a low-thousands corpus.
- Security: the index is compared **literally** (`String.startsWith` / `Pattern.quote`)
because `originalFilename` is user-controlled and may carry regex metacharacters — an
unquoted pattern would be a ReDoS / regex-injection vector (CWE-1333 / CWE-625). The
date-label sub-patterns use only bounded, non-nested quantifiers.
- **File-replaced documents are treated as manual, by design.** The index is
`originalFilename`, which `updateDocument` reassigns to the uploaded file's name on a
file-replace. After a replace the stored title no longer matches `build(currentState)`, so
neither save-time nor backfill rewrites it. This is the accepted fail-safe of overloading
`originalFilename` rather than adding a dedicated `catalogIndex` column.
- The save-time heuristic risk is zero (exact match); the backfill heuristic can, by its
documented FR-004 rule, treat `{index} {valid date label} {anything}` as machine-built
and rewrite the trailing segment. This is the accepted trade for cleaning the backlog
without the lost pre-edit state.
## Alternatives considered
- **A dedicated `catalogIndex` column** instead of overloading `originalFilename` — rejected;
it adds a migration and a second source of truth for the index for no current benefit, and
the file-replace fail-safe is acceptable.
- **A heuristic at save-time too** (instead of the exact match) — rejected; the stored title
is available pre-edit, so an exact comparison is strictly better (no false positives).
- **A live title preview in the edit form** — rejected (FR-005); a static helper line is
calmer for the 60+ audience and avoids a second client-side mirror of the formula.
- **Collapsing the frontend `formatDocumentDate` into the backend** — out of scope; the
Java/TS date-label split is the deliberate #666 design, pinned by a shared fixture.

View File

@@ -0,0 +1,64 @@
# ADR-032 — Person-delete referential integrity lives in the database, and the cascade never reaches `documents`
**Date:** 2026-06-06
**Status:** Accepted
**Issue:** #684 (move person-delete FK detach to database-level `ON DELETE`)
**Milestone:**
---
## Context
Deleting a `Person` had to detach the two FKs into `persons` that lacked any `ON DELETE`
behaviour: `documents.sender_id` and `document_receivers.person_id` (both from V1).
`PersonService.deletePerson` and `mergePersons` did this in Java — nulling the sender and
deleting receiver join rows before `deleteById` — so the integrity guarantee lived in
application code. Any other delete path (a future endpoint, a manual `psql`, a batch job)
could still orphan rows or fail with an FK-violation 500.
A related soft reference made it worse: `transcription_block_mentioned_persons.person_id`
was a UUID column with **no FK** (V56, a deliberate "no FK" choice), so a person delete left
dangling `@`-mention rows. The literal `@DisplayName` lives in `transcription_blocks.text`,
so only the *link* was ever at stake — not the visible name.
## Decision
Move person-delete integrity into the database (migration V71) and thin the service to a
plain `deleteById`:
- `documents.sender_id``ON DELETE SET NULL` (`documents.senderText` preserves the raw
textual attribution, so nulling the link loses no historical record).
- `document_receivers.person_id``ON DELETE CASCADE` (the symmetric completion of V14,
which gave the `document_id` side the same).
- `transcription_block_mentioned_persons.person_id` → a real FK with `ON DELETE CASCADE`,
reversing V56's "no FK" decision. The read renderer already degrades a `@DisplayName` with
no sidecar row to plain escaped text, so removing the link is invisible to the reader.
**Cascade-boundary invariant:** the cascade stays strictly at the join/reference layer and
**never reaches `documents` rows** — a cascade into `documents` would destroy historical
letters. This is pinned by a non-negotiable document-survival assertion in
`PersonRepositoryTest`.
## Consequences
- A person delete is safe from every path, not just `PersonService`. The service and merge
stay thin (`deleteById` + the cascade); `reassignSenderToNull` and `deleteReceiverReferences`
are deleted.
- This *fixes* the pre-existing dead-link-on-deleted-person case — it is not a purely
invisible refactor. Note the read renderer strips the `@` prefix when it emits a live
mention link, but the degraded (deleted-person) path leaves the literal `@Name` in the
block text as-is — the reader sees `@Auguste Raddatz` as plain text, never a dead link.
- DB cascades run below `AuditService`, so the row-level cleanup is intentionally not
audit-logged; the person-delete action itself is still logged at the service layer.
- The V71 FK validation requires cleaning pre-existing orphan mention rows first; the
migration does this in a `DO` block that logs the purge count via `RAISE NOTICE`.
## Alternatives considered
- **Keep integrity in Java** — rejected; it only protects the one code path and re-breaks the
moment a second delete path appears.
- **Cascade `documents.sender_id`** — rejected; it would delete historical letters when a
sender is removed. `SET NULL` keeps the letter and its `senderText`.
- **Leave the mention sidecar FK-less (honour V56)** — rejected; the "no FK" rationale was
stale, the name survives in the block text regardless, and the FK removes the orphan-row
class of bug.

View File

@@ -0,0 +1,148 @@
# ADR-033 — Tag-name resolution tolerates case-collisions: exact-case first, then a deterministic lowest-id fallback, and never a `unique(lower(name))` constraint
**Date:** 2026-06-06
**Status:** Accepted
**Issue:** #730 (document with a case-colliding tag cannot be saved — `findByNameIgnoreCase` `NonUniqueResultException`)
**Milestone:**
---
## Context
`TagService.findOrCreate(name)` is the single point that turns a tag **name** into a `Tag`
row. The document edit form, bulk-edit, and the upload batch all round-trip tag **names**
(the edit form sends `tags.map(t => t.name).join(',')`) and re-resolve them on **every**
save through `resolveTags → findOrCreate`. The old implementation resolved with
`tagRepository.findByNameIgnoreCase(name)`, a derived query returning `Optional<Tag>`.
That signature encodes an invariant the data does **not** hold: that a name is globally
unique case-insensitively. The canonical tag tree legitimately contains names that differ
only by case — a parent container and its same-named lowercase **child** (`Geburt` /
`Geburt/geburt`, `Weihnachten` / `Weihnachten/weihnachten`, …), or two siblings
(`Reise/Reisepläne` / `Reise/reisepläne`). Each is a distinct node with its own
`source_ref` (the stable identity, per ADR-025) and its own document attachments — **not** an
accidental duplicate. When two rows matched case-insensitively, Hibernate threw
`NonUniqueResultException``IncorrectResultSizeDataAccessException` → a generic HTTP 500.
The effect was severe and opaque: every document carrying one of ~10 colliding tags (≈180
document-tag attachments on staging) became **un-editable** — any field change failed on save
because the whole tag set is re-resolved — and the user saw only "an unexpected error", with
no hint that a tag was the cause.
This is a **lookup** problem, not a data problem: the collisions are valid canonical nodes
and must be preserved.
## Decision
### 1. Resolution is exact-case first, then a non-throwing deterministic fallback
`findOrCreate` resolves in three ordered steps and never throws on a collision:
1. `findByName(cleanName)`**exact-case** derived query. If present, return it. The edit
round-trip replays the stored name verbatim, so the exact-case row is the right binding
(typing the bare child name `weihnachten` binds to the child; `Weihnachten` binds to the
parent container).
2. else `findAllByNameIgnoreCase(cleanName)` — the **plural** case-insensitive list. If
non-empty, return the element with the **lowest `id`** (`min(comparing(Tag::getId))`).
3. else create the tag (an orphan: null `sourceRef`/`parentId`).
The two repository methods are deliberately **two distinctly-named methods** — Spring Data
cannot disambiguate an `Optional<Tag>` from a `List<Tag>` derived query by return type alone.
The throwing `Optional<Tag> findByNameIgnoreCase` is **deleted** so the non-unique-throwing
call cannot be reintroduced; `findOrCreate` was its only production caller.
### 2. The tie-break is `id`, and it is load-bearing
`id` is a stable, always-present, unique column, so "lowest id" is a total, deterministic
order over the candidates: the same name resolves to the same row on every call, forever,
without throwing. This matters only in the free-text authoring path (step 2), where no
exact-case row exists yet two case-folding rows do.
### 3. No `unique(lower(name))` constraint — and a load-bearing comment says so
A global case-insensitive uniqueness constraint is **wrong**: it would reject the legitimate
parent/child canonical nodes. It would also **fail to apply** against the existing rows,
turning a code-only deploy into a failed Flyway migration that blocks startup. A comment at
both `findOrCreate` and the repository methods records this so the constraint is not "helpfully"
added later.
## Consequences
- **Code-only, zero migration, fully reversible** (roll back the JAR). No tag data is touched;
the colliding rows stay exactly as the canonical importer produced them.
- One change fixes all three write paths — single-document edit, bulk-edit, and upload batch —
because they all funnel through `resolveTags → findOrCreate`, which stays the single source
of truth (resolution logic is **not** hoisted into `DocumentService`).
- **Free-text tag semantics under collision are accepted as-is** (issue #730, option A): the
bare word `weihnachten` binds to the deep child and `Weihnachten` to the parent container.
Correct for the edit round-trip and acceptable for authoring; making the typeahead show the
tree path so an author can tell a container from its same-named child is a separate
follow-up.
- The wire response stays opaque: after the fix this path no longer throws
`IncorrectResultSizeDataAccessException`, and `GlobalExceptionHandler`'s generic handler maps
any stray one to `INTERNAL_ERROR` with no Hibernate/SQL leak — so no dedicated handler was
added.
- **The sibling Person path is fixed the same way — see the Person extension below (#731).**
- Postgres `LOWER()` folding of umlauts (`ü`/`ä`) is the actual correctness hinge of the
fallback and cannot be proven by a mocked repo, so it is pinned by a Testcontainers
`postgres:16-alpine` test on a `Glückwünsche`/`glückwünsche` pair; a plain-ASCII test would
stay green while the bug reappeared for umlaut tags.
## Person extension (#731)
The Person domain carried the same latent throw on **two** user-influenced lookup surfaces, and
is fixed with the same exact-case-first, non-throwing pattern — but with a deliberately
**different fallback per surface**, because the two paths have different consequences.
- **Alias path — `PersonService.findOrCreateByAlias` — deterministic lowest-id (mirrors tag).**
`findByAliasIgnoreCase` (`Optional`) is replaced by `findByAlias` (exact) → `findAllByAliasIgnoreCase`
(plural, lowest id) → the existing create-when-absent branch (INSTITUTION/GROUP and the
maiden-name alias are preserved verbatim). There is no human in the importer loop and the path
creates-on-absent anyway, so a deterministic guess is the right behaviour — exactly like tags.
- **Name/sender path — `PersonService.findByName` — bail to null on ambiguity (the new wrinkle).**
Used only by `DocumentService.storeDocument` to resolve the upload **sender** from the parsed
filename. `findByFirstNameIgnoreCaseAndLastNameIgnoreCase` (`Optional`) is replaced by
`findByFirstNameAndLastName` (exact) → `findAllByFirstNameIgnoreCaseAndLastNameIgnoreCase`
(plural). Resolution returns the exact-case match, else the single case-insensitive match, else
— on **two or more** matches — **empty**. The sender is left unset rather than guessing.
**Why this diverges from the alias (and tag) decision:** the archive's value is correct
provenance. A confidently-wrong pre-filled `Hans Müller` is worse than an empty field, because a
senior reviewer will not re-check a value that is already filled in, whereas an empty sender
routes the document into the "needs completion" state (`metadataComplete=false`) for a human to
assign. The load-bearing comment at `findByName` records this so a future "consistency cleanup"
does not reintroduce the confidently-wrong-sender bug by switching it to lowest-id.
- **Fail-closed on a null first name.** A parsed filename can lack a first name. The two new name
methods use explicit HQL equality (`= :firstName`) rather than a derived
`…IgnoreCase` query, because Spring Data folds a null derived-query argument to `first_name IS
NULL` — which would silently widen the match and pull a last-name-only / institution row in as a
"sender" (a quiet provenance-integrity defect). With HQL equality a null binds as `= NULL`,
which never matches, so a null first name resolves to **no sender**. This is pinned by a
real-Postgres repository test.
- **Scope — "ambiguous" is case-insensitive only.** Both exact-case lookups (`findByAlias`,
`findByFirstNameAndLastName`) return `Optional`, so two **byte-identical same-case** rows would
still throw `NonUniqueResultException`. That is a true data anomaly, deliberately out of scope
(it is not a case-collision), and it surfaces as the opaque `INTERNAL_ERROR` — never a silently
wrong row — so it is no worse than any other unexpected error and needs no extra handling here.
- **Same stance as tags otherwise:** no `unique(lower(alias))` / `unique(lower(name))` constraint
(collisions are valid human labels; `source_ref` is the stable identity per ADR-025), no
merge/dedupe, code-only and reversible, and no shared `resolveExactThenCi(...)` helper — the
two Person paths have different fallbacks, so the exact→CI→fallback logic is inlined at each
with its load-bearing comment (KISS).
## Alternatives considered
- **A `unique(lower(name))` index** — rejected: the collisions are valid canonical nodes, and
the migration would fail against the existing data and block startup.
- **Merging/deduping the colliding tags** — rejected: each has a distinct `source_ref`, tree
position, and real document attachments; they are not duplicates.
- **Round-tripping tag IDs instead of names** so resolution can't be ambiguous at all — the
cleaner long-term shape (removes the name-as-key smell), but a larger change with frontend
surface; deferred to #732. The lookup fix here is the minimal correct unblock.
- **Hoisting resolution into `DocumentService.resolveTags`** — rejected: it would duplicate the
rule across the edit, bulk-edit, and import paths and let them drift; `findOrCreate` stays
the one owner.

View File

@@ -0,0 +1,53 @@
# ADR-034 — Remove NL/smart-search (supersedes ADR-028 ×2, ADR-034-ollama, ADR-035)
**Date:** 2026-06-07
**Status:** Accepted
**Issue:** #772
**Supersedes:** ADR-028 (nl-search-ollama), ADR-028 (ollama-docker-compose-service), ADR-034 (ollama-production-deployment-and-keep-alive), ADR-035 (rule-based-nlp-service)
---
## Context
The natural-language search feature ("KI-Suche" / smart search) allowed users to enter
free-form queries like *"Was hat Walter an Emma im Krieg geschrieben?"* and have them
interpreted by an LLM into structured filters (persons, tags, date range, keywords).
The feature went through two major iterations:
1. **Ollama integration** (ADR-028): an `ollama` Docker service running a local LLM
(llama3.2/gemma3) parsed queries via a JSON-mode prompt.
2. **Rule-based NLP service** (ADR-035): after Ollama proved too slow and unreliable on
CPU-only hardware, a Python FastAPI microservice (`nlp-service`, port 8001) replaced
it with deterministic regex + spaCy parsing plus a lightweight LLM call.
Both approaches shared the same fundamental problem: inference on the production server
(Hetzner Serverbörse, no GPU, 64 GB RAM, i7-6700) was too slow to be useful, with
typical query latencies of 1030 seconds. Users got better and faster results from
the existing keyword search with date/person/tag filters.
## Decision
**Remove the NL search feature entirely.** The Python `nlp-service` microservice, the
Spring Boot `search/` package (`NlSearchController`, `NlQueryParserService`,
`RestClientNlpClient`, `NlSearchRateLimiter`, and all supporting classes), the frontend
NL search components (`SmartModeToggle`, `SmartSearchStatus`, `InterpretationChipRow`,
`DisambiguationPicker`), the related Docker Compose services, Prometheus scrape job,
Grafana dashboard, and all i18n keys are removed.
The existing structured search (FTS keyword + person/tag/date/directional filters) is
sufficient for the archive's current audience and search workload.
## Consequences
- **Capability removed:** users can no longer enter free-form natural-language queries.
They must use the structured filter bar (keyword text box + person/tag/date/directional
dropdowns). For documents where these filters are sufficient, there is no regression.
- **Operational simplification:** the Docker Compose stack loses two services
(`nlp-service` and previously `ollama`/`ollama-model-init`). Memory budget on the
production host is freed. No external model weights need to be kept warm.
- **Future reinstatement:** if a GPU-capable host becomes available, re-implementing
server-side LLM inference would be straightforward given the clean separation of the
`NlSearchController` entry point. However, this ADR deliberately avoids leaving dead
infrastructure or stub code in place — start clean if and when that becomes viable.
- **No data or schema change:** only query/endpoint code and Docker services are removed.
The `documents`, `persons`, and `tags` tables and their FTS indexes are untouched.

View File

@@ -9,10 +9,12 @@ Person(member, "Family Member", "Access by administrator invite. Searches, brows
System(familienarchiv, "Familienarchiv", "Web application for digitising, organising, and searching family documents")
System_Ext(mail, "Email Service", "SMTP server. Delivers notification emails (mentions, replies) and password-reset links.")
System_Ext(glitchtip, "GlitchTip", "Self-hosted error tracking (Sentry-compatible). Receives frontend and backend error events with stack traces.")
System_Ext(ollama, "Ollama (self-hosted)", "Local LLM inference server (qwen2.5:7b). Parses natural-language search queries into structured filters. Runs in the same Docker Compose stack.")
Rel(admin, familienarchiv, "Manages via browser", "HTTPS")
Rel(member, familienarchiv, "Searches, reads, and transcribes via browser", "HTTPS")
Rel(familienarchiv, mail, "Sends notification and password-reset emails (optional)", "SMTP")
Rel(familienarchiv, glitchtip, "Sends error events with errorId and stack trace", "HTTPS")
Rel(familienarchiv, ollama, "NL query parsing for natural-language search", "HTTP / REST (internal)")
@enduml

View File

@@ -18,7 +18,7 @@ System_Boundary(archiv, "Familienarchiv (Docker Compose)") {
}
System_Boundary(observability, "Observability Stack (/opt/familienarchiv/docker-compose.observability.yml)") {
Container(prometheus, "Prometheus", "prom/prometheus:v3.4.0", "Scrapes metrics from backend management port 8081 (/actuator/prometheus), node-exporter, and cAdvisor. Retention: 30 days.")
Container(prometheus, "Prometheus", "prom/prometheus:v3.4.0", "Scrapes metrics from backend (8081 /actuator/prometheus), OCR service (8000 /metrics), node-exporter, and cAdvisor. Retention: 30 days.")
Container(node_exporter, "Node Exporter", "prom/node-exporter:v1.9.0", "Host-level CPU, memory, disk, and network metrics.")
Container(cadvisor, "cAdvisor", "gcr.io/cadvisor/cadvisor:v0.52.1", "Per-container resource metrics.")
Container(loki, "Loki", "grafana/loki:3.4.2", "Stores log streams from all containers.")
@@ -45,6 +45,7 @@ Rel(promtail, loki, "Pushes log streams", "HTTP/Loki push API")
Rel(backend, tempo, "Sends distributed traces via OTLP", "HTTP / OTLP / port 4318 (archiv-net)")
Rel(prometheus, backend, "Scrapes JVM + HTTP metrics", "HTTP 8081 /actuator/prometheus")
Rel(prometheus, ocr, "Scrapes OCR + http_* metrics", "HTTP 8000 /metrics")
Rel(grafana, prometheus, "Queries metrics", "HTTP 9090")
Rel(grafana, loki, "Queries logs", "HTTP 3100")
Rel(grafana, tempo, "Queries traces", "HTTP 3200")

View File

@@ -9,15 +9,17 @@ ContainerDb(minio, "Object Storage", "MinIO (S3-compatible)")
System_Boundary(backend, "API Backend (Spring Boot)") {
Component(docCtrl, "DocumentController", "Spring MVC — /api/documents", "CRUD for documents: search, get by ID, update metadata, upload/download file, batch metadata updates, and per-month density aggregation for the timeline filter widget.")
Component(adminCtrl, "AdminController", "Spring MVC — /api/admin", "Triggers the asynchronous canonical import (requires ADMIN permission). Reports import state (IDLE/RUNNING/DONE/FAILED).")
Component(docSvc, "DocumentService", "Spring Service", "Core document business logic: store, update, search. Resolves persons and tags, delegates file I/O to FileService, builds dynamic JPA Specifications, and integrates with audit logging.")
Component(adminCtrl, "AdminController", "Spring MVC — /api/admin", "Triggers the asynchronous canonical import (requires ADMIN permission). Reports import state (IDLE/RUNNING/DONE/FAILED). Hosts the one-shot maintenance backfills (versions, file-hashes, titles) — synchronous, ADMIN-only.")
Component(docSvc, "DocumentService", "Spring Service", "Core document business logic: store, update, search. On update, regenerates an unchanged auto-title from the new date/location (exact old-vs-new match, #726); exposes backfillTitles() to clean already-stale titles in one sweep. Resolves persons and tags, delegates file I/O to FileService, builds dynamic JPA Specifications, and integrates with audit logging.")
Component(fileSvc, "FileService", "Spring Service", "Wraps AWS SDK v2 S3Client. Uploads files with UUID-keyed paths, computes SHA-256 hash, downloads with content-type detection, and generates presigned URLs for OCR access.")
Component(importOrch, "CanonicalImportOrchestrator", "Spring Service — @Async", "Runs the four canonical loaders in an explicit dependency DAG (TagTree → PersonRegister → PersonTree → Document). Smoke-checks all four artifacts before starting, owns the IDLE/RUNNING/DONE/FAILED state machine, fails closed on a malformed artifact.")
Component(tagTreeLoader, "TagTreeImporter", "Spring Component", "Upserts the tag hierarchy from canonical-tag-tree.xlsx via TagService (by canonical tag_path).")
Component(personRegLoader, "PersonRegisterImporter", "Spring Component", "Upserts register persons from canonical-persons.xlsx via PersonService (by normalizer person_id).")
Component(personTreeLoader, "PersonTreeImporter", "Spring Component", "Upserts tree persons + relationships from canonical-persons-tree.json via PersonService and RelationshipService.")
Component(docLoader, "DocumentImporter", "Spring Component", "Loads canonical-documents.xlsx: routes attribution register-first (raw cell always retained in sender_text/receiver_text), parses clean dates, builds an honest precision-aware title via DocumentTitleFormatter, keeps the S3 upload + thumbnail plumbing, and resolves each PDF by index (importDir/<index>.pdf) guarded by strict index validation + canonical-path containment + %PDF magic-byte check (no recursive walk).")
Component(titleFmt, "DocumentTitleFormatter", "Pure helper", "Formats the date label baked into an import title at exactly the data's precision (MONTH -> 'Juni 1916', never a fabricated day). Mirrors the frontend formatDocumentDate; both are pinned to docs/date-label-fixtures.json (#666).")
Component(docLoader, "DocumentImporter", "Spring Component", "Loads canonical-documents.xlsx: routes attribution register-first (raw cell always retained in sender_text/receiver_text), parses clean dates, builds the title via DocumentTitleFactory, keeps the S3 upload + thumbnail plumbing, and resolves each PDF by index (importDir/<index>.pdf) guarded by strict index validation + canonical-path containment + %PDF magic-byte check (no recursive walk).")
Component(titleFactory, "DocumentTitleFactory", "Spring Component", "Single source of truth for the auto-title {index} {dateLabel} {location} (#726). The document package owns this formula; importer, save-time regeneration, and the backfill all build through it so they never diverge.")
Component(titleFmt, "DocumentTitleFormatter", "Pure helper (document pkg)", "Formats the date label at exactly the data's precision (MONTH -> 'Juni 1916', never a fabricated day). Mirrors the frontend formatDocumentDate; both are pinned to docs/date-label-fixtures.json (#666).")
Component(titleMatcher, "DocumentTitleBackfillMatcher", "Pure helper", "Backfill-only heuristic deciding whether a STORED title is machine-generated (overwritable) vs hand-written prose. Index matched literally (no regex injection / ReDoS); fail-closed.")
Component(sheetReader, "CanonicalSheetReader", "POI helper", "Maps a canonical .xlsx by header name (no positional indices), splits pipe-delimited list columns, fails closed (IMPORT_ARTIFACT_INVALID) on a missing required header.")
Component(minioConf, "MinioConfig", "Spring @Configuration", "Creates the S3Client and S3Presigner beans with path-style access for MinIO. Validates MinIO connectivity on startup.")
Component(docRepo, "DocumentRepository", "Spring Data JPA", "Queries documents with Specification-based dynamic search, full-text search with ranking and match highlighting, and transcription pipeline queue projections.")
@@ -44,7 +46,11 @@ Rel(importOrch, docLoader, "4. Loads documents")
Rel(tagTreeLoader, sheetReader, "Reads canonical .xlsx")
Rel(personRegLoader, sheetReader, "Reads canonical .xlsx")
Rel(docLoader, sheetReader, "Reads canonical .xlsx")
Rel(docLoader, titleFmt, "Builds honest title date")
Rel(docLoader, titleFactory, "Builds the auto-title")
Rel(docSvc, titleFactory, "Regenerates auto-title (save-time + backfill)")
Rel(docSvc, titleMatcher, "Backfill overwrite test")
Rel(titleFactory, titleFmt, "Formats the honest date label")
Rel(adminCtrl, docSvc, "backfillTitles() / backfillFileHashes()")
Rel(tagTreeLoader, tagSvc, "Upserts tags by source_ref")
Rel(personRegLoader, personSvc, "Upserts persons by source_ref")
Rel(personTreeLoader, personSvc, "Upserts persons by source_ref")

View File

@@ -10,6 +10,11 @@ System_Boundary(frontend, "Web Frontend (SvelteKit / SSR)") {
Component(homePage, "/ (Home / Search)", "SvelteKit Route", "Loader: parses URL params (q, from, to, senderId, receiverId, tags), fetches /api/documents/search and /api/persons. Renders search form with full-text, date range, sender/receiver typeahead, and tag filters.")
Component(docsListPageTs, "/documents/+page.ts", "SvelteKit Client Loader", "Client-side load gated by matchMedia('(min-width: 1024px)') and ?view query. Fetches /api/documents/density only on desktop (Tailwind lg breakpoint) and outside calendar view; degrades to empty buckets on network failure.")
Component(timelineFilter, "TimelineDensityFilter.svelte", "Svelte Component", "Per-month density bars above the document list. Click selects a single month, emits onchange({from, to}) using YYYY-MM-DD boundaries. Hidden on mobile and tablet (below lg, 1024px) and in calendar view.")
Component(searchFilterBar, "SearchFilterBar.svelte", "Svelte Component", "Search/filter card on /documents. Hosts the keyword input, sort, advanced filters, and the smart-mode toggle. In smart mode submits the NL query on Enter via onSmartSearch instead of the live keyword search.")
Component(smartToggle, "search/SmartModeToggle.svelte", "Svelte Component", "Toggle pill (KI/Text) inside the search input. aria-pressed; switches between keyword and NL (smart) search modes.")
Component(chipRow, "search/InterpretationChipRow.svelte", "Svelte Component", "Renders NL interpretation chips (Absender / directional / Zeitraum / Stichwort). Removing a chip emits onRemoveChip; the page re-runs a keyword GET with the remaining params.")
Component(smartStatus, "search/SmartSearchStatus.svelte", "Svelte Component", "Full-area panels for NL search: loading (role=status), 503 SMART_SEARCH_UNAVAILABLE (with keyword fallback), 429 SMART_SEARCH_RATE_LIMITED.")
Component(disambig, "search/DisambiguationPicker.svelte", "Svelte Component", "Accessible single-select disclosure for ambiguous person names; selecting a candidate re-runs the search via GET.")
Component(docDetail, "/documents/[id]", "SvelteKit Route", "Loader: GET /api/documents/{id}. Page: metadata panel, inline file viewer, transcription editor, annotation layer, and comment thread.")
Component(docEdit, "/documents/[id]/edit", "SvelteKit Route", "Edit form with PersonTypeahead, TagInput, date/location fields. Form action: PUT /api/documents/{id}.")
Component(docNew, "/documents/new", "SvelteKit Route", "Upload form for a new document. Loader: GET /api/persons. Form action: POST /api/documents with multipart file.")
@@ -25,6 +30,12 @@ Rel(user, homePage, "Searches and browses", "HTTPS / Browser")
Rel(homePage, backend, "GET /api/documents/search, GET /api/persons", "HTTP / JSON")
Rel(docsListPageTs, backend, "GET /api/documents/density (desktop only, ≥1024px)", "HTTP / JSON")
Rel(homePage, timelineFilter, "Mounts above the result list")
Rel(homePage, searchFilterBar, "Mounts the search/filter card")
Rel(searchFilterBar, smartToggle, "Embeds the smart-mode toggle in the input")
Rel(homePage, backend, "POST /api/search/nl (smart mode)", "HTTP / JSON")
Rel(homePage, smartStatus, "Renders loading / 503 / 429 panels")
Rel(homePage, chipRow, "Renders interpretation chips; handles chip removal")
Rel(homePage, disambig, "Renders the picker when names are ambiguous")
Rel(docsListPageTs, timelineFilter, "Provides density / minDate / maxDate props")
Rel(docDetail, backend, "GET /api/documents/{id}, GET /api/documents/{id}/file", "HTTP / JSON + Binary")
Rel(docEdit, backend, "PUT /api/documents/{id}", "HTTP / Multipart")

View File

@@ -260,7 +260,7 @@ package "Transcription" {
entity transcription_block_mentioned_persons {
block_id : UUID <<FK>>
person_id : UUID NOT NULL
person_id : UUID NOT NULL <<FK>>
--
display_name : VARCHAR(200) NOT NULL
}
@@ -386,9 +386,9 @@ invite_token_group_ids }o--|| invite_tokens : invite_token_id
invite_token_group_ids }o--|| user_groups : group_id
' Document relationships
documents }o--o| persons : sender_id
documents }o--o| persons : sender_id (ON DELETE SET NULL)
document_receivers }o--|| documents : document_id
document_receivers }o--|| persons : person_id
document_receivers }o--|| persons : person_id (ON DELETE CASCADE)
document_tags }o--|| documents : document_id
document_tags }o--|| tag : tag_id
document_versions }o--|| documents : document_id
@@ -420,6 +420,7 @@ transcription_blocks }o--o| app_users : updated_by
transcription_block_versions }o--|| transcription_blocks : block_id
transcription_block_versions }o--o| app_users : changed_by
transcription_block_mentioned_persons }o--|| transcription_blocks : block_id
transcription_block_mentioned_persons }o--|| persons : person_id (ON DELETE CASCADE)
' OCR relationships
ocr_job_documents }o--|| ocr_jobs : job_id

View File

@@ -79,9 +79,9 @@ invite_token_group_ids }o--|| invite_tokens : invite_token_id
invite_token_group_ids }o--|| user_groups : group_id
' Document relationships
documents }o--o| persons : sender_id
documents }o--o| persons : sender_id (ON DELETE SET NULL)
document_receivers }o--|| documents : document_id
document_receivers }o--|| persons : person_id
document_receivers }o--|| persons : person_id (ON DELETE CASCADE)
document_tags }o--|| documents : document_id
document_tags }o--|| tag : tag_id
document_versions }o--|| documents : document_id
@@ -113,6 +113,7 @@ transcription_blocks }o--o| app_users : updated_by
transcription_block_versions }o--|| transcription_blocks : block_id
transcription_block_versions }o--o| app_users : changed_by
transcription_block_mentioned_persons }o--|| transcription_blocks : block_id
transcription_block_mentioned_persons }o--|| persons : person_id (ON DELETE CASCADE)
' OCR relationships
ocr_job_documents }o--|| ocr_jobs : job_id

View File

@@ -20,24 +20,19 @@ The observability stack (Prometheus, Loki, Grafana, Tempo, GlitchTip) ships as a
---
## VPS Sizing Recommendations
## Server Sizing
### Recommended: Hetzner CX32
### Current Production Server: Hetzner Dedicated (Serverbörse)
**Specs**: 4 vCPU, 8 GB RAM, 80 GB SSD · **Cost**: 17 EUR/mo
**Specs**: Intel Core i7-6700 (4C/8T, 3.4 GHz), 64 GB RAM · acquired via Hetzner server auction
Sufficient for the application stack (Postgres, MinIO, OCR with `mem_limit: 12g`, backend, frontend, Caddy) on a CX32 today. Once the observability stack lands (Prometheus/Loki/Grafana/Alertmanager add ~2 GB) consider a CX42.
Comfortably handles the full application stack (Postgres, MinIO, OCR with `mem_limit: 12g`, backend, frontend, Caddy, full observability stack) with headroom to spare. The 64 GB RAM means OCR, Ollama inference, and the observability stack can all run concurrently without memory pressure.
### When to Upgrade: Hetzner CX42
### When to Reconsider Hardware
**Specs**: 8 vCPU, 16 GB RAM · **Cost**: 29 EUR/mo
Upgrade when:
- Observability stack adds memory pressure (Loki + Grafana with >30 days retention)
- OCR throughput needs scaling beyond a single-node Surya/Kraken setup
- Real user load profiled in Grafana shows response-time degradation
Never upgrade the VPS tier before profiling — most perceived performance issues are application bugs, not resource constraints.
- CPU is Skylake (2015) — single-threaded performance is the likely bottleneck before RAM
- Profile with Grafana dashboards before concluding hardware is the constraint
- Most perceived performance issues are application bugs (unindexed queries, N+1 loads), not resource limits
---
@@ -45,12 +40,11 @@ Never upgrade the VPS tier before profiling — most perceived performance issue
| Service | Cost |
|---|---|
| Hetzner CX32 VPS | 17.00 EUR |
| Hetzner dedicated server (Serverbörse, i7-6700, 64 GB RAM) | see invoice |
| Hetzner DNS | 0.00 EUR |
| Hetzner SMTP relay | ~1.00 EUR |
| **Total** | **~18 EUR/mo** |
MinIO data lives on the VPS disk (no Object Storage line item yet). The Hetzner OBS migration would add ~5 EUR/mo at ~200 GB.
MinIO data lives on the server disk (no Object Storage line item yet). The Hetzner OBS migration would add ~5 EUR/mo at ~200 GB.
Equivalent SaaS stack: 200300 EUR/mo.

View File

@@ -0,0 +1,808 @@
<!DOCTYPE html>
<html lang="de">
<head>
<meta charset="UTF-8"/>
<meta name="viewport" content="width=device-width, initial-scale=1.0"/>
<title>Lesereisen — Journey-Editor · Familienarchiv</title>
<link href="https://fonts.googleapis.com/css2?family=Fraunces:opsz,wght@9..144,300;9..144,400;9..144,500&family=DM+Sans:wght@300;400;500;600&family=DM+Mono:wght@400;500&display=swap" rel="stylesheet"/>
<style>
:root{--color-page:#FAFAF7;--color-surface:#F5F4EE;--color-subtle:#EDECEA;--color-border:#D8D7D0;--color-text-muted:#6B6A63;--color-text:#1C1C18;--navy:#012851;--mint:#A1DCD8;--sand:#F0EFE9;--turquoise:#00C7B1;--blue-tint:#E6F1FB;--blue:#2D7DD2;--blue-dark:#185FA5;--green-tint:#E8F5EA;--green:#3D8C4A;--green-dark:#2E6E39;--orange-tint:#FEF0E6;--orange:#E8862A;--orange-dark:#B46820;--font-display:'Fraunces',Georgia,serif;--font-sans:'DM Sans',system-ui,sans-serif;--font-mono:'DM Mono',monospace;--radius-sm:4px;--radius-md:6px;--radius-lg:10px;--radius-xl:16px;--shadow-card:0 1px 3px rgba(28,28,24,.06),0 1px 2px rgba(28,28,24,.04);--shadow-raised:0 4px 12px rgba(28,28,24,.08),0 2px 4px rgba(28,28,24,.04);--shadow-overlay:0 8px 32px rgba(28,28,24,.12),0 2px 8px rgba(28,28,24,.06);}
*,*::before,*::after{box-sizing:border-box;margin:0;padding:0;}
body{font-family:var(--font-sans);background:#E8E7E2;color:var(--color-text);font-size:14px;line-height:1.6;}
.doc{max-width:1200px;margin:0 auto;padding:48px 40px 120px;}
.doc-header{display:flex;justify-content:space-between;align-items:flex-end;padding-bottom:28px;border-bottom:1px solid var(--color-border);margin-bottom:48px;background:var(--color-page);margin:-48px -40px 48px;padding:48px 40px 28px;border-radius:var(--radius-xl) var(--radius-xl) 0 0;}
.doc-header h1{font-family:var(--font-display);font-size:28px;font-weight:500;letter-spacing:-.02em;margin-bottom:4px;}
.doc-header p{font-size:13px;color:var(--color-text-muted);max-width:680px;}
.doc-meta{font-family:var(--font-mono);font-size:11px;color:var(--color-text-muted);text-align:right;line-height:1.9;}
.pill{display:inline-block;padding:2px 8px;border-radius:var(--radius-sm);font-size:10px;font-weight:500;letter-spacing:.05em;}
.pill-o{background:var(--orange-tint);color:var(--orange-dark);}
.section{margin-bottom:64px;}
.section-title{font-size:10px;font-weight:500;letter-spacing:.12em;text-transform:uppercase;color:var(--color-text-muted);padding-bottom:10px;border-bottom:1px solid var(--color-border);margin-bottom:24px;}
.prose{font-size:13px;color:var(--color-text-muted);line-height:1.65;max-width:720px;margin-bottom:20px;}
.jh{padding:20px 24px;border-radius:var(--radius-xl);margin-bottom:40px;display:flex;align-items:center;gap:16px;}
.jh .jn{font-family:var(--font-display);font-size:48px;font-weight:300;line-height:1;opacity:.5;}
.jh h2{font-family:var(--font-display);font-size:22px;font-weight:500;letter-spacing:-.02em;margin-bottom:4px;}
.jh p{font-size:13px;line-height:1.5;}.jh .fl{font-family:var(--font-mono);font-size:11px;margin-top:6px;opacity:.7;}
.jh-o{background:var(--orange-tint);border:1px solid #F0C99A;}
.jh-o .jn{color:var(--orange);}
.jh-o p,.jh-o .fl{color:var(--orange-dark);}
.scr{margin-bottom:56px;}
.scr-head{display:flex;justify-content:space-between;align-items:center;margin-bottom:6px;}
.scr-head h3{font-family:var(--font-display);font-size:20px;font-weight:500;letter-spacing:-.02em;}
.scr-id{font-family:var(--font-mono);font-size:11px;color:var(--color-text-muted);padding:2px 8px;border:1px solid var(--color-border);border-radius:var(--radius-sm);background:var(--color-page);}
.scr-desc{font-size:12px;color:var(--color-text-muted);line-height:1.6;max-width:720px;margin-bottom:6px;}
.scr-var{font-size:11px;color:var(--color-text-muted);margin-bottom:20px;}.scr-var strong{color:var(--color-text);}
.previews{display:flex;gap:32px;flex-wrap:wrap;justify-content:center;align-items:flex-start;margin-bottom:20px;}
.prev-col{display:flex;flex-direction:column;align-items:center;gap:10px;}
.bp-lbl{font-family:var(--font-mono);font-size:10px;color:var(--color-text-muted);}
.desk{width:100%;max-width:1040px;background:var(--color-page);border-radius:var(--radius-xl);overflow:hidden;box-shadow:var(--shadow-overlay),0 0 0 1px rgba(0,0,0,.06);display:flex;flex-direction:column;}
.phone{width:320px;flex-shrink:0;background:var(--color-page);border-radius:36px;overflow:hidden;box-shadow:var(--shadow-overlay),0 0 0 1px rgba(0,0,0,.07);display:flex;flex-direction:column;border:6px solid #1C1C18;}
.pst{padding:10px 20px 0;display:flex;justify-content:space-between;align-items:center;font-size:12px;background:var(--color-page);}.pst b{font-weight:600;}.pst span{font-size:10px;}
.pb{flex:1;overflow-y:auto;display:flex;flex-direction:column;}
.fa-nav{height:32px;background:var(--navy);display:flex;align-items:center;padding:0 12px;gap:8px;flex-shrink:0;}
.fa-logo{font-size:7px;font-weight:900;color:#fff;letter-spacing:.8px;border-bottom:2px solid var(--mint);padding-bottom:1px;}
.fa-link{font-size:5.5px;color:rgba(255,255,255,.4);font-weight:700;text-transform:uppercase;letter-spacing:.05em;}
.fa-link.active{color:var(--mint);}
.fa-nav-r{margin-left:auto;display:flex;gap:5px;align-items:center;}
.fa-av{width:16px;height:16px;background:rgba(255,255,255,.1);border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:5px;font-weight:800;color:rgba(255,255,255,.5);}
.m-nav{height:26px;background:var(--navy);display:flex;align-items:center;padding:0 10px;gap:6px;flex-shrink:0;}
.m-logo{font-size:6px;font-weight:900;color:#fff;letter-spacing:.7px;border-bottom:1.5px solid var(--mint);padding-bottom:1px;}
.m-nav-r{margin-left:auto;display:flex;gap:4px;align-items:center;}
.m-av{width:14px;height:14px;background:rgba(255,255,255,.1);border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:4.5px;font-weight:800;color:rgba(255,255,255,.5);}
.m-ham{display:flex;flex-direction:column;gap:2px;width:12px;}
.m-ham span{height:1.5px;background:rgba(255,255,255,.6);border-radius:1px;}
/* ── impl-ref table ── */
.agent{background:var(--color-text);color:#E8E8E2;padding:24px;border-radius:var(--radius-lg);margin-top:20px;}
.agent h4{font-size:9px;font-weight:500;letter-spacing:.1em;text-transform:uppercase;color:#5A5A55;margin-bottom:12px;}
.at{width:100%;border-collapse:collapse;font-family:var(--font-mono);font-size:10px;}
.at thead tr{border-bottom:1px solid #2A2A26;}
.at th{text-align:left;padding:6px 10px;font-size:8px;font-weight:500;letter-spacing:.08em;text-transform:uppercase;color:#5A5A55;font-family:var(--font-sans);}
.at td{padding:5px 10px;border-bottom:1px solid #1E1E1A;vertical-align:top;line-height:1.5;}
.at tr:last-child td{border-bottom:none;}
.at td:first-child{color:#7A7A72;}
.at td:nth-child(2){color:#E8E8E2;font-weight:500;}
.at td:nth-child(3){color:#5A5A55;}
.at .grp td{padding-top:14px;font-family:var(--font-sans);font-size:8px;font-weight:500;letter-spacing:.08em;text-transform:uppercase;color:#3A3A36;}
/* ── LLM guide ── */
.llm{background:var(--color-page);border:2px solid var(--navy);border-radius:var(--radius-xl);padding:32px 40px;margin-top:64px;}
.llm h2{font-family:var(--font-display);font-size:22px;font-weight:500;letter-spacing:-.02em;margin-bottom:8px;color:var(--navy);}
.llm h3{font-size:14px;font-weight:600;margin:20px 0 8px;}
.llm h4{font-size:12px;font-weight:600;margin:14px 0 6px;color:var(--color-text-muted);}
.llm p,.llm li{font-size:13px;color:var(--color-text-muted);line-height:1.65;}
.llm ul,.llm ol{padding-left:20px;margin-bottom:12px;}
.llm li{margin-bottom:4px;}
.llm code{font-family:var(--font-mono);font-size:11px;background:var(--color-surface);padding:1px 5px;border-radius:3px;}
.llm table{width:100%;border-collapse:collapse;margin:12px 0;font-size:12px;}
.llm th,.llm td{text-align:left;padding:6px 10px;border-bottom:1px solid var(--color-border);}
.llm th{font-weight:500;color:var(--color-text);font-size:11px;text-transform:uppercase;letter-spacing:.05em;}
.llm td{color:var(--color-text-muted);}
/* ── Editor chrome (shared with writer spec) ── */
.ed-topbar{background:#fff;border-bottom:1px solid #e4e2d7;display:flex;align-items:center;padding:0 14px;gap:8px;height:38px;flex-shrink:0;}
.ed-back{width:22px;height:22px;border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:9px;color:var(--color-text-muted);flex-shrink:0;}
.ed-title-label{font-family:var(--font-sans);font-size:10px;font-weight:500;color:var(--color-text);flex:1;}
.ed-status-pill{display:inline-flex;align-items:center;padding:2px 7px;border-radius:20px;font-size:8px;font-weight:700;letter-spacing:.06em;text-transform:uppercase;flex-shrink:0;}
.ed-status-draft{background:#F0EFE9;color:#6B6A63;border:1px solid #D8D7D0;}
.ed-status-pub{background:var(--green-tint);color:var(--green-dark);border:1px solid #A0D8A8;}
.ed-delete-link{font-size:8px;font-weight:600;color:#DC4C3E;margin-left:8px;}
.ed-split{display:flex;flex:1;overflow:hidden;}
.ed-sidebar{width:210px;flex-shrink:0;border-left:1px solid #e4e2d7;background:#fff;display:flex;flex-direction:column;overflow-y:auto;}
.ed-sb-section{padding:12px 12px 10px;}
.ed-sb-title{font-size:8px;font-weight:700;letter-spacing:.1em;text-transform:uppercase;color:var(--color-text-muted);margin-bottom:8px;}
.ed-sb-divider{height:1px;background:#e4e2d7;}
.ed-search-row{display:flex;align-items:center;gap:6px;background:var(--color-surface);border:1px solid var(--color-border);border-radius:var(--radius-sm);padding:4px 8px;margin-bottom:6px;}
.ed-search-input{font-size:9px;color:var(--color-text-muted);}
.ed-chip{display:inline-flex;align-items:center;gap:4px;padding:3px 7px;background:var(--sand);border:1px solid var(--color-border);border-radius:12px;font-size:8px;font-weight:500;color:var(--color-text);margin:0 4px 4px 0;}
.ed-chip-x{color:var(--color-text-muted);font-size:9px;cursor:pointer;margin-left:2px;}
.ed-hint{font-size:8px;color:var(--color-text-muted);line-height:1.5;margin-top:4px;}
.ed-savebar{background:#fff;border-top:1px solid #e4e2d7;padding:9px 14px;display:flex;align-items:center;justify-content:space-between;flex-shrink:0;gap:10px;}
.ed-savebar-hint{font-size:8px;color:var(--color-text-muted);}
.ed-savebar-actions{display:flex;align-items:center;gap:7px;}
.ed-btn-ghost{font-size:9px;font-weight:600;padding:5px 12px;border-radius:var(--radius-sm);border:1px solid var(--color-border);color:var(--color-text);background:#fff;cursor:pointer;white-space:nowrap;}
.ed-btn-ghost.retract{color:#B46820;border-color:#E8D5B0;}
.ed-btn-primary{font-size:9px;font-weight:600;padding:5px 12px;border-radius:var(--radius-sm);background:var(--navy);color:#fff;border:none;cursor:pointer;white-space:nowrap;}
/* ── Journey Editor main area ── */
.je-main{flex:1;display:flex;flex-direction:column;padding:14px 16px;overflow-y:auto;gap:8px;background:var(--color-page);}
.je-title-input{font-family:var(--font-display);font-size:15px;font-weight:400;color:var(--color-text);border:none;border-bottom:1px solid var(--color-border);padding:4px 0 6px;width:100%;outline:none;background:transparent;letter-spacing:-.01em;}
.je-title-input.placeholder{color:var(--color-text-muted);font-style:italic;}
.je-sep{height:1px;background:var(--color-border);margin:2px 0;}
.je-intro-area{font-family:Georgia,serif;font-size:9px;line-height:1.7;color:var(--color-text-muted);font-style:italic;border:none;padding:5px 0;width:100%;outline:none;background:transparent;min-height:36px;resize:none;}
.je-intro-label{font-size:7.5px;font-weight:600;letter-spacing:.07em;text-transform:uppercase;color:var(--color-text-muted);margin-bottom:2px;}
.je-list-label{font-size:7.5px;font-weight:600;letter-spacing:.07em;text-transform:uppercase;color:var(--color-text-muted);margin-bottom:5px;margin-top:4px;}
/* ── Item rows ── */
.je-item{display:flex;align-items:stretch;gap:0;background:#fff;border:1px solid #E4E2D7;border-radius:4px;margin-bottom:5px;overflow:hidden;}
.je-drag{width:16px;background:#F5F4EE;border-right:1px solid #E4E2D7;display:flex;align-items:center;justify-content:center;cursor:grab;flex-shrink:0;}
.je-drag-dots{display:flex;flex-direction:column;gap:2px;}
.je-drag-dot{width:3px;height:3px;border-radius:50%;background:#C4C3BC;}
.je-num{width:20px;display:flex;align-items:flex-start;justify-content:center;padding-top:8px;font-size:8px;font-weight:700;color:#9B9A93;flex-shrink:0;}
.je-body{flex:1;padding:7px 8px 7px 4px;}
.je-doc-title{font-size:9px;font-weight:600;color:var(--navy);line-height:1.3;margin-bottom:2px;}
.je-doc-meta{font-size:7.5px;color:var(--color-text-muted);margin-bottom:5px;}
.je-note-area{width:100%;min-height:32px;font-family:Georgia,serif;font-size:8px;line-height:1.55;color:var(--color-text);font-style:italic;border:1px solid var(--color-border);border-radius:3px;background:var(--color-surface);padding:4px 6px;resize:none;outline:none;}
.je-note-add{font-size:7.5px;font-weight:600;color:var(--blue);cursor:pointer;display:inline-flex;align-items:center;gap:2px;}
.je-remove{width:24px;display:flex;align-items:flex-start;justify-content:center;padding-top:7px;flex-shrink:0;}
.je-remove-x{font-size:11px;color:#C4C3BC;cursor:pointer;line-height:1;font-weight:300;}
.je-interlude-bg{background:var(--orange-tint);border-color:#F0C99A;}
.je-interlude-icon{font-size:8px;color:var(--orange);margin-bottom:2px;}
.je-interlude-area{width:100%;min-height:36px;font-family:Georgia,serif;font-size:8px;line-height:1.6;color:var(--color-text);font-style:italic;border:1px solid #F0C99A;border-radius:3px;background:rgba(255,255,255,.6);padding:4px 6px;resize:none;outline:none;}
.je-empty{padding:16px;text-align:center;border:1px dashed var(--color-border);border-radius:4px;background:var(--color-surface);}
.je-empty-text{font-family:Georgia,serif;font-size:8px;color:var(--color-text-muted);font-style:italic;}
/* ── Add bar ── */
.je-add-bar{display:flex;gap:7px;padding:6px 0 4px;}
.je-add-btn{font-size:8px;font-weight:600;padding:5px 10px;border-radius:3px;border:1px dashed var(--color-border);color:var(--color-text-muted);background:transparent;cursor:pointer;display:flex;align-items:center;gap:3px;}
.je-add-btn:hover{border-color:var(--navy);color:var(--navy);}
/* ── Inline note editing state (highlight) ── */
.je-note-editing{border-color:var(--navy);background:#fff;}
/* ── Mobile journey editor ── */
.mob-topbar{background:#fff;border-bottom:1px solid #e4e2d7;display:flex;align-items:center;padding:0 10px;gap:6px;height:34px;flex-shrink:0;}
.mob-back{font-size:8px;color:var(--color-text-muted);}
.mob-label{font-family:var(--font-sans);font-size:9px;font-weight:500;color:var(--color-text);flex:1;}
.mob-body{flex:1;overflow-y:auto;padding:10px 12px;display:flex;flex-direction:column;gap:7px;background:var(--color-page);}
.mob-title-input{font-family:var(--font-display);font-size:13px;color:var(--color-text-muted);font-style:italic;border:none;border-bottom:1px solid var(--color-border);padding:3px 0 5px;width:100%;background:transparent;outline:none;}
.mob-collapsible{background:#fff;border:1px solid #e4e2d7;border-radius:3px;overflow:hidden;}
.mob-coll-hdr{display:flex;align-items:center;justify-content:space-between;padding:7px 9px;font-size:8.5px;font-weight:600;color:var(--color-text);}
.mob-coll-chevron{font-size:9px;color:var(--color-text-muted);}
.mob-savebar{background:#fff;border-top:1px solid #e4e2d7;padding:8px 10px;display:flex;gap:6px;flex-shrink:0;}
.mob-btn{font-size:8.5px;font-weight:600;padding:7px 0;border-radius:3px;text-align:center;flex:1;}
.mob-btn-ghost{border:1px solid var(--color-border);color:var(--color-text);background:#fff;}
.mob-btn-primary{background:var(--navy);color:#fff;border:none;}
.mob-je-item{display:flex;align-items:stretch;gap:0;background:#fff;border:1px solid #E4E2D7;border-radius:3px;margin-bottom:4px;overflow:hidden;}
.mob-je-drag{width:14px;background:#F5F4EE;border-right:1px solid #E4E2D7;display:flex;align-items:center;justify-content:center;}
.mob-je-body{flex:1;padding:6px 7px;}
.mob-je-title{font-size:8.5px;font-weight:600;color:var(--navy);line-height:1.3;margin-bottom:1px;}
.mob-je-meta{font-size:7px;color:var(--color-text-muted);}
.mob-je-note{margin-top:4px;padding:3px 5px;background:var(--color-surface);border-left:2px solid var(--mint);font-size:7.5px;font-style:italic;color:var(--color-text-muted);}
.mob-je-interlude{background:var(--orange-tint);border-color:#F0C99A;}
.mob-je-interlude-text{font-size:7.5px;font-style:italic;color:var(--color-text);}
@media(max-width:900px){.doc{padding:24px 16px 80px;}}
</style>
</head>
<body>
<div class="doc">
<!-- ═══ DOC HEADER ═══ -->
<div class="doc-header">
<div>
<h1>Lesereisen — Journey-Editor</h1>
<p>Kuratierungs-Oberfläche für <code>JourneyEditor</code> auf <code>/geschichten/[id]/edit</code> (wenn <code>type === 'JOURNEY'</code>). Geordnete Briefliste mit Drag-to-Reorder, Dokumenten-Picker, Interlude-Notizen und Inline-Annotation-Editing. Ersetzt den TipTap-Editor für den Journey-Typ.</p>
</div>
<div class="doc-meta">
Familienarchiv<br/>
<span class="pill pill-o">Final Spec</span><br/>
2026-06-07 &middot; @leonievoss<br/>
<span style="font-size:10px;margin-top:4px;display:inline-block;">Issue #753</span>
</div>
</div>
<!-- ═══ JOURNEY HEADER ═══ -->
<div class="jh jh-o">
<div class="jn">E</div>
<div>
<h2>Journey-Editor</h2>
<p>BLOG_WRITERs kuratieren eine geordnete Briefsequenz — Briefe hinzufügen, Zwischentexte einfügen, Reihenfolge per Drag anpassen, Notizen inline bearbeiten.</p>
<div class="fl">/geschichten/[id]/edit (type === 'JOURNEY')</div>
</div>
</div>
<!-- ═══ KONZEPT ═══ -->
<div class="section">
<div class="section-title">Konzept</div>
<p class="prose">Der <code>JourneyEditor</code> ist eine parallele Implementierung zum bestehenden <code>GeschichteEditor</code> und wird auf derselben Edit-Route eingeblendet wenn <code>type === 'JOURNEY'</code>. Das Split-Layout (70/30) bleibt erhalten: links die Briefliste, rechts die Sidebar mit Personen und Status.</p>
<p class="prose">Die linke Fläche zeigt: oben einen optionalen Einleitungs-Textarea (<code>body</code>), darunter die geordnete Itemliste, ganz unten eine Aktionsleiste mit „+ Brief hinzufügen" und „+ Zwischentext hinzufügen". Jedes Item hat einen Drag-Handle, eine Positionsnummer, den Inhalt und einen Entfernen-Button.</p>
<p class="prose">Dokument-Items zeigen Titel und Kurz-Metadaten. Eine „Notiz hinzufügen/bearbeiten"-Aktion expandiert ein Textarea direkt unterhalb des Items — kein Modal, kein separates Formular. Interlude-Items (reiner Zwischentext) zeigen direkt ein editierbares Textarea mit orangenem Hintergrund zur klaren visuellen Unterscheidung.</p>
<p class="prose">Speicheraktionen: Speichern (bei veröffentlichter Journey) oder „Entwurf speichern" + „Veröffentlichen" (bei DRAFT). Die Savebarlogik ist identisch zum GeschichteEditor. Alle Mutationen lösen sofort einen API-Call aus und aktualisieren den lokalen Zustand optimistisch — kein separates Save für einzelne Items.</p>
</div>
<!-- ═══ SCREEN LE-1: EMPTY EDITOR ═══ -->
<div class="section">
<div class="section-title">Screens — Leerer Editor</div>
<div class="scr">
<div class="scr-head">
<h3>LE-1 — Journey-Editor leer</h3>
<span class="scr-id">Issue #753 · LE-1</span>
</div>
<p class="scr-desc">Ausgangszustand einer neuen oder leeren Lesereise. Titel-Input oben. Darunter ein optionaler Einleitungs-Textarea. Leere Itemliste mit Leerstate-Text. Aktionsleiste mit zwei Buttons. Sidebar: Personen-Verknüpfung und Status-Anzeige. Keine Items → „Veröffentlichen" noch nicht aktiv (Disabled-Hint erscheint).</p>
<p class="scr-var"><strong>Varianten:</strong> Neuer Entwurf ohne Titel (hier gezeigt) · Mit Titel, leere Liste</p>
<div class="previews">
<div class="prev-col" style="width:100%;max-width:1040px;">
<span class="bp-lbl">Desktop — 1040px · Neuer Entwurf</span>
<div class="desk" style="min-height:500px;">
<div class="fa-nav">
<span class="fa-logo">ARCHIV</span>
<span style="width:1px;height:14px;background:rgba(255,255,255,.1);margin:0 2px;"></span>
<span class="fa-link">Dokumente</span>
<span class="fa-link">Personen</span>
<span class="fa-link active">Geschichten</span>
<span class="fa-link">Chronik</span>
<div class="fa-nav-r">
<div class="fa-av" style="background:#012851;color:var(--mint);font-size:5px;font-weight:800;">MR</div>
</div>
</div>
<div class="ed-topbar">
<div class="ed-back">&#8592;</div>
<div class="ed-title-label" style="display:flex;align-items:center;gap:6px;">
Neue Lesereise
<span style="font-size:7px;font-weight:600;letter-spacing:.06em;text-transform:uppercase;color:var(--orange-dark);background:var(--orange-tint);border:1px solid #F0C99A;padding:1px 5px;border-radius:3px;">REISE</span>
</div>
<div class="ed-status-pill ed-status-draft">ENTWURF</div>
</div>
<div class="ed-split">
<!-- Left: Journey editor area -->
<div class="je-main">
<input class="je-title-input placeholder" type="text" value="" placeholder="Titel der Lesereise" readonly/>
<div class="je-sep"></div>
<div>
<div class="je-intro-label">Einleitung (optional)</div>
<textarea class="je-intro-area" placeholder="Optionaler Einleitungstext für diese Lesereise…" readonly></textarea>
</div>
<div class="je-sep"></div>
<div class="je-list-label">Briefe &amp; Zwischentexte</div>
<div class="je-empty">
<div class="je-empty-text">Noch keine Einträge. Füge den ersten Brief oder einen Zwischentext hinzu.</div>
</div>
<div class="je-add-bar">
<button class="je-add-btn">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Brief hinzufügen
</button>
<button class="je-add-btn">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Zwischentext hinzufügen
</button>
</div>
</div>
<!-- Right: Sidebar -->
<div class="ed-sidebar">
<div class="ed-sb-section">
<div class="ed-sb-title">Personen</div>
<div class="ed-search-row">
<span style="font-size:9px;color:var(--color-text-muted);">&#128269;</span>
<div class="ed-search-input">Person suchen…</div>
</div>
<div class="ed-hint">Welche historischen Personen kommen in dieser Lesereise vor?</div>
</div>
<div class="ed-sb-divider"></div>
<div class="ed-sb-section">
<div class="ed-sb-title">Status</div>
<div class="ed-status-pill ed-status-draft" style="font-size:9px;">ENTWURF</div>
<div class="ed-hint" style="margin-top:6px;">Noch nicht öffentlich sichtbar. Füge mindestens einen Brief hinzu, um zu veröffentlichen.</div>
</div>
</div>
</div>
<div class="ed-savebar">
<span class="ed-savebar-hint">Alle Änderungen werden als Entwurf gespeichert.</span>
<div class="ed-savebar-actions">
<button class="ed-btn-ghost">Entwurf speichern</button>
<button class="ed-btn-primary" style="opacity:.4;cursor:not-allowed;">Veröffentlichen</button>
</div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LE-1 Leerer Editor</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Seitenstruktur</td></tr>
<tr><td>Bedingte Verzweigung</td><td>{#if geschichte.type === 'JOURNEY'}&lt;JourneyEditor /&gt;{:else}&lt;GeschichteEditor /&gt;{/if}</td><td>in edit/+page.svelte; Props: geschichte: Geschichte</td></tr>
<tr><td>Split-Layout</td><td>flex flex-1 overflow-hidden (gleich wie GeschichteEditor)</td><td>70/30; Sidebar identisch</td></tr>
<tr><td>Topbar-Badge</td><td>„REISE" Pill neben dem Titel-Label</td><td>orange; kein interaktives Element; zeigt Typ</td></tr>
<tr class="grp"><td colspan="3">Titel-Input</td></tr>
<tr><td>Titel-Input</td><td>font-serif text-2xl border-b border-line pb-2 w-full bg-transparent outline-none</td><td>bind:value={title}; gleiche Validierung wie GeschichteEditor (required)</td></tr>
<tr class="grp"><td colspan="3">Einleitungs-Textarea</td></tr>
<tr><td>Intro-Textarea</td><td>font-serif text-sm italic text-ink-3 leading-relaxed w-full resize-none bg-transparent outline-none border-none py-1</td><td>bind:value={body}; plaintext; auto-resize per rows-attr oder JS</td></tr>
<tr><td>Label</td><td>text-[10px] font-bold uppercase tracking-widest text-ink-3 mb-1</td><td>„EINLEITUNG (OPTIONAL)"</td></tr>
<tr class="grp"><td colspan="3">Leerstate</td></tr>
<tr><td>Leerstate-Container</td><td>py-8 text-center border border-dashed border-line rounded-sm bg-surface</td><td>verschwindet sobald erstes Item vorhanden</td></tr>
<tr><td>Leerstate-Text</td><td>font-serif text-xs text-ink-3 italic</td><td></td></tr>
<tr class="grp"><td colspan="3">Veröffentlichen-Button</td></tr>
<tr><td>Disabled-Zustand</td><td>disabled={items.length === 0 || !title.trim()}</td><td>opacity-40 + cursor-not-allowed; keine Tooltip nötig — Sidebar-Hint erklärt es</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ SCREEN LE-2: EDITOR WITH ITEMS ═══ -->
<div class="section">
<div class="section-title">Screens — Editor mit Einträgen</div>
<div class="scr">
<div class="scr-head">
<h3>LE-2 — Journey-Editor mit Einträgen</h3>
<span class="scr-id">Issue #753 · LE-2</span>
</div>
<p class="scr-desc">Gefüllte Itemliste mit gemischten Typen: Dokument-Item ohne Notiz, Interlude-Item (reiner Zwischentext), Dokument-Item mit bestehender Notiz. Jedes Item zeigt Drag-Handle links, Positionsnummer, Inhalt und Entfernen-Button. Aktionsleiste bleibt unter der Liste sichtbar.</p>
<p class="scr-var"><strong>Varianten:</strong> Veröffentlichte Journey (hier gezeigt) · Entwurf · Mobile</p>
<div class="previews">
<div class="prev-col" style="width:100%;max-width:1040px;">
<span class="bp-lbl">Desktop — 1040px · VERÖFFENTLICHT</span>
<div class="desk" style="min-height:580px;">
<div class="fa-nav">
<span class="fa-logo">ARCHIV</span>
<span style="width:1px;height:14px;background:rgba(255,255,255,.1);margin:0 2px;"></span>
<span class="fa-link">Dokumente</span>
<span class="fa-link">Personen</span>
<span class="fa-link active">Geschichten</span>
<div class="fa-nav-r">
<div class="fa-av" style="background:#012851;color:var(--mint);font-size:5px;font-weight:800;">KR</div>
</div>
</div>
<div class="ed-topbar">
<div class="ed-back">&#8592;</div>
<div class="ed-title-label" style="display:flex;align-items:center;gap:6px;">
Lesereise bearbeiten
<span style="font-size:7px;font-weight:600;letter-spacing:.06em;text-transform:uppercase;color:var(--orange-dark);background:var(--orange-tint);border:1px solid #F0C99A;padding:1px 5px;border-radius:3px;">REISE</span>
</div>
<div class="ed-status-pill ed-status-pub">VERÖFFENTLICHT</div>
<span class="ed-delete-link">Löschen</span>
</div>
<div class="ed-split">
<!-- Left: Journey editor area with items -->
<div class="je-main">
<input class="je-title-input" type="text" value="Briefe aus Breslau 19381942" readonly/>
<div class="je-sep"></div>
<div>
<div class="je-intro-label">Einleitung (optional)</div>
<textarea class="je-intro-area" readonly style="color:var(--color-text);">Der Briefwechsel zwischen Franz Raddatz und seiner Schwester Emma umspannt vier Jahre — von den letzten Friedenssommern bis zum Ende des Krieges.</textarea>
</div>
<div class="je-sep"></div>
<div class="je-list-label">Briefe &amp; Zwischentexte</div>
<!-- Item 1: Document, no note -->
<div class="je-item">
<div class="je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="je-num">1</div>
<div class="je-body">
<div class="je-doc-title">Brief vom 12. Juli 1938</div>
<div class="je-doc-meta">12. Juli 1938 &middot; von Franz Raddatz an Emma Müller</div>
<div class="je-note-add">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Notiz hinzufügen
</div>
</div>
<div class="je-remove"><div class="je-remove-x">&#215;</div></div>
</div>
<!-- Item 2: Interlude -->
<div class="je-item je-interlude-bg">
<div class="je-drag" style="background:rgba(232,134,42,.08);border-right-color:#F0C99A;">
<div class="je-drag-dots">
<div class="je-drag-dot" style="background:#D4A574;"></div><div class="je-drag-dot" style="background:#D4A574;"></div>
<div class="je-drag-dot" style="background:#D4A574;"></div><div class="je-drag-dot" style="background:#D4A574;"></div>
</div>
</div>
<div class="je-num" style="color:var(--orange-dark);">
<svg width="10" height="10" viewBox="0 0 12 12" fill="none" style="margin-top:7px;"><path d="M2 4h8M2 7h5" stroke="var(--orange)" stroke-width="1.5" stroke-linecap="round"/></svg>
</div>
<div class="je-body" style="padding-top:6px;">
<div style="font-size:7.5px;font-weight:600;letter-spacing:.06em;text-transform:uppercase;color:var(--orange-dark);margin-bottom:4px;">Zwischentext</div>
<textarea class="je-interlude-area" readonly>Im Sommer 1938 schrieb Franz voller Zuversicht — er hatte kaum eine Ahnung, wie bald sich die Welt um ihn herum verändern würde.</textarea>
</div>
<div class="je-remove"><div class="je-remove-x" style="color:#D4A574;">&#215;</div></div>
</div>
<!-- Item 3: Document with note -->
<div class="je-item">
<div class="je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="je-num">2</div>
<div class="je-body">
<div class="je-doc-title">Postkarte aus Breslau, August 1938</div>
<div class="je-doc-meta" style="margin-bottom:5px;">22. Aug. 1938 &middot; von Franz Raddatz an Emma Müller</div>
<textarea class="je-note-area" readonly>Diese Karte ist ungewöhnlich kurz für Franz — vier Zeilen, fast hastig. Ein Zeichen der aufkommenden Unruhe in den Nachrichten?</textarea>
<div class="je-note-add" style="margin-top:3px;color:var(--color-text-muted);">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M2 2l6 6M8 2l-6 6" stroke="currentColor" stroke-width="1.3" stroke-linecap="round"/></svg>
Notiz entfernen
</div>
</div>
<div class="je-remove"><div class="je-remove-x">&#215;</div></div>
</div>
<!-- Item 4: Document, no note -->
<div class="je-item">
<div class="je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="je-num">3</div>
<div class="je-body">
<div class="je-doc-title">Brief vom 3. September 1939</div>
<div class="je-doc-meta">3. Sept. 1939 &middot; von Emma Müller an Franz Raddatz</div>
<div class="je-note-add">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Notiz hinzufügen
</div>
</div>
<div class="je-remove"><div class="je-remove-x">&#215;</div></div>
</div>
<!-- Add bar -->
<div class="je-add-bar">
<button class="je-add-btn">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Brief hinzufügen
</button>
<button class="je-add-btn">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Zwischentext hinzufügen
</button>
</div>
</div>
<!-- Right: Sidebar -->
<div class="ed-sidebar">
<div class="ed-sb-section">
<div class="ed-sb-title">Personen</div>
<div class="ed-search-row">
<span style="font-size:9px;color:var(--color-text-muted);">&#128269;</span>
<div class="ed-search-input">Person suchen…</div>
</div>
<div style="display:flex;flex-wrap:wrap;margin-top:4px;">
<span class="ed-chip">
<span style="width:10px;height:10px;border-radius:50%;background:#012851;display:flex;align-items:center;justify-content:center;font-size:5px;font-weight:800;color:var(--mint);">FR</span>
Franz Raddatz
<span class="ed-chip-x">&#215;</span>
</span>
<span class="ed-chip">
<span style="width:10px;height:10px;border-radius:50%;background:#534AB7;display:flex;align-items:center;justify-content:center;font-size:5px;font-weight:800;color:#fff;">EM</span>
Emma Müller
<span class="ed-chip-x">&#215;</span>
</span>
</div>
</div>
<div class="ed-sb-divider"></div>
<div class="ed-sb-section">
<div class="ed-sb-title">Status</div>
<div class="ed-status-pill ed-status-pub" style="font-size:9px;">VERÖFFENTLICHT</div>
<div class="ed-hint" style="margin-top:6px;">Änderungen gehen sofort live.</div>
</div>
</div>
</div>
<div class="ed-savebar">
<span class="ed-savebar-hint">Änderungen sofort live — Leser sehen die aktuelle Version.</span>
<div class="ed-savebar-actions">
<button class="ed-btn-ghost retract">Zurück zu Entwurf</button>
<button class="ed-btn-primary">Speichern</button>
</div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LE-2 Items-Liste</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Item-Zeile allgemein</td></tr>
<tr><td>Item-Container</td><td>flex items-stretch bg-white border border-line rounded-sm mb-2 overflow-hidden</td><td>interlude: bg-orange-50 border-orange-200</td></tr>
<tr><td>Drag-Handle</td><td>w-4 bg-surface border-r border-line flex items-center justify-center cursor-grab shrink-0</td><td>aria-label="Reihenfolge ändern"; cursor-grabbing während Drag</td></tr>
<tr><td>Positions-Nr.</td><td>w-5 text-[10px] font-bold text-ink-3 flex items-start justify-center pt-2 shrink-0</td><td>aus Array-Index, nicht item.position</td></tr>
<tr><td>Entfernen-Button</td><td>w-6 flex items-start justify-center pt-2 shrink-0</td><td>× aria-label="Eintrag entfernen"; hover: text-red-500; Confirm nur wenn note vorhanden</td></tr>
<tr class="grp"><td colspan="3">Dokument-Item</td></tr>
<tr><td>Brieftitel</td><td>text-[11px] font-semibold text-ink leading-snug mb-0.5</td><td>document.title</td></tr>
<tr><td>Briefmeta</td><td>text-xs text-ink-3</td><td>formatDate(doc.documentDate) · "von X" oder "von X an Y"</td></tr>
<tr><td>Notiz-Textarea (sichtbar)</td><td>w-full min-h-[40px] font-serif text-xs italic bg-surface border border-line rounded-sm p-1.5 resize-none focus:border-primary focus:bg-white mt-2</td><td>auto-expand; bind:value={item.note}</td></tr>
<tr><td>„Notiz hinzufügen" Link</td><td>text-xs font-semibold text-blue-600 inline-flex items-center gap-1 mt-1</td><td>togglet Notiz-Textarea</td></tr>
<tr><td>„Notiz entfernen" Link</td><td>text-xs text-ink-3 inline-flex items-center gap-1 mt-1</td><td>zeigt sich wenn note.trim() nicht leer; setzt note = '' und blendet Textarea aus</td></tr>
<tr class="grp"><td colspan="3">Interlude-Item</td></tr>
<tr><td>Interlude-Container</td><td>bg-orange-50 border-orange-200 (überschreibt Item-Container)</td><td>kein Positions-Kreis; Positions-Spalte zeigt Icon statt Zahl</td></tr>
<tr><td>Label „Zwischentext"</td><td>text-[9px] font-bold uppercase tracking-widest text-orange-700 mb-1</td><td>immer sichtbar; nicht editierbar</td></tr>
<tr><td>Zwischentext-Textarea</td><td>w-full min-h-[44px] font-serif text-xs italic bg-white/60 border border-orange-200 rounded-sm p-1.5 resize-none focus:border-orange-400</td><td>bind:value={item.note}; auto-expand; min 44px für Touch-Target</td></tr>
<tr class="grp"><td colspan="3">Aktionsleiste</td></tr>
<tr><td>Add Bar</td><td>flex gap-2 pt-2 pb-1</td><td>immer unten sichtbar, auch wenn Liste gefüllt</td></tr>
<tr><td>„Brief hinzufügen" Button</td><td>border border-dashed border-line rounded-sm px-3 py-1.5 text-xs font-semibold text-ink-2 hover:border-primary hover:text-primary flex items-center gap-1</td><td>öffnet existierende DocumentPicker-Komponente als Dropdown/Modal</td></tr>
<tr><td>„Zwischentext hinzufügen" Button</td><td>gleich wie Brief-Button</td><td>fügt neues Interlude-Item am Ende ein; Fokus auf das neue Textarea</td></tr>
<tr class="grp"><td colspan="3">Drag-to-Reorder</td></tr>
<tr><td>Bibliothek</td><td>@dnd-kit/core oder svelte-dnd-action (bereits im Projekt prüfen)</td><td>kein neues Package ohne Absprache</td></tr>
<tr><td>Reorder-API-Call</td><td>PUT /api/geschichten/{id}/items/reorder — body: [{id, position}] für alle Items</td><td>nach jedem Drop ausgelöst; optimistisch: lokalen State sofort aktualisieren</td></tr>
<tr><td>Accessibility</td><td>Drag-Handle: role="button" tabIndex=0; Keyboard: Space startet Drag, Arrow hoch/runter verschiebt, Space/Enter bestätigt, Esc abbricht</td><td>WCAG 2.1 SC 2.1.1</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ SCREEN LE-3: INLINE NOTE EDITING ═══ -->
<div class="section">
<div class="section-title">Screens — Inline-Notiz-Editing</div>
<div class="scr">
<div class="scr-head">
<h3>LE-3 — Notiz-Textarea wird geöffnet</h3>
<span class="scr-id">Issue #753 · LE-3</span>
</div>
<p class="scr-desc">Wenn der Nutzer auf „Notiz hinzufügen" klickt, expandiert das Item um ein Textarea direkt unterhalb der Briefmeta — kein Modal. Der Fokus springt automatisch in das Textarea. Das Textarea hat einen blauen Fokusring als Orientierungshilfe. Ein API-PATCH wird beim Verlassen des Textareas (blur) ausgelöst, nicht bei jedem Tastendruck.</p>
<p class="scr-var"><strong>Inset-Ansicht — kein vollständiger Seiten-Mockup nötig</strong></p>
<div class="previews">
<div class="prev-col" style="width:100%;max-width:560px;">
<span class="bp-lbl">Inset — Notiz-Textarea geöffnet (Fokus)</span>
<div style="background:#E8E7E2;padding:16px;border-radius:var(--radius-xl);">
<!-- Item before (no note) -->
<div class="je-item" style="margin-bottom:5px;">
<div class="je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="je-num">1</div>
<div class="je-body">
<div class="je-doc-title">Brief vom 12. Juli 1938</div>
<div class="je-doc-meta">12. Juli 1938 &middot; Franz → Emma</div>
<div class="je-note-add">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/></svg>
Notiz hinzufügen
</div>
</div>
<div class="je-remove"><div class="je-remove-x">&#215;</div></div>
</div>
<!-- Item with opened note textarea (focused) -->
<div class="je-item">
<div class="je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="je-num">2</div>
<div class="je-body">
<div class="je-doc-title">Postkarte aus Breslau, August 1938</div>
<div class="je-doc-meta" style="margin-bottom:5px;">22. Aug. 1938 &middot; Franz → Emma</div>
<!-- Focused textarea -->
<textarea class="je-note-area je-note-editing" style="outline:none;box-shadow:0 0 0 2px rgba(1,40,81,.2);" readonly placeholder="Kuratoren-Notiz für diesen Brief…">|</textarea>
<div style="font-size:7px;color:var(--color-text-muted);margin-top:3px;">Wird gespeichert, wenn du das Feld verlässt.</div>
<div class="je-note-add" style="color:var(--color-text-muted);margin-top:2px;">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M2 2l6 6M8 2l-6 6" stroke="currentColor" stroke-width="1.3" stroke-linecap="round"/></svg>
Notiz entfernen
</div>
</div>
<div class="je-remove"><div class="je-remove-x">&#215;</div></div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LE-3 Inline-Notiz</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Toggleverhalten</td></tr>
<tr><td>Lokaler State</td><td>let noteOpen = item.note !== null and item.note !== ''</td><td>öffnet sich automatisch wenn Notiz bereits vorhanden</td></tr>
<tr><td>„Notiz hinzufügen" Klick</td><td>noteOpen = true; tick().then(() => noteTextarea.focus())</td><td>Fokus nach Svelte-Tick um DOM-Update abzuwarten</td></tr>
<tr><td>Textarea blur-Handler</td><td>on:blur={() => saveNote(item.id, note)}</td><td>PATCH /api/geschichten/{id}/items/{itemId} mit {note}</td></tr>
<tr><td>Leere Notiz on blur</td><td>wenn note.trim() === '' → noteOpen = false; note = null</td><td>verhindert leere Notizen im Backend</td></tr>
<tr class="grp"><td colspan="3">Fokus-Styling</td></tr>
<tr><td>Fokus-Ring</td><td>focus:border-primary focus:ring-2 focus:ring-primary/20 focus:bg-white</td><td>sichtbarer Ring für Keyboard-Navigation; ring-offset für Abstand</td></tr>
<tr><td>Spar-Hint</td><td>text-[9px] text-ink-3 mt-1</td><td>„Wird gespeichert, wenn du das Feld verlässt."; verschwindet wenn noteOpen = false</td></tr>
<tr class="grp"><td colspan="3">Barrierefreiheit</td></tr>
<tr><td>aria-label Textarea</td><td>aria-label="Kuratoren-Notiz für {document.title}"</td><td>spezifisch; Screen-Reader nennt Brief-Kontext</td></tr>
<tr><td>aria-expanded Toggle</td><td>aria-expanded={noteOpen} auf „Notiz hinzufügen"-Button</td><td>kommuniziert Expand-State</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ SCREEN LE-4: MOBILE ═══ -->
<div class="section">
<div class="section-title">Screens — Mobile Editor</div>
<div class="scr">
<div class="scr-head">
<h3>LE-4 — Mobile Journey-Editor</h3>
<span class="scr-id">Issue #753 · LE-4</span>
</div>
<p class="scr-desc">Auf Mobile (320px) entfällt die Sidebar-Split. Die Personen- und Status-Sektion werden als ausklappbare Sektionen unter der Itemliste gezeigt. Drag-to-Reorder ist auf Mobile durch Long-Press aktiviert. Die Aktionsleiste scrollt mit dem Inhalt.</p>
<p class="scr-var"><strong>Primäre Zielgruppe für den Editor: Desktop/Tablet. Mobile ist sekundär — alle Funktionen erreichbar, aber Drag ist schwerer bedienbar.</strong></p>
<div class="previews">
<div class="prev-col">
<span class="bp-lbl">Mobile — 320px · mit Einträgen</span>
<div class="phone" style="min-height:580px;">
<div class="pst"><b>9:41</b><span>&#9679;&#9679;&#9679;</span></div>
<div class="pb">
<div class="m-nav">
<span class="m-logo">ARCHIV</span>
<div class="m-nav-r">
<div class="m-av">KR</div>
<div class="m-ham"><span></span><span></span><span></span></div>
</div>
</div>
<div class="mob-topbar">
<span class="mob-back">&#8592;</span>
<span class="mob-label">Lesereise bearbeiten</span>
<div class="ed-status-pill ed-status-pub" style="font-size:7px;padding:1px 5px;">VER&Ouml;FF.</div>
</div>
<div class="mob-body">
<input class="mob-title-input" type="text" value="Briefe aus Breslau 19381942" readonly/>
<!-- Item 1: Document -->
<div class="mob-je-item">
<div class="mob-je-drag">
<div class="je-drag-dots">
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
<div class="je-drag-dot"></div><div class="je-drag-dot"></div>
</div>
</div>
<div class="mob-je-body">
<div class="mob-je-title">Brief vom 12. Juli 1938</div>
<div class="mob-je-meta">12. Juli 1938 &middot; Franz → Emma</div>
</div>
<div style="padding:6px 6px 0 0;font-size:10px;color:#C4C3BC;">&#215;</div>
</div>
<!-- Item 2: Interlude -->
<div class="mob-je-item mob-je-interlude">
<div class="mob-je-drag" style="background:rgba(232,134,42,.08);border-right-color:#F0C99A;"></div>
<div class="mob-je-body">
<div style="font-size:6.5px;font-weight:700;letter-spacing:.06em;text-transform:uppercase;color:var(--orange-dark);margin-bottom:3px;">Zwischentext</div>
<div class="mob-je-interlude-text">Im Sommer 1938 schrieb Franz voller Zuversicht…</div>
</div>
<div style="padding:6px 6px 0 0;font-size:10px;color:#D4A574;">&#215;</div>
</div>
<!-- Item 3: Document with note -->
<div class="mob-je-item">
<div class="mob-je-drag"></div>
<div class="mob-je-body">
<div class="mob-je-title">Postkarte Aug. 1938</div>
<div class="mob-je-meta">22. Aug. 1938 &middot; Franz → Emma</div>
<div class="mob-je-note">Diese Karte ist ungewöhnlich kurz für Franz…</div>
</div>
<div style="padding:6px 6px 0 0;font-size:10px;color:#C4C3BC;">&#215;</div>
</div>
<!-- Add bar -->
<div style="display:flex;gap:5px;padding:4px 0;">
<button class="je-add-btn" style="flex:1;font-size:7.5px;padding:6px 8px;justify-content:center;">+ Brief</button>
<button class="je-add-btn" style="flex:1;font-size:7.5px;padding:6px 8px;justify-content:center;">+ Zwischentext</button>
</div>
<!-- Collapsible: Personen -->
<div class="mob-collapsible">
<div class="mob-coll-hdr">
Personen
<span class="mob-coll-chevron">&#8250;</span>
</div>
</div>
<!-- Collapsible: Status -->
<div class="mob-collapsible">
<div class="mob-coll-hdr">
Status &amp; Speichern
<span class="mob-coll-chevron">&#8250;</span>
</div>
</div>
</div>
<div class="mob-savebar">
<button class="mob-btn mob-btn-ghost" style="font-size:8px;flex:0 0 auto;padding:7px 10px;">Entwurf</button>
<button class="mob-btn mob-btn-primary">Speichern</button>
</div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LE-4 Mobile</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Layout-Anpassungen</td></tr>
<tr><td>Split entfällt</td><td>@media (max-width: 768px): flex-col; Sidebar-Sektionen als Collapsibles am Ende</td><td>gleich wie GeschichteEditor auf Mobile</td></tr>
<tr><td>Collapsibles</td><td>details/summary oder eigene boolean-Toggle; Personen + Status separat</td><td>geschlossen beim ersten Laden; Fokus öffnet</td></tr>
<tr class="grp"><td colspan="3">Touch &amp; Drag</td></tr>
<tr><td>Drag auf Mobile</td><td>Long-Press (500ms) auf dem Drag-Handle aktiviert Drag</td><td>dnd-kit unterstützt Touch nativ; kein separates Config nötig</td></tr>
<tr><td>Touch Target Items</td><td>min-h-[44px] für jede Item-Zeile</td><td>WCAG 2.2 AA; durch Padding gesichert</td></tr>
<tr><td>Add-Buttons</td><td>flex-1; volle verfügbare Breite geteilt</td><td>min-h-[44px] als Touch-Target</td></tr>
<tr class="grp"><td colspan="3">Savebar</td></tr>
<tr><td>Savebar Mobile</td><td>flex gap-2; „Zurück zu Entwurf" komprimiert zu „Entwurf"</td><td>Volltext passt nicht auf 320px</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ LLM IMPLEMENTATION GUIDE ═══ -->
<div class="llm">
<h2>Implementation Guide — Journey-Editor</h2>
<h3>Neue Komponente</h3>
<table>
<thead><tr><th>Datei</th><th>Typ</th><th>Beschreibung</th></tr></thead>
<tbody>
<tr><td><code>src/lib/geschichte/JourneyEditor.svelte</code></td><td>Svelte-Komponente</td><td>Hauptkomponente; Props: <code>geschichte: Geschichte</code></td></tr>
<tr><td><code>src/lib/geschichte/JourneyItemRow.svelte</code></td><td>Svelte-Komponente</td><td>Eine Zeile (Dokument oder Interlude); Props: <code>item: JourneyItem, position: number</code>, Events: <code>remove, noteChange</code></td></tr>
</tbody>
</table>
<h3>Edit-Page-Integration</h3>
<ul>
<li><code>GeschichteEditor.svelte</code> erhält ein neues Prop <code>type: GeschichteType</code>.</li>
<li>Wenn <code>type === 'JOURNEY'</code>: rendere <code>JourneyEditor</code> statt TipTap-Editor. Die Sidebar (Personen, Status, Savebar) bleibt identisch.</li>
<li>Die Savebar-Logik ist in der Edit-Page (<code>+page.svelte</code>) verankert — <code>JourneyEditor</code> gibt nur Änderungen nach oben (Svelte-Events oder bindable Props), die Seite hält den Save-State.</li>
</ul>
<h3>API-Calls</h3>
<table>
<thead><tr><th>Aktion</th><th>Endpoint</th><th>Body</th></tr></thead>
<tbody>
<tr><td>Brief hinzufügen</td><td><code>POST /api/geschichten/{id}/items</code></td><td><code>{documentId: UUID}</code></td></tr>
<tr><td>Zwischentext hinzufügen</td><td><code>POST /api/geschichten/{id}/items</code></td><td><code>{note: string}</code></td></tr>
<tr><td>Notiz speichern/bearbeiten</td><td><code>PATCH /api/geschichten/{id}/items/{itemId}</code></td><td><code>{note: string | null}</code></td></tr>
<tr><td>Item entfernen</td><td><code>DELETE /api/geschichten/{id}/items/{itemId}</code></td><td></td></tr>
<tr><td>Reihenfolge speichern</td><td><code>PUT /api/geschichten/{id}/items/reorder</code></td><td><code>[{id: UUID, position: number}]</code></td></tr>
</tbody>
</table>
<h3>Optimistische Updates</h3>
<ul>
<li>Alle Mutationen (add, remove, reorder, noteChange) aktualisieren den lokalen State <em>sofort</em>, der API-Call läuft parallel.</li>
<li>Bei Fehler: lokalen State zurückrollen und einen <code>aria-live="polite"</code>-Fehlerhinweis anzeigen.</li>
<li>Notiz-Saving ist ein Sonderfall: es gibt kein optimistisches Update da der Wert bereits live im Textarea ist — nur blur → PATCH.</li>
</ul>
<h3>DocumentPicker-Integration</h3>
<ul>
<li>Der „Brief hinzufügen"-Button öffnet die bestehende <code>DocumentPicker</code>-Komponente (prüfe <code>$lib/document/</code> auf vorhandene Typeahead-Komponenten).</li>
<li>Nach Auswahl eines Dokuments: <code>POST /items</code> mit <code>documentId</code>, neues Item wird an das Ende der Liste angehängt und eingeblendet.</li>
<li>Bereits in der Journey enthaltene Dokumente: in der Picker-Ergebnisliste mit einem „Bereits enthalten"-Hinweis markieren und deaktivieren.</li>
</ul>
<h3>Drag-to-Reorder</h3>
<ul>
<li>Bibliothek: prüfe zunächst ob <code>@dnd-kit/core</code> oder <code>svelte-dnd-action</code> bereits im <code>package.json</code> ist. Kein neues Package einführen ohne Absprache.</li>
<li>Nach dem Drop: neue Reihenfolge als Array <code>[{id, position}]</code> berechnen (position = index * 10 lässt Lücken für künftige Inserts) und <code>PUT /items/reorder</code> senden.</li>
<li>Keyboard-Drag: Space/Enter startet, Arrow Up/Down verschiebt, Space/Enter bestätigt, Escape abbricht. Screenreader-Announcement: „Eintrag X von Position Y nach Z verschoben".</li>
</ul>
<h3>Barrierefreiheit</h3>
<ul>
<li>Items-Liste: <code>&lt;ol&gt;</code>-Element — kommuniziert die Ordnung an Screenreader.</li>
<li>Drag-Handle: <code>role="button"</code>, <code>tabindex="0"</code>, <code>aria-label="Reihenfolge von '{title}' ändern"</code>.</li>
<li>Entfernen-Button: <code>aria-label="'{title}' entfernen"</code>; kein reines ×-Zeichen ohne Label.</li>
<li>Notiz-Textarea: <code>aria-label="Kuratoren-Notiz für '{title}'"</code>.</li>
<li>Touch-Targets: alle interaktiven Elemente min 44×44px (WCAG 2.2 AA).</li>
<li>Fokusring: <code>focus-visible:ring-2 focus-visible:ring-primary</code> auf allen Buttons und Textareas.</li>
</ul>
<h3>Abgrenzung zu GeschichteEditor</h3>
<ul>
<li>TipTap wird für JOURNEY <em>nicht</em> geladen — kein unnötiger Bundle-Load.</li>
<li>Die Sidebar (Personen, Status) ist für beide Typen identisch — kein Duplikat, die Sidebar-Komponente wird geteilt.</li>
<li>Savebar-Logik (DRAFT/PUBLISHED/Retract) ist identisch — JourneyEditor ändert sie nicht.</li>
<li><code>Geschichte.body</code> dient für JOURNEY als Einleitungstext (Plaintext, kein HTML). Kein Rich-Text-Rendering auf der Leseseite nötig.</li>
</ul>
</div>
</div>
</body>
</html>

View File

@@ -0,0 +1,727 @@
<!DOCTYPE html>
<html lang="de">
<head>
<meta charset="UTF-8"/>
<meta name="viewport" content="width=device-width, initial-scale=1.0"/>
<title>Lesereisen — Reader-Integration · Familienarchiv</title>
<link href="https://fonts.googleapis.com/css2?family=Fraunces:opsz,wght@9..144,300;9..144,400;9..144,500&family=DM+Sans:wght@300;400;500;600&family=DM+Mono:wght@400;500&display=swap" rel="stylesheet"/>
<style>
:root{--color-page:#FAFAF7;--color-surface:#F5F4EE;--color-subtle:#EDECEA;--color-border:#D8D7D0;--color-text-muted:#6B6A63;--color-text:#1C1C18;--navy:#012851;--mint:#A1DCD8;--sand:#F0EFE9;--turquoise:#00C7B1;--blue-tint:#E6F1FB;--blue:#2D7DD2;--blue-dark:#185FA5;--green-tint:#E8F5EA;--green:#3D8C4A;--green-dark:#2E6E39;--orange-tint:#FEF0E6;--orange:#E8862A;--orange-dark:#B46820;--font-display:'Fraunces',Georgia,serif;--font-sans:'DM Sans',system-ui,sans-serif;--font-mono:'DM Mono',monospace;--radius-sm:4px;--radius-md:6px;--radius-lg:10px;--radius-xl:16px;--shadow-card:0 1px 3px rgba(28,28,24,.06),0 1px 2px rgba(28,28,24,.04);--shadow-raised:0 4px 12px rgba(28,28,24,.08),0 2px 4px rgba(28,28,24,.04);--shadow-overlay:0 8px 32px rgba(28,28,24,.12),0 2px 8px rgba(28,28,24,.06);}
*,*::before,*::after{box-sizing:border-box;margin:0;padding:0;}
body{font-family:var(--font-sans);background:#E8E7E2;color:var(--color-text);font-size:14px;line-height:1.6;}
.doc{max-width:1200px;margin:0 auto;padding:48px 40px 120px;}
.doc-header{display:flex;justify-content:space-between;align-items:flex-end;padding-bottom:28px;border-bottom:1px solid var(--color-border);margin-bottom:48px;background:var(--color-page);margin:-48px -40px 48px;padding:48px 40px 28px;border-radius:var(--radius-xl) var(--radius-xl) 0 0;}
.doc-header h1{font-family:var(--font-display);font-size:28px;font-weight:500;letter-spacing:-.02em;margin-bottom:4px;}
.doc-header p{font-size:13px;color:var(--color-text-muted);max-width:680px;}
.doc-meta{font-family:var(--font-mono);font-size:11px;color:var(--color-text-muted);text-align:right;line-height:1.9;}
.pill{display:inline-block;padding:2px 8px;border-radius:var(--radius-sm);font-size:10px;font-weight:500;letter-spacing:.05em;}
.pill-o{background:var(--orange-tint);color:var(--orange-dark);}
.section{margin-bottom:64px;}
.section-title{font-size:10px;font-weight:500;letter-spacing:.12em;text-transform:uppercase;color:var(--color-text-muted);padding-bottom:10px;border-bottom:1px solid var(--color-border);margin-bottom:24px;}
.prose{font-size:13px;color:var(--color-text-muted);line-height:1.65;max-width:720px;margin-bottom:20px;}
.jh{padding:20px 24px;border-radius:var(--radius-xl);margin-bottom:40px;display:flex;align-items:center;gap:16px;}
.jh .jn{font-family:var(--font-display);font-size:48px;font-weight:300;line-height:1;opacity:.5;}
.jh h2{font-family:var(--font-display);font-size:22px;font-weight:500;letter-spacing:-.02em;margin-bottom:4px;}
.jh p{font-size:13px;line-height:1.5;}.jh .fl{font-family:var(--font-mono);font-size:11px;margin-top:6px;opacity:.7;}
.jh-o{background:var(--orange-tint);border:1px solid #F0C99A;}
.jh-o .jn{color:var(--orange);}
.jh-o p,.jh-o .fl{color:var(--orange-dark);}
.scr{margin-bottom:56px;}
.scr-head{display:flex;justify-content:space-between;align-items:center;margin-bottom:6px;}
.scr-head h3{font-family:var(--font-display);font-size:20px;font-weight:500;letter-spacing:-.02em;}
.scr-id{font-family:var(--font-mono);font-size:11px;color:var(--color-text-muted);padding:2px 8px;border:1px solid var(--color-border);border-radius:var(--radius-sm);background:var(--color-page);}
.scr-desc{font-size:12px;color:var(--color-text-muted);line-height:1.6;max-width:720px;margin-bottom:6px;}
.scr-var{font-size:11px;color:var(--color-text-muted);margin-bottom:20px;}.scr-var strong{color:var(--color-text);}
.previews{display:flex;gap:32px;flex-wrap:wrap;justify-content:center;align-items:flex-start;margin-bottom:20px;}
.prev-col{display:flex;flex-direction:column;align-items:center;gap:10px;}
.bp-lbl{font-family:var(--font-mono);font-size:10px;color:var(--color-text-muted);}
.desk{width:100%;max-width:1040px;background:var(--color-page);border-radius:var(--radius-xl);overflow:hidden;box-shadow:var(--shadow-overlay),0 0 0 1px rgba(0,0,0,.06);display:flex;flex-direction:column;}
.phone{width:320px;flex-shrink:0;background:var(--color-page);border-radius:36px;overflow:hidden;box-shadow:var(--shadow-overlay),0 0 0 1px rgba(0,0,0,.07);display:flex;flex-direction:column;border:6px solid #1C1C18;}
.pst{padding:10px 20px 0;display:flex;justify-content:space-between;align-items:center;font-size:12px;background:var(--color-page);}.pst b{font-weight:600;}.pst span{font-size:10px;}
.pb{flex:1;overflow-y:auto;display:flex;flex-direction:column;}
.fa-nav{height:32px;background:var(--navy);display:flex;align-items:center;padding:0 12px;gap:8px;flex-shrink:0;}
.fa-logo{font-size:7px;font-weight:900;color:#fff;letter-spacing:.8px;border-bottom:2px solid var(--mint);padding-bottom:1px;}
.fa-link{font-size:5.5px;color:rgba(255,255,255,.4);font-weight:700;text-transform:uppercase;letter-spacing:.05em;}
.fa-link.active{color:var(--mint);}
.fa-nav-r{margin-left:auto;display:flex;gap:5px;align-items:center;}
.fa-av{width:16px;height:16px;background:rgba(255,255,255,.1);border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:5px;font-weight:800;color:rgba(255,255,255,.5);}
.m-nav{height:26px;background:var(--navy);display:flex;align-items:center;padding:0 10px;gap:6px;flex-shrink:0;}
.m-logo{font-size:6px;font-weight:900;color:#fff;letter-spacing:.7px;border-bottom:1.5px solid var(--mint);padding-bottom:1px;}
.m-nav-r{margin-left:auto;display:flex;gap:4px;align-items:center;}
.m-av{width:14px;height:14px;background:rgba(255,255,255,.1);border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:4.5px;font-weight:800;color:rgba(255,255,255,.5);}
.m-ham{display:flex;flex-direction:column;gap:2px;width:12px;}
.m-ham span{height:1.5px;background:rgba(255,255,255,.6);border-radius:1px;}
/* ── impl-ref table ── */
.agent{background:var(--color-text);color:#E8E8E2;padding:24px;border-radius:var(--radius-lg);margin-top:20px;}
.agent h4{font-size:9px;font-weight:500;letter-spacing:.1em;text-transform:uppercase;color:#5A5A55;margin-bottom:12px;}
.at{width:100%;border-collapse:collapse;font-family:var(--font-mono);font-size:10px;}
.at thead tr{border-bottom:1px solid #2A2A26;}
.at th{text-align:left;padding:6px 10px;font-size:8px;font-weight:500;letter-spacing:.08em;text-transform:uppercase;color:#5A5A55;font-family:var(--font-sans);}
.at td{padding:5px 10px;border-bottom:1px solid #1E1E1A;vertical-align:top;line-height:1.5;}
.at tr:last-child td{border-bottom:none;}
.at td:first-child{color:#7A7A72;}
.at td:nth-child(2){color:#E8E8E2;font-weight:500;}
.at td:nth-child(3){color:#5A5A55;}
.at .grp td{padding-top:14px;font-family:var(--font-sans);font-size:8px;font-weight:500;letter-spacing:.08em;text-transform:uppercase;color:#3A3A36;}
/* ── LLM guide ── */
.llm{background:var(--color-page);border:2px solid var(--navy);border-radius:var(--radius-xl);padding:32px 40px;margin-top:64px;}
.llm h2{font-family:var(--font-display);font-size:22px;font-weight:500;letter-spacing:-.02em;margin-bottom:8px;color:var(--navy);}
.llm h3{font-size:14px;font-weight:600;margin:20px 0 8px;}
.llm h4{font-size:12px;font-weight:600;margin:14px 0 6px;color:var(--color-text-muted);}
.llm p,.llm li{font-size:13px;color:var(--color-text-muted);line-height:1.65;}
.llm ul,.llm ol{padding-left:20px;margin-bottom:12px;}
.llm li{margin-bottom:4px;}
.llm code{font-family:var(--font-mono);font-size:11px;background:var(--color-surface);padding:1px 5px;border-radius:3px;}
.llm table{width:100%;border-collapse:collapse;margin:12px 0;font-size:12px;}
.llm th,.llm td{text-align:left;padding:6px 10px;border-bottom:1px solid var(--color-border);}
.llm th{font-weight:500;color:var(--color-text);font-size:11px;text-transform:uppercase;letter-spacing:.05em;}
.llm td{color:var(--color-text-muted);}
/* ── List row (re-used from reader-journey spec) ── */
.g-list-card{background:#fff;border:1px solid #E4E2D7;border-radius:4px;box-shadow:var(--shadow-card);overflow:hidden;}
.g-row{display:flex;gap:0;border-bottom:1px solid #F0EFE9;}
.g-row:last-child{border-bottom:none;}
.g-meta{width:88px;flex-shrink:0;padding:10px 10px 10px 12px;display:flex;flex-direction:column;gap:3px;border-right:1px solid #F0EFE9;}
.g-content{padding:10px 14px 10px 12px;flex:1;min-width:0;}
.g-av{width:22px;height:22px;border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:7px;font-weight:800;color:#fff;flex-shrink:0;margin-bottom:3px;}
.av-navy{background:#012851;} .av-purple{background:#534AB7;} .av-teal{background:#0E9488;}
.g-author{font-size:7px;font-weight:700;color:#1C1C18;line-height:1.3;}
.g-date{font-size:6.5px;color:#6B6A63;}
.g-chip{display:inline-flex;align-items:center;gap:2px;padding:1px 5px;background:#F5F4EE;border:1px solid #D8D7D0;border-radius:10px;font-size:6px;font-weight:500;color:#1C1C18;margin-top:2px;max-width:76px;overflow:hidden;text-overflow:ellipsis;white-space:nowrap;}
.g-title{font-family:Georgia,serif;font-size:11px;color:#012851;line-height:1.4;margin-bottom:2px;}
.g-excerpt{font-size:7.5px;color:#6B6A63;line-height:1.55;display:-webkit-box;-webkit-line-clamp:2;-webkit-box-orient:vertical;overflow:hidden;}
.g-filters{display:flex;gap:5px;align-items:center;padding:8px 12px;background:var(--color-page);border-bottom:1px solid #EDECEA;flex-wrap:wrap;}
.g-pill{display:inline-flex;align-items:center;padding:2px 8px;border-radius:10px;font-size:6.5px;font-weight:700;border:1px solid #D8D7D0;color:#6B6A63;background:transparent;}
.g-pill.active{background:#012851;color:#fff;border-color:#012851;}
.g-page-hdr{display:flex;justify-content:space-between;align-items:center;padding:10px 14px 6px;}
.g-page-title{font-family:Georgia,serif;font-size:16px;font-weight:400;color:#012851;}
.g-new-btn{font-size:7px;font-weight:700;padding:4px 10px;border-radius:3px;background:#012851;color:#fff;border:none;display:flex;align-items:center;gap:3px;}
/* ── Journey badge in list ── */
.j-badge{display:inline-flex;align-items:center;padding:1px 5px;border-radius:3px;font-size:5.5px;font-weight:700;letter-spacing:.07em;text-transform:uppercase;background:var(--orange-tint);color:var(--orange-dark);border:1px solid #F0C99A;margin-top:2px;}
/* ── Type selector cards ── */
.type-selector{display:flex;gap:12px;justify-content:center;padding:20px 24px;flex:1;align-items:center;background:#E8E7E2;}
.type-selector-inner{max-width:520px;width:100%;}
.type-selector-q{font-family:Georgia,serif;font-size:12px;font-weight:400;color:#6B6A63;text-align:center;margin-bottom:14px;}
.type-cards{display:flex;gap:10px;}
.type-card{flex:1;border:1px solid #D8D7D0;border-radius:6px;padding:12px 14px;cursor:pointer;background:#fff;display:flex;flex-direction:column;gap:5px;}
.type-card.selected{border-color:var(--orange);background:var(--orange-tint);box-shadow:0 0 0 2px rgba(232,134,42,.15);}
.type-card-icon{font-size:16px;margin-bottom:2px;}
.type-card-title{font-family:Georgia,serif;font-size:11px;font-weight:400;color:var(--navy);}
.type-card-desc{font-size:7.5px;color:#6B6A63;line-height:1.55;}
.type-card-check{width:14px;height:14px;border-radius:50%;background:var(--orange);display:flex;align-items:center;justify-content:center;margin-top:4px;align-self:flex-end;}
.type-card-check svg{width:8px;height:8px;}
.type-next-bar{display:flex;justify-content:flex-end;padding:8px 24px;background:#fff;border-top:1px solid #E4E2D7;}
.type-next-btn{font-size:8px;font-weight:700;padding:5px 14px;border-radius:3px;background:var(--navy);color:#fff;border:none;display:flex;align-items:center;gap:3px;}
/* ── Journey reader ── */
.jr-article{background:var(--color-page);border-radius:6px;padding:16px 20px;max-width:640px;margin:0 auto;}
.jr-back{font-size:7px;color:#6B6A63;margin-bottom:10px;display:flex;align-items:center;gap:2px;}
.jr-badge{display:inline-flex;align-items:center;padding:1px 6px;border-radius:3px;font-size:6px;font-weight:700;letter-spacing:.08em;text-transform:uppercase;background:var(--orange-tint);color:var(--orange-dark);border:1px solid #F0C99A;margin-bottom:5px;}
.jr-title{font-family:Georgia,serif;font-size:18px;font-weight:400;color:#012851;line-height:1.3;margin-bottom:8px;}
.jr-metabar{display:flex;align-items:center;gap:6px;padding-bottom:8px;border-bottom:1px solid #EDECEA;margin-bottom:10px;}
.jr-metabar-r{margin-left:auto;display:flex;align-items:center;gap:6px;}
.jr-edit-btn{font-size:6.5px;font-weight:600;padding:2px 7px;border:1px solid #D8D7D0;border-radius:3px;color:#1C1C18;background:transparent;}
.jr-intro{font-family:Georgia,serif;font-size:8.5px;line-height:1.75;color:#6B6A63;font-style:italic;margin-bottom:12px;padding-bottom:10px;border-bottom:1px dashed #EDECEA;}
/* Journey items in reader */
.jr-item{display:flex;gap:7px;margin-bottom:9px;align-items:flex-start;}
.jr-num{width:18px;height:18px;border-radius:50%;background:#012851;color:#fff;display:flex;align-items:center;justify-content:center;font-size:7px;font-weight:700;flex-shrink:0;margin-top:1px;}
.jr-card{flex:1;background:#fff;border:1px solid #E4E2D7;border-radius:4px;padding:7px 9px;}
.jr-card-title{font-family:Georgia,serif;font-size:9px;color:#012851;line-height:1.3;margin-bottom:2px;font-weight:400;}
.jr-card-meta{font-size:6.5px;color:#6B6A63;margin-bottom:5px;}
.jr-card-link{font-size:7px;font-weight:600;color:#012851;display:flex;align-items:center;gap:2px;}
.jr-annotation{margin-top:6px;padding:5px 7px;border-left:2px solid var(--mint);background:#F5F4EE;border-radius:0 3px 3px 0;}
.jr-annotation-text{font-size:7.5px;font-style:italic;color:#6B6A63;line-height:1.55;}
.jr-interlude{margin:10px 0 10px 25px;padding:7px 9px;border-left:2px solid var(--orange);background:var(--orange-tint);border-radius:0 4px 4px 0;}
.jr-interlude-text{font-size:8px;font-style:italic;color:#1C1C18;line-height:1.65;}
/* Mobile list row */
.m-row{padding:9px 10px;border-bottom:1px solid #F0EFE9;background:#fff;}
.m-row-top{display:flex;align-items:center;gap:5px;margin-bottom:3px;}
.m-author-name{font-size:7px;font-weight:700;color:#1C1C18;}
.m-date{font-size:6.5px;color:#6B6A63;margin-left:auto;}
.m-title{font-family:Georgia,serif;font-size:10px;color:#012851;line-height:1.4;margin-bottom:2px;}
.m-excerpt{font-size:7px;color:#6B6A63;line-height:1.5;display:-webkit-box;-webkit-line-clamp:2;-webkit-box-orient:vertical;overflow:hidden;}
.m-filters{display:flex;gap:4px;padding:6px 10px;background:var(--color-page);border-bottom:1px solid #EDECEA;overflow-x:auto;flex-wrap:nowrap;}
.m-filters::-webkit-scrollbar{display:none;}
/* Mobile journey reader */
.mjr-article{background:#fff;border-radius:6px;padding:12px 12px 16px;}
.mjr-back{font-size:7px;color:#6B6A63;margin-bottom:7px;display:flex;align-items:center;gap:2px;}
.mjr-badge{display:inline-flex;padding:1px 5px;border-radius:3px;font-size:5.5px;font-weight:700;letter-spacing:.07em;text-transform:uppercase;background:var(--orange-tint);color:var(--orange-dark);border:1px solid #F0C99A;margin-bottom:4px;}
.mjr-title{font-family:Georgia,serif;font-size:14px;font-weight:400;color:#012851;line-height:1.3;margin-bottom:6px;}
.mjr-metabar{display:flex;align-items:center;gap:5px;padding-bottom:6px;border-bottom:1px solid #EDECEA;margin-bottom:8px;}
.mjr-intro{font-family:Georgia,serif;font-size:8px;line-height:1.7;color:#6B6A63;font-style:italic;margin-bottom:9px;padding-bottom:7px;border-bottom:1px dashed #EDECEA;}
.mjr-item{display:flex;gap:5px;margin-bottom:7px;align-items:flex-start;}
.mjr-num{width:14px;height:14px;border-radius:50%;background:#012851;color:#fff;display:flex;align-items:center;justify-content:center;font-size:6px;font-weight:700;flex-shrink:0;margin-top:1px;}
.mjr-card{flex:1;background:#F5F4EE;border:1px solid #E4E2D7;border-radius:4px;padding:5px 7px;}
.mjr-card-title{font-family:Georgia,serif;font-size:8.5px;color:#012851;line-height:1.3;margin-bottom:1px;}
.mjr-card-meta{font-size:6px;color:#6B6A63;margin-bottom:4px;}
.mjr-card-link{font-size:6.5px;font-weight:600;color:#012851;}
.mjr-interlude{margin:7px 0 7px 19px;padding:5px 7px;border-left:2px solid var(--orange);background:var(--orange-tint);border-radius:0 3px 3px 0;}
.mjr-interlude-text{font-size:7.5px;font-style:italic;color:#1C1C18;line-height:1.6;}
/* ── Editor topbar (type selector screen) ── */
.ed-topbar{background:#fff;border-bottom:1px solid #e4e2d7;display:flex;align-items:center;padding:0 14px;gap:8px;height:38px;flex-shrink:0;}
.ed-back{width:22px;height:22px;border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:9px;color:var(--color-text-muted);flex-shrink:0;}
.ed-title-label{font-family:var(--font-sans);font-size:10px;font-weight:500;color:var(--color-text);flex:1;}
.ed-status-pill{display:inline-flex;align-items:center;padding:2px 7px;border-radius:20px;font-size:8px;font-weight:700;letter-spacing:.06em;text-transform:uppercase;flex-shrink:0;}
.ed-status-draft{background:#F0EFE9;color:#6B6A63;border:1px solid #D8D7D0;}
@media(max-width:900px){.doc{padding:24px 16px 80px;}}
</style>
</head>
<body>
<div class="doc">
<!-- ═══ DOC HEADER ═══ -->
<div class="doc-header">
<div>
<h1>Lesereisen — Reader-Integration</h1>
<p>Typauswahl bei <code>/geschichten/new</code>, Journey-Badge auf der Übersichtsliste und die neue geordnete Leseansicht auf <code>/geschichten/[id]</code> wenn <code>type === 'JOURNEY'</code>. Bestehende Story-Ansichten bleiben unverändert.</p>
</div>
<div class="doc-meta">
Familienarchiv<br/>
<span class="pill pill-o">Final Spec</span><br/>
2026-06-07 &middot; @leonievoss<br/>
<span style="font-size:10px;margin-top:4px;display:inline-block;">Issue #752</span>
</div>
</div>
<!-- ═══ JOURNEY HEADER ═══ -->
<div class="jh jh-o">
<div class="jn">R</div>
<div>
<h2>Lesereisen — Reader</h2>
<p>Alle angemeldeten Familienmitglieder können Lesereisen entdecken und in Briefsequenzen mit Kuratoren-Notizen eintauchen. BLOG_WRITERs sehen zusätzlich Bearbeiten/Löschen-Aktionen.</p>
<div class="fl">/geschichten &middot; /geschichten/new &middot; /geschichten/[id]</div>
</div>
</div>
<!-- ═══ KONZEPT ═══ -->
<div class="section">
<div class="section-title">Konzept</div>
<p class="prose">Eine <em>Lesereise</em> ist eine <code>Geschichte</code> mit <code>type === 'JOURNEY'</code>. Ihr Kerninhalt ist eine geordnete Sequenz von Briefen (<code>JourneyItem</code>s mit <code>document_id</code>) und Zwischentexten (<code>JourneyItem</code>s ohne <code>document_id</code>). Das optionale Feld <code>body</code> dient als Einleitung/Preface.</p>
<p class="prose">Diese Spec deckt drei Änderungen ab: (1) die Typauswahl auf <code>/geschichten/new</code> als vorgelagerter Schritt, (2) das „REISE"-Badge in der Übersichtsliste, und (3) die neue Journey-Leseansicht auf der Detailseite, die den bestehenden Prosa-Body durch eine nummerierte Briefliste ersetzt.</p>
<p class="prose">Dokument-Items zeigen Titel, Datum, Sender→Empfänger und einen Link zum Brief. Optionale Kuratoren-Notizen erscheinen als Annotation mit Mint-Linker-Rand unter dem Briefeintrag. Interlude-Items (kein Dokument) erscheinen als eingerückte Absätze mit orangenem linken Rand — klar vom Dokumenttyp unterscheidbar, aber harmonisch im Lesefluss.</p>
</div>
<!-- ═══ SCREEN LR-0: TYPE SELECTOR ═══ -->
<div class="section">
<div class="section-title">Screens — Typauswahl</div>
<div class="scr">
<div class="scr-head">
<h3>LR-0 — Typauswahl /geschichten/new</h3>
<span class="scr-id">Issue #752 · LR-0</span>
</div>
<p class="scr-desc">Neuer vorgelagerter Schritt beim Erstellen einer Geschichte. Zwei Karten zur Auswahl: „Geschichte" (Prosa) und „Lesereise" (Briefsequenz). Die ausgewählte Karte wird hervorgehoben. Erst nach Auswahl wird der „Weiter"-Button aktiv. Auswahl bleibt im URL-Param erhalten (<code>?type=JOURNEY</code>).</p>
<p class="scr-var"><strong>Varianten:</strong> Keine Auswahl (Weiter-Button inaktiv) · Lesereise gewählt (hier gezeigt) · Geschichte gewählt</p>
<div class="previews">
<div class="prev-col" style="width:100%;max-width:1040px;">
<span class="bp-lbl">Desktop — 1040px · Lesereise gewählt</span>
<div class="desk" style="min-height:320px;">
<div class="fa-nav">
<span class="fa-logo">ARCHIV</span>
<span style="width:1px;height:14px;background:rgba(255,255,255,.1);margin:0 2px;"></span>
<span class="fa-link">Dokumente</span>
<span class="fa-link">Personen</span>
<span class="fa-link active">Geschichten</span>
<span class="fa-link">Chronik</span>
<div class="fa-nav-r">
<div class="fa-av" style="background:#012851;color:var(--mint);font-size:5px;font-weight:800;">MR</div>
</div>
</div>
<div class="ed-topbar">
<div class="ed-back">&#8592;</div>
<div class="ed-title-label">Neue Geschichte</div>
<div class="ed-status-pill ed-status-draft">ENTWURF</div>
</div>
<div class="type-selector">
<div class="type-selector-inner">
<div class="type-selector-q">Was möchtest du erstellen?</div>
<div class="type-cards">
<!-- Story card -->
<div class="type-card">
<div class="type-card-icon">✍️</div>
<div class="type-card-title">Geschichte</div>
<div class="type-card-desc">Freier Prosatext über Familienerlebnisse, Erinnerungen oder historische Einordnungen — mit verlinkten Personen und Dokumenten.</div>
</div>
<!-- Journey card (selected) -->
<div class="type-card selected">
<div class="type-card-icon">📜</div>
<div class="type-card-title">Lesereise</div>
<div class="type-card-desc">Geordnete Briefsequenz mit optionalen Kuratoren-Notizen zwischen den Briefen — für chronologische Korrespondenz-Sammlungen.</div>
<div class="type-card-check">
<svg viewBox="0 0 10 10" fill="none"><path d="M2 5l2.5 2.5L8 3" stroke="#fff" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
</div>
</div>
</div>
</div>
</div>
<div class="type-next-bar">
<button class="type-next-btn">
Weiter
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M4 2l4 3-4 3" stroke="#fff" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
</button>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LR-0 Typauswahl</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Layout</td></tr>
<tr><td>Selector area</td><td>flex flex-1 items-center justify-center bg-canvas px-6 py-10</td><td>zentriert, füllt restliche Höhe</td></tr>
<tr><td>Frage</td><td>font-serif text-sm text-ink-2 text-center mb-4</td><td></td></tr>
<tr><td>Karten-Grid</td><td>flex gap-4</td><td>2 gleich breite Karten; auf Mobile flex-col</td></tr>
<tr class="grp"><td colspan="3">Type-Karte</td></tr>
<tr><td>Karte (inaktiv)</td><td>border border-line rounded-md p-4 bg-white cursor-pointer hover:border-primary hover:bg-surface</td><td>focus-visible:ring-2 focus-visible:ring-primary</td></tr>
<tr><td>Karte (ausgewählt)</td><td>border-2 border-orange-500 bg-orange-50 shadow-sm</td><td>aria-pressed="true"; kein Tailwind-Kürzel — nutze CSS-var(--orange)</td></tr>
<tr><td>Check-Kreis</td><td>w-5 h-5 rounded-full bg-orange-500 flex items-center justify-center self-end mt-2</td><td>nur sichtbar wenn ausgewählt</td></tr>
<tr><td>Kartentitel</td><td>font-serif text-sm text-ink</td><td></td></tr>
<tr><td>Kartenbeschreibung</td><td>text-xs text-ink-3 leading-relaxed mt-1</td><td></td></tr>
<tr class="grp"><td colspan="3">Navigation</td></tr>
<tr><td>Weiter-Button</td><td>rounded border border-primary bg-primary text-white px-4 py-2 text-sm font-medium disabled:opacity-40</td><td>disabled wenn keine Karte ausgewählt</td></tr>
<tr><td>URL-Param</td><td>?type=STORY | ?type=JOURNEY</td><td>per goto() nach Klick auf Weiter; lesefreundlich bookmarkbar</td></tr>
<tr><td>Mobile</td><td>flex-col Karten; volle Breite</td><td>kein Scrollbedarf auf 320px</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ SCREEN LR-1: LIST WITH BADGE ═══ -->
<div class="section">
<div class="section-title">Screens — Übersichtsliste</div>
<div class="scr">
<div class="scr-head">
<h3>LR-1 — Reise-Badge in /geschichten</h3>
<span class="scr-id">Issue #752 · LR-1</span>
</div>
<p class="scr-desc">Die Übersichtsliste erhält ein kleines „REISE"-Badge in der Metaspalte einer Journey-Zeile — unterhalb von Datum und Personenchip. Zeilen mit <code>type === 'STORY'</code> bleiben unverändert. Das Badge ist nicht klickbar, dient als reine visuelle Unterscheidung.</p>
<p class="scr-var"><strong>Varianten:</strong> Mischte Liste (hier gezeigt) · Nur-Journey-Filter · Nur-Story-Ansicht (unverändert)</p>
<div class="previews">
<!-- Desktop -->
<div class="prev-col" style="width:100%;max-width:1040px;">
<span class="bp-lbl">Desktop — 1040px · gemischte Liste</span>
<div class="desk">
<div class="fa-nav">
<span class="fa-logo">ARCHIV</span>
<span style="width:1px;height:14px;background:rgba(255,255,255,.1);margin:0 2px;"></span>
<span class="fa-link">Dokumente</span>
<span class="fa-link">Personen</span>
<span class="fa-link active">Geschichten</span>
<span class="fa-link">Chronik</span>
<div class="fa-nav-r">
<div class="fa-av" style="background:#012851;color:var(--mint);font-size:5px;font-weight:800;">MR</div>
</div>
</div>
<div style="background:#E8E7E2;flex:1;padding:14px 16px;">
<div class="g-page-hdr" style="padding:0 0 8px;">
<span class="g-page-title">Geschichten</span>
<button class="g-new-btn">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M5 1v8M1 5h8" stroke="#fff" stroke-width="2" stroke-linecap="round"/></svg>
Neue Geschichte
</button>
</div>
<div class="g-list-card">
<div class="g-filters">
<span class="g-pill active">Alle</span>
<span class="g-pill">Franz Raddatz</span>
<span class="g-pill">Emma Müller</span>
<span class="g-pill" style="border-style:dashed;color:#6B6A63;">+ Person wählen</span>
</div>
<!-- Row 1: Story (no badge) -->
<div class="g-row">
<div class="g-meta">
<div class="g-av av-navy">MR</div>
<div class="g-author">Maria Raddatz</div>
<div class="g-date">14. März 2025</div>
<span class="g-chip">
<span style="width:8px;height:8px;border-radius:50%;background:#012851;display:inline-flex;align-items:center;justify-content:center;font-size:4.5px;font-weight:800;color:var(--mint);flex-shrink:0;">FR</span>
Franz Raddatz
</span>
</div>
<div class="g-content">
<div class="g-title">Der Sommer in Breslau</div>
<div class="g-excerpt">Oma erzählte oft vom letzten Sommer vor dem Krieg, als die Familie noch vollständig zusammen war und niemand ahnte, was kommen würde…</div>
</div>
</div>
<!-- Row 2: Journey (badge!) -->
<div class="g-row">
<div class="g-meta">
<div class="g-av av-purple">KR</div>
<div class="g-author">Klaus Raddatz</div>
<div class="g-date">15. Mai 2025</div>
<span class="g-chip">
<span style="width:8px;height:8px;border-radius:50%;background:#012851;display:inline-flex;align-items:center;justify-content:center;font-size:4.5px;font-weight:800;color:var(--mint);flex-shrink:0;">FR</span>
Franz Raddatz
</span>
<span class="j-badge">REISE</span>
</div>
<div class="g-content">
<div class="g-title">Briefe aus Breslau 19381942</div>
<div class="g-excerpt">Eine Lesereise durch den Briefwechsel zwischen Franz und Emma — von den letzten Friedenssommern bis zum Ende des Krieges.</div>
</div>
</div>
<!-- Row 3: Story -->
<div class="g-row">
<div class="g-meta">
<div class="g-av av-teal">GK</div>
<div class="g-author">Gertrud Koch</div>
<div class="g-date">18. Okt. 2024</div>
<span class="g-chip">
<span style="width:8px;height:8px;border-radius:50%;background:#534AB7;display:inline-flex;align-items:center;justify-content:center;font-size:4.5px;font-weight:800;color:#fff;flex-shrink:0;">EM</span>
Emma Müller
</span>
</div>
<div class="g-content">
<div class="g-title">Die Hochzeit im Krieg</div>
<div class="g-excerpt">1943, mitten im Chaos — Emma bestand darauf, dass das Fest stattfand. Ihr Bruder kam auf Fronturlaub, drei Tage nur, aber es reichte…</div>
</div>
</div>
</div>
</div>
</div>
</div>
<!-- Mobile -->
<div class="prev-col">
<span class="bp-lbl">Mobile — 320px</span>
<div class="phone">
<div class="pst"><b>9:41</b><span>&#9679;&#9679;&#9679;</span></div>
<div class="pb">
<div class="m-nav">
<span class="m-logo">ARCHIV</span>
<div class="m-nav-r">
<div class="m-av">MR</div>
<div class="m-ham"><span></span><span></span><span></span></div>
</div>
</div>
<div style="background:#E8E7E2;flex:1;display:flex;flex-direction:column;">
<div style="padding:8px 10px 4px;">
<span style="font-family:Georgia,serif;font-size:13px;color:#012851;">Geschichten</span>
</div>
<div class="m-filters">
<span class="g-pill active" style="font-size:6px;padding:2px 7px;">Alle</span>
<span class="g-pill" style="font-size:6px;padding:2px 7px;">Franz Raddatz</span>
<span class="g-pill" style="font-size:6px;padding:2px 7px;border-style:dashed;">+ Person…</span>
</div>
<div style="background:#fff;flex:1;">
<!-- Story row -->
<div class="m-row">
<div class="m-row-top">
<div class="g-av av-navy" style="width:16px;height:16px;font-size:5.5px;">MR</div>
<span class="m-author-name">Maria Raddatz</span>
<span class="m-date">14. Mrz. 2025</span>
</div>
<div class="m-title">Der Sommer in Breslau</div>
<div class="m-excerpt">Oma erzählte oft vom letzten Sommer vor dem Krieg…</div>
</div>
<!-- Journey row (badge) -->
<div class="m-row">
<div class="m-row-top">
<div class="g-av av-purple" style="width:16px;height:16px;font-size:5.5px;">KR</div>
<span class="m-author-name">Klaus Raddatz</span>
<span class="m-date">15. Mai 2025</span>
</div>
<div style="display:flex;align-items:center;gap:5px;margin-bottom:3px;">
<div class="m-title" style="margin-bottom:0;">Briefe aus Breslau 19381942</div>
<span class="j-badge" style="flex-shrink:0;">REISE</span>
</div>
<div class="m-excerpt">Eine Lesereise durch den Briefwechsel zwischen Franz und Emma…</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LR-1 Journey-Badge in der Liste</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Badge</td></tr>
<tr><td>Journey badge</td><td>inline-flex items-center px-1.5 py-px rounded-sm text-[10px] font-bold uppercase tracking-wide bg-orange-50 text-orange-700 border border-orange-200</td><td>nur wenn type === 'JOURNEY'</td></tr>
<tr><td>Position Desktop</td><td>unterhalb Datum-Text und Personenchip in der Metaspalte (g-meta)</td><td>kein extra Abstand nötig — gap-1 der Flex-Spalte reicht</td></tr>
<tr><td>Position Mobile</td><td>inline flex items-center gap-1.5 neben Titel</td><td>Titel + Badge in einem flex-Wrapper; badge shrink-0</td></tr>
<tr><td>aria-label</td><td>aria-label="Lesereise"</td><td>Badge ist span, kein interaktives Element</td></tr>
<tr class="grp"><td colspan="3">Bedingte Logik</td></tr>
<tr><td>Svelte guard</td><td>{#if geschichte.type === 'JOURNEY'}&lt;span …&gt;REISE&lt;/span&gt;{/if}</td><td>kein Badge für STORY</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ SCREEN LR-2: JOURNEY READER ═══ -->
<div class="section">
<div class="section-title">Screens — Journey-Leseansicht</div>
<div class="scr">
<div class="scr-head">
<h3>LR-2 — Journey-Detail /geschichten/[id]</h3>
<span class="scr-id">Issue #752 · LR-2</span>
</div>
<p class="scr-desc">Wenn <code>type === 'JOURNEY'</code> ersetzt die geordnete Briefliste den Prosa-Body. Optional zeigt ein Einleitungsabsatz (<code>body</code>) vor den Items. Jedes Item ist entweder ein Briefeintrag (Kartentitel, Datum, Link) oder ein Interlude-Absatz (orangener linker Rand, kursiv). Die Reihenfolge ergibt sich von oben nach unten — keine Nummern. Briefeinträge können eine optionale Kuratoren-Annotation unter dem Link zeigen.</p>
<p class="scr-var"><strong>Varianten:</strong> Leserin ohne Schreibrecht · BLOG_WRITER (Bearbeiten/Löschen sichtbar — hier gezeigt) · Mobile</p>
<div class="previews">
<!-- Desktop -->
<div class="prev-col" style="width:100%;max-width:1040px;">
<span class="bp-lbl">Desktop — 1040px · BLOG_WRITER-Ansicht</span>
<div class="desk" style="min-height:600px;">
<div class="fa-nav">
<span class="fa-logo">ARCHIV</span>
<span style="width:1px;height:14px;background:rgba(255,255,255,.1);margin:0 2px;"></span>
<span class="fa-link">Dokumente</span>
<span class="fa-link">Personen</span>
<span class="fa-link active">Geschichten</span>
<span class="fa-link">Chronik</span>
<div class="fa-nav-r">
<div class="fa-av" style="background:#012851;color:var(--mint);font-size:5px;font-weight:800;">MR</div>
</div>
</div>
<div style="background:#E8E7E2;flex:1;padding:16px 20px;">
<div class="jr-article">
<div class="jr-back">
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M6 2L2 5l4 3" stroke="#6B6A63" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
Zurück zu Geschichten
</div>
<div class="jr-badge">LESEREISE</div>
<div class="jr-title">Briefe aus Breslau 19381942</div>
<div class="jr-metabar">
<div class="g-av av-purple" style="width:20px;height:20px;font-size:6.5px;">KR</div>
<div>
<div style="font-size:7.5px;font-weight:700;color:#1C1C18;line-height:1.2;">Klaus Raddatz</div>
<div style="font-size:6.5px;color:#6B6A63;">zusammengestellt am 15. Mai 2025</div>
</div>
<div class="jr-metabar-r">
<button class="jr-edit-btn">Bearbeiten</button>
<span style="font-size:6.5px;font-weight:600;color:#DC4C3E;">Löschen</span>
</div>
</div>
<!-- Intro -->
<div class="jr-intro">Der Briefwechsel zwischen Franz Raddatz und seiner Schwester Emma umspannt vier Jahre — von den letzten unbeschwerten Sommerwochen 1938 bis zum Kriegsende. Diese Lesereise folgt den Briefen in chronologischer Reihenfolge.</div>
<!-- Item 1: Document, no annotation -->
<div class="jr-item">
<div class="jr-card">
<div class="jr-card-title">Brief vom 12. Juli 1938</div>
<div class="jr-card-meta">12. Juli 1938 &middot; von Franz Raddatz an Emma Müller</div>
<div class="jr-card-link">
<svg width="8" height="8" viewBox="0 0 10 12" fill="none"><rect x="1" y="1" width="8" height="10" rx="1" stroke="#012851" stroke-width="1"/><path d="M3 4h4M3 6.5h4M3 9h2" stroke="#012851" stroke-width=".7" stroke-linecap="round"/></svg>
Brief öffnen
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M4 2l4 3-4 3" stroke="#012851" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
</div>
</div>
</div>
<!-- Interlude -->
<div class="jr-interlude">
<div class="jr-interlude-text">Im Sommer 1938 schrieb Franz voller Zuversicht — er hatte kaum eine Ahnung, wie bald sich die Welt um ihn herum verändern würde. Seine Briefe aus dieser Zeit tragen eine Leichtigkeit, die in den späteren Kriegsjahren vollständig verschwindet.</div>
</div>
<!-- Item 2: Document with annotation -->
<div class="jr-item">
<div class="jr-card">
<div class="jr-card-title">Postkarte aus Breslau, August 1938</div>
<div class="jr-card-meta">22. Aug. 1938 &middot; von Franz Raddatz an Emma Müller</div>
<div class="jr-card-link">
<svg width="8" height="8" viewBox="0 0 10 12" fill="none"><rect x="1" y="1" width="8" height="10" rx="1" stroke="#012851" stroke-width="1"/><path d="M3 4h4M3 6.5h4M3 9h2" stroke="#012851" stroke-width=".7" stroke-linecap="round"/></svg>
Brief öffnen
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M4 2l4 3-4 3" stroke="#012851" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
</div>
<div class="jr-annotation">
<div class="jr-annotation-text">Diese Karte ist ungewöhnlich kurz für Franz — vier Zeilen, fast hastig. Ein Zeichen der aufkommenden Unruhe in den Nachrichten, oder schlicht die Hitze des Augusts?</div>
</div>
</div>
</div>
<!-- Item 3: Document -->
<div class="jr-item">
<div class="jr-card">
<div class="jr-card-title">Brief vom 3. September 1939</div>
<div class="jr-card-meta">3. Sept. 1939 &middot; von Emma Müller an Franz Raddatz</div>
<div class="jr-card-link">
<svg width="8" height="8" viewBox="0 0 10 12" fill="none"><rect x="1" y="1" width="8" height="10" rx="1" stroke="#012851" stroke-width="1"/><path d="M3 4h4M3 6.5h4M3 9h2" stroke="#012851" stroke-width=".7" stroke-linecap="round"/></svg>
Brief öffnen
<svg width="7" height="7" viewBox="0 0 10 10" fill="none"><path d="M4 2l4 3-4 3" stroke="#012851" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round"/></svg>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<!-- Mobile -->
<div class="prev-col">
<span class="bp-lbl">Mobile — 320px · Leserin</span>
<div class="phone" style="min-height:520px;">
<div class="pst"><b>9:41</b><span>&#9679;&#9679;&#9679;</span></div>
<div class="pb">
<div class="m-nav">
<span class="m-logo">ARCHIV</span>
<div class="m-nav-r">
<div class="m-av">MR</div>
<div class="m-ham"><span></span><span></span><span></span></div>
</div>
</div>
<div style="background:#E8E7E2;flex:1;padding:10px;">
<div class="mjr-article">
<div class="mjr-back">
<svg width="6" height="6" viewBox="0 0 10 10" fill="none"><path d="M6 2L2 5l4 3" stroke="#6B6A63" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"/></svg>
Zurück
</div>
<div class="mjr-badge">LESEREISE</div>
<div class="mjr-title">Briefe aus Breslau 19381942</div>
<div class="mjr-metabar">
<div class="g-av av-purple" style="width:16px;height:16px;font-size:5.5px;flex-shrink:0;">KR</div>
<div>
<div style="font-size:7px;font-weight:700;color:#1C1C18;">Klaus Raddatz</div>
<div style="font-size:6px;color:#6B6A63;">15. Mai 2025</div>
</div>
<div style="margin-left:auto;font-size:12px;color:#6B6A63;">···</div>
</div>
<div style="height:1px;background:#EDECEA;margin-bottom:8px;"></div>
<div class="mjr-intro">Der Briefwechsel zwischen Franz und Emma — von 1938 bis Kriegsende.</div>
<!-- Item 1 -->
<div class="mjr-item">
<div class="mjr-card">
<div class="mjr-card-title">Brief vom 12. Juli 1938</div>
<div class="mjr-card-meta">12. Juli 1938 · Franz → Emma</div>
<div class="mjr-card-link">Brief öffnen →</div>
</div>
</div>
<!-- Interlude -->
<div class="mjr-interlude">
<div class="mjr-interlude-text">Im Sommer 1938 schrieb Franz voller Zuversicht — er hatte kaum eine Ahnung, wie bald sich die Welt um ihn herum verändern würde.</div>
</div>
<!-- Item 2 -->
<div class="mjr-item">
<div class="mjr-card">
<div class="mjr-card-title">Postkarte Aug. 1938</div>
<div class="mjr-card-meta">22. Aug. 1938 · Franz → Emma</div>
<div class="mjr-card-link">Brief öffnen →</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div class="agent">
<h4>impl-ref — LR-2 Journey-Leseansicht</h4>
<table class="at">
<thead><tr><th>Element</th><th>Wert</th><th>Hinweise</th></tr></thead>
<tbody>
<tr class="grp"><td colspan="3">Seitenstruktur</td></tr>
<tr><td>Bedingte Logik</td><td>{#if geschichte.type === 'JOURNEY'} JourneyReader {:else} StoryReader {/if}</td><td>in +page.svelte von /geschichten/[id]</td></tr>
<tr><td>Artikel-Container</td><td>max-w-3xl mx-auto px-4 py-8</td><td>gleich wie StoryReader</td></tr>
<tr><td>Journey-Badge</td><td>inline-flex px-2 py-px rounded-sm text-[10px] font-bold uppercase tracking-widest bg-orange-50 text-orange-700 border border-orange-200 mb-2</td><td>über dem Titel; nicht für STORY</td></tr>
<tr><td>Titel</td><td>font-serif text-3xl text-ink leading-tight mb-4</td><td>gleich wie Story</td></tr>
<tr><td>Metabar</td><td>flex items-center gap-3 pb-4 border-b border-subtle mb-4</td><td>gleich wie Story</td></tr>
<tr><td>Bearbeiten/Löschen</td><td>nur BLOG_WRITE; auf Mobile im ··· BottomSheet</td><td>gleich wie Story</td></tr>
<tr class="grp"><td colspan="3">Intro-Absatz</td></tr>
<tr><td>Intro (body)</td><td>font-serif text-sm text-ink-2 italic leading-relaxed mb-6 pb-4 border-b border-dashed border-subtle</td><td>nur rendern wenn body nicht leer; kein HTML-Rendering — plaintext</td></tr>
<tr class="grp"><td colspan="3">Dokument-Item</td></tr>
<tr><td>Item-Zeile</td><td>mb-3</td><td>kein flex nötig — Karte ist full-width</td></tr>
<tr><td>Dokumentkarte</td><td>bg-white border border-line rounded-sm p-3</td><td></td></tr>
<tr><td>Brieftitel</td><td>font-serif text-sm text-ink leading-snug mb-0.5</td><td>document.title</td></tr>
<tr><td>Briefmeta</td><td>text-xs text-ink-3 mb-2</td><td>formatDate(document.documentDate) · "von X an Y"</td></tr>
<tr><td>Brief öffnen Link</td><td>inline-flex items-center gap-1 text-xs font-semibold text-ink hover:text-primary</td><td>href="/documents/{item.document.id}"</td></tr>
<tr class="grp"><td colspan="3">Kuratoren-Annotation</td></tr>
<tr><td>Annotation</td><td>mt-3 pl-3 border-l-2 border-mint bg-surface rounded-r-sm py-1.5 pr-2</td><td>nur rendern wenn item.note vorhanden</td></tr>
<tr><td>Annotations-Text</td><td>text-xs italic text-ink-2 leading-relaxed</td><td></td></tr>
<tr class="grp"><td colspan="3">Interlude-Item</td></tr>
<tr><td>Interlude-Block</td><td>pl-3 border-l-2 border-orange-400 bg-orange-50 rounded-r-sm py-2 pr-3 my-4</td><td>item.document === null</td></tr>
<tr><td>Interlude-Text</td><td>text-xs italic text-ink leading-relaxed</td><td>item.note; plaintext</td></tr>
<tr class="grp"><td colspan="3">Mobile</td></tr>
<tr><td>··· Menü</td><td>ml-auto text-ink-3; öffnet BottomSheet mit Bearbeiten + Löschen</td><td>BLOG_WRITE; gleich wie Story</td></tr>
<tr><td>Touch Target (Brief öffnen)</td><td>min-h-[44px] durch padding auf der Karte</td><td>WCAG 2.2 AA</td></tr>
</tbody>
</table>
</div>
</div>
</div>
<!-- ═══ LLM IMPLEMENTATION GUIDE ═══ -->
<div class="llm">
<h2>Implementation Guide — Lesereisen Reader</h2>
<h3>Geänderte Views und Routen</h3>
<table>
<thead><tr><th>View</th><th>Route</th><th>Änderung</th></tr></thead>
<tbody>
<tr><td>Neue Geschichte</td><td>/geschichten/new</td><td>Neuer Typauswahl-Schritt als first render; setzt ?type=STORY|JOURNEY</td></tr>
<tr><td>Geschichten-Liste</td><td>/geschichten</td><td>Journey-Badge in GeschichtenCard wenn type === 'JOURNEY'</td></tr>
<tr><td>Geschichte-Detail</td><td>/geschichten/[id]</td><td>Bedingte Verzweigung: JourneyReader | StoryReader</td></tr>
</tbody>
</table>
<h3>Neue Komponenten</h3>
<ul>
<li><code>JourneyReader.svelte</code> — rendert Intro + Items-Liste; Props: <code>geschichte: GeschichteDetail</code></li>
<li><code>JourneyItemCard.svelte</code> — ein Dokument-Item mit optionaler Annotation; Props: <code>item: JourneyItem, position: number</code></li>
<li><code>JourneyInterlude.svelte</code> — ein reiner Text-Interlude; Props: <code>note: string</code></li>
</ul>
<h3>Datenmodell (nach #750)</h3>
<ul>
<li><code>GeschichteType: 'STORY' | 'JOURNEY'</code></li>
<li><code>JourneyItem: { id: UUID, position: number, document: DocumentSummary | null, note: string | null }</code></li>
<li><code>Geschichte.items</code> — geordnete Liste (nach <code>position</code> ASC); für STORY leer</li>
<li><code>Geschichte.body</code> — für JOURNEY der optionale Einleitungstext (plaintext, kein HTML); für STORY der Rich-Text-Body</li>
</ul>
<h3>Typauswahl — Implementierungshinweise</h3>
<ul>
<li>Die Typauswahl ist ein Schritt INNERHALB der <code>/geschichten/new</code>-Route — kein eigener URL, kein <code>goto()</code>. Zustand <code>let selectedType: GeschichteType | null = null</code> in der Komponente.</li>
<li>Erst wenn <code>selectedType !== null</code> ist der „Weiter"-Button aktiviert (<code>disabled={!selectedType}</code>).</li>
<li>Nach Klick auf „Weiter": wenn <code>selectedType === 'JOURNEY'</code><code>goto('/geschichten/new?type=JOURNEY')</code> und zeige den Journey-Editor (aus Issue #753); wenn <code>STORY</code> → bestehender GeschichteEditor (unverändert).</li>
<li>Die Karten verwenden <code>role="radio"</code> und <code>aria-checked</code> für Accessibility. Keyboard: Arrow-Keys wechseln zwischen den Karten, Space/Enter wählt aus.</li>
</ul>
<h3>Journey-Badge — Implementierungshinweise</h3>
<ul>
<li>Badge nur in <code>GeschichtenCard.svelte</code> hinzufügen — keine Änderung an der Listenlogik oder dem API-Aufruf.</li>
<li>Text: „REISE" (Kurzform für die Metaspalte); <code>aria-label="Lesereise"</code> für den Badge-Span.</li>
</ul>
<h3>Journey-Reader — Implementierungshinweise</h3>
<ul>
<li>Items werden bereits geordnet vom Backend geliefert (<code>ORDER BY position ASC</code>). Keine client-seitige Sortierung nötig.</li>
<li>Ein Item ist Interlude wenn <code>item.document === null</code>. In diesem Fall: <code>JourneyInterlude</code>-Komponente rendern.</li>
<li>Der Intro-Absatz (<code>body</code>) wird als Plaintext gerendert — <em>nicht</em> als innerHTML. Im Editor wird es als einfaches Textarea gespeichert, kein HTML.</li>
</ul>
<h3>Berechtigungen</h3>
<ul>
<li>„Bearbeiten" und „Löschen" nur für <code>currentUser.permissions.includes('BLOG_WRITE')</code> — gleich wie Story.</li>
<li>Auf Mobile: Bearbeiten/Löschen im BottomSheet hinter ··· — gleich wie Story.</li>
</ul>
<h3>Barrierefreiheit</h3>
<ul>
<li>Items-Liste: <code>&lt;ol&gt;</code> semantisch für die geordnete Briefliste. Interludes sind <code>&lt;li&gt;</code>-Elemente mit <code>aria-label="Kuratorennotiz"</code>.</li>
<li>„Brief öffnen"-Link: beschreibender Text mit Briefdatum im <code>aria-label</code>, z.B. <code>aria-label="Brief vom 12. Juli 1938 öffnen"</code>.</li>
<li>Touch-Targets: jede Dokumentkarte hat mindestens 44px Höhe durch den Padding der Karte.</li>
<li>Fokusring: <code>focus-visible:ring-2 focus-visible:ring-primary</code> auf allen Links.</li>
</ul>
</div>
</div>
</body>
</html>

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,418 @@
<!DOCTYPE html>
<html lang="de">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Zeitstrahl — Ereignis-Editor &amp; Brief-Gruppierung · Quick-Action im Dokument · Familienarchiv</title>
<link href="https://fonts.googleapis.com/css2?family=Tinos:ital,wght@0,400;0,700;1,400&family=Montserrat:wght@400;500;600;700;800;900&display=swap" rel="stylesheet">
<style>
*,*::before,*::after{box-sizing:border-box;margin:0;padding:0}
body{font-family:'Montserrat',system-ui,sans-serif;background:#ECEAE4;color:#1A1A1A;line-height:1.5;font-size:13px}
.page{max-width:1320px;margin:0 auto;padding:48px 32px 120px}
.mh{padding-bottom:24px;border-bottom:3px solid #012851;margin-bottom:48px}
.mh h1{font-size:23px;font-weight:900;color:#012851;letter-spacing:-.4px}
.mh p{font-size:13px;color:#555;max-width:790px;line-height:1.75;margin-top:8px}
.mh .byline{font-size:9px;color:#999;font-weight:700;letter-spacing:1.5px;text-transform:uppercase;margin-top:10px}
.tag-row{display:flex;gap:6px;margin-top:12px;flex-wrap:wrap}
.tg{background:#012851;color:#a1dcd8;padding:2px 8px;border-radius:2px;font-size:8px;font-weight:700;letter-spacing:.8px;text-transform:uppercase}
.tg.mint{background:#a1dcd8;color:#012851}
.tg.slate{background:#607080;color:#e8edf2}
.sh{margin:60px 0 22px;padding-bottom:12px;border-bottom:2px solid #E0DDD6}
.sh h2{font-size:17px;font-weight:900;color:#012851}
.sh p{font-size:12.5px;color:#666;margin-top:5px;max-width:790px;line-height:1.65}
.callout{padding:13px 17px;border-radius:4px;font-size:12px;line-height:1.65;margin-bottom:18px}
.callout.navy{background:rgba(1,40,81,.06);border-left:3px solid #012851;color:#333}
.callout.mint{background:rgba(161,220,216,.18);border-left:3px solid #00c7b1;color:#1f3a3a}
.callout code{font-family:'Courier New',monospace;font-size:11px;background:#fff;padding:1px 4px;border-radius:2px}
.callout strong{font-weight:800;color:#012851}
/* desktop chrome */
.dchrome{background:#F0EFE9;border:1.5px solid #C4C0BA;border-radius:9px;overflow:hidden;box-shadow:0 8px 30px rgba(0,0,0,.12)}
.dbar{height:22px;background:#E2DFD8;border-bottom:1px solid #C4C0BA;display:flex;align-items:center;gap:4px;padding:0 9px}
.ddot{width:7px;height:7px;border-radius:50%;background:#C4BFB8}
.durl{flex:1;height:10px;background:#D2CEC8;border-radius:5px;margin:0 8px;max-width:360px}
.dnav{height:32px;background:#012851;display:flex;align-items:center;gap:13px;padding:0 16px}
.dnav .nlogo{font-family:'Tinos',serif;font-size:10px;color:#fff;font-weight:700}
.dnav .nlink{font-size:7.5px;color:rgba(255,255,255,.5);font-weight:700;text-transform:uppercase;letter-spacing:.4px}
.dnav .nlink.on{color:#fff;border-bottom:2px solid #a1dcd8;padding-bottom:9px}
.dnav .av{margin-left:auto;width:18px;height:18px;border-radius:50%;background:rgba(255,255,255,.12);display:flex;align-items:center;justify-content:center;font-size:6px;font-weight:800;color:rgba(255,255,255,.55)}
.dcanvas{background:#f0efe9;padding:20px 26px 26px}
/* form atoms (mirror real Tailwind look) */
.lbl{font-size:8.5px;font-weight:800;letter-spacing:1.2px;text-transform:uppercase;color:#6b7280;margin-bottom:6px}
.inp{border:1px solid #e4e2d7;border-radius:4px;background:#fff;padding:8px 11px;font-size:12px;color:#012851}
.inp.title{font-family:'Tinos',serif;font-size:22px;font-weight:700;padding:11px 13px}
.card{border:1px solid #e4e2d7;border-radius:6px;background:#fff;padding:15px;box-shadow:0 1px 3px rgba(0,0,0,.04)}
.card h3{font-size:8.5px;font-weight:800;letter-spacing:1.2px;text-transform:uppercase;color:#6b7280;margin-bottom:3px}
.card .hint{font-size:9.5px;color:#9a9a96;margin-bottom:9px}
.seg{display:inline-flex;border:1px solid #012851;border-radius:5px;overflow:hidden}
.seg span{font-size:10px;font-weight:700;padding:5px 13px;color:#012851}
.seg span.on{background:#012851;color:#fff}
.seg span+span{border-left:1px solid #012851}
.chips{border:1px solid #e4e2d7;border-radius:5px;background:#fff;padding:7px;display:flex;flex-wrap:wrap;gap:6px;align-items:center}
.chip-sel{display:inline-flex;align-items:center;gap:4px;background:#f5f4ef;border-radius:4px;padding:3px 8px;font-size:10px;color:#012851}
.chip-sel .x{color:#012851;opacity:.45;font-size:11px}
.chip-in{flex:1;min-width:90px;font-size:10.5px;color:#9a9a96;padding:3px 4px}
.btn{display:inline-flex;align-items:center;gap:6px;height:38px;padding:0 16px;border-radius:5px;font-size:12px;font-weight:600}
.btn.primary{background:#012851;color:#fff}
.btn.ghost{background:#fff;border:1px solid #e4e2d7;color:#012851}
.btn.danger{background:#fff;border:1px solid #e7c9c4;color:#c0392b}
.tagchip{display:inline-flex;align-items:center;gap:3px;font-size:8px;border-radius:9px;padding:2px 7px;color:#a0522d;background:#f6ece6}
.tagchip i{width:6px;height:6px;border-radius:2px;background:#a0522d;display:inline-block}
/* annotation callouts on mockups */
.anno{display:flex;gap:9px;align-items:flex-start;font-size:11.5px;color:#4a4a46;line-height:1.55;margin-bottom:7px}
.anno .n{flex-shrink:0;width:18px;height:18px;border-radius:50%;background:#012851;color:#a1dcd8;font-size:9px;font-weight:800;display:flex;align-items:center;justify-content:center;margin-top:1px}
.anno b{color:#012851}
.anno code{font-size:10px;background:#F0EFE9;padding:1px 4px;border-radius:2px}
.annogrid{display:grid;grid-template-columns:1fr 1fr;gap:6px 26px;margin-top:14px}
/* dropdown */
.dd{border:1px solid #e4e2d7;border-radius:6px;background:#fff;box-shadow:0 6px 18px rgba(0,0,0,.12);overflow:hidden;margin-top:3px}
.dd .opt{padding:7px 11px;font-size:11px;color:#012851;border-bottom:1px solid #f3f1ea;cursor:pointer}
.dd .opt:hover,.dd .opt.hl{background:#f5f4ef}
.dd .opt:last-child{border-bottom:none}
.dd .opt .d{font-size:8.5px;color:#9a9a96}
/* states grid */
.states{display:grid;grid-template-columns:repeat(2,1fr);gap:18px}
.state{border:1px solid #E0DDD6;border-radius:8px;background:#fff;overflow:hidden}
.state .sh2{background:#F4F2EC;border-bottom:1px solid #E0DDD6;padding:7px 12px;font-size:8.5px;font-weight:800;letter-spacing:.8px;text-transform:uppercase;color:#012851;display:flex;justify-content:space-between}
.state .sb{padding:13px}
.cap{font-size:11px;color:#888;font-style:italic;line-height:1.55;margin-top:8px}
/* impl-ref */
.impl-ref{background:#fff;border:1px solid #E0DDD6;border-radius:7px;overflow:hidden;margin-top:8px}
.impl-ref table{width:100%;border-collapse:collapse}
.impl-ref th{background:#012851;color:#fff;padding:8px 13px;text-align:left;font-size:8px;font-weight:800;letter-spacing:.6px;text-transform:uppercase}
.impl-ref td{padding:8px 13px;border-bottom:1px solid #F0EEE8;vertical-align:top;font-size:11px;color:#444;line-height:1.55}
.impl-ref tr:nth-child(even) td{background:#FAFAF7}
.impl-ref td:first-child{font-weight:700;color:#012851;white-space:nowrap;width:185px}
.impl-ref td code{font-size:9.5px;background:#F0EFE9;padding:1px 4px;border-radius:2px;font-family:'Courier New',monospace;color:#333}
hr{border:none;border-top:2px dashed #C8C4BE;margin:52px 0}
.note{font-size:11px;color:#888;font-style:italic;margin-top:10px;line-height:1.6}
</style>
</head>
<body>
<div class="page">
<!-- ══ MASTHEAD ══ -->
<div class="mh">
<h1>Ereignis-Editor &amp; Brief-Gruppierung · Quick-Action im Dokument</h1>
<p>Wie kuratierte Zeitstrahl-Ereignisse entstehen und wie Briefe gruppiert werden — von zwei Seiten in ein Datenmodell (<code style="font-family:monospace;font-size:12px">TimelineEvent.documents</code>): der <strong>Ereignis-Editor</strong> unter <code style="font-family:monospace;font-size:12px">/zeitstrahl/events/[id]/edit</code> (Kurator baut, verlinkt viele Briefe) und die <strong>Quick-Action im Dokument-Detail</strong> (beim Lesen schnell zuordnen). Beide bauen auf bereits ausgelieferten Komponenten auf.</p>
<div class="tag-row">
<span class="tg">Milestone #14 · Zeitstrahl</span>
<span class="tg mint">Reuse: GeschichteEditor · DocumentMultiSelect · PersonMultiSelect</span>
<span class="tg slate">WRITE_ALL</span>
</div>
<div class="byline">Familienarchiv · 2026-06-08 · Leonie Voss, UX Lead · gegründet auf Code: GeschichteEditor.svelte · DocumentMetadataDrawer.svelte · DocumentMultiSelect.svelte</div>
</div>
<!-- ══ 1 · TWO ENTRY POINTS ══ -->
<div class="sh">
<h2>1 · Zwei Einstiegspunkte, ein Datenmodell</h2>
<p>Manuelle Gruppierung = ein <code>TimelineEvent</code> mit verknüpften Dokumenten. Kuratoren arbeiten in beide Richtungen — wir bauen beide, statt eine zu erzwingen.</p>
</div>
<div class="states" style="grid-template-columns:1fr 1fr">
<div class="callout navy" style="margin:0">
<strong>A · Ereignis-zuerst</strong> — der Kurator baut den Zeitstrahl. <code>/zeitstrahl/events/new · [id]/edit</code> mit Dokument-Mehrfach-Picker = <b>Bulk-Linking</b> vieler Briefe auf einmal. Spiegelt 1:1 den <code>GeschichteEditor</code> (gleiche zwei-Spalten-Form, Sidebar-Picker, Sticky-Save-Bar).
</div>
<div class="callout mint" style="margin:0">
<strong>B · Dokument-zuerst</strong> — beim Lesen eines Briefs. Quick-Action im Dokument-Detail: bestehendes Ereignis wählen <i>oder</i> neu anlegen, verlinkt diesen einen Brief. Spiegelt die bestehende <b>Geschichten-Spalte</b> im Details-Drawer (<code>DocumentMetadataDrawer.svelte</code>).
</div>
</div>
<!-- ══ 2 · EVENT EDITOR ══ -->
<div class="sh">
<h2>2 · Ereignis-Editor — <code style="font-size:14px">/zeitstrahl/events/[id]/edit</code></h2>
<p>Form-Actions-Muster, gegated mit <code>WRITE_ALL</code>. Layout &amp; Verhalten 1:1 vom <code>GeschichteEditor</code> übernommen: Hauptspalte + Sidebar (<code>lg:grid-cols-[2fr_1fr]</code>), Sticky-Save-Bar, <code>beforeNavigate</code>-Warnung bei ungespeicherten Änderungen.</p>
</div>
<div class="dchrome" style="margin-bottom:16px">
<div class="dbar"><div class="ddot"></div><div class="ddot"></div><div class="ddot"></div><div class="durl"></div></div>
<div class="dnav"><span class="nlogo">Familienarchiv</span><span class="nlink">Dokumente</span><span class="nlink">Personen</span><span class="nlink on">Zeitstrahl</span><span class="nlink">Stammbaum</span><span class="av">KR</span></div>
<div class="dcanvas">
<div style="font-size:8px;font-weight:800;letter-spacing:1px;text-transform:uppercase;color:#8a8a86;margin-bottom:10px"> Zurück zum Zeitstrahl</div>
<div style="font-family:'Tinos',serif;font-size:18px;font-weight:700;color:#012851;margin-bottom:16px">Ereignis bearbeiten</div>
<div style="display:grid;grid-template-columns:2fr 1fr;gap:22px">
<!-- MAIN COLUMN -->
<div style="display:flex;flex-direction:column;gap:16px">
<!-- ① title -->
<div>
<div class="inp title" style="width:100%">Briefe von der Front</div>
</div>
<!-- ② type + ③ date/precision -->
<div style="display:flex;gap:22px;flex-wrap:wrap">
<div>
<div class="lbl">② Typ</div>
<div class="seg"><span class="on">Persönlich</span><span>Historisch</span></div>
</div>
<div>
<div class="lbl">③ Datum · Präzision</div>
<div style="display:flex;gap:6px;align-items:center">
<div class="inp" style="width:120px">1915</div>
<div class="inp" style="display:flex;align-items:center;gap:18px;color:#6b7280">Jahr <span style="font-size:8px"></span></div>
</div>
<div style="font-size:9px;color:#9a9a96;margin-top:5px;font-style:italic">Bei „Zeitspanne" erscheint ein zweites End-Datum-Feld. Bei „ca." / „Saison" passt sich nur das Label an.</div>
</div>
</div>
<!-- ④ description -->
<div>
<div class="lbl">④ Beschreibung <span style="color:#bbb;font-weight:600">· optional</span></div>
<div class="inp" style="width:100%;min-height:96px;color:#4a4a46;font-family:'Tinos',serif;line-height:1.6">Karls Feldpost von der Westfront, 1915 — wöchentliche Briefe an Elfriede und den neugeborenen Hans. Eine zusammenhängende Korrespondenz, die hier als Cluster gebündelt wird …</div>
<div style="font-size:9px;color:#9a9a96;margin-top:5px;font-style:italic">Schlichtes Textfeld (kein Rich-Text wie Geschichten) — Ereignisse sind kurze Notizen, keine Langform.</div>
</div>
</div>
<!-- SIDEBAR -->
<aside style="display:flex;flex-direction:column;gap:16px">
<!-- ⑤ linked letters = grouping -->
<div class="card" style="border-color:#a1dcd8;box-shadow:0 2px 10px rgba(161,220,216,.3)">
<h3 style="color:#012851">⑤ Verknüpfte Briefe · 24</h3>
<div class="hint">Diese Briefe bilden den Cluster. <code style="font-size:9px">DocumentMultiSelect</code></div>
<div class="chips">
<span class="chip-sel">✉ Westfront-Brief · Mär 1915 <span class="x">×</span></span>
<span class="chip-sel">✉ Feldpost Verdun · Jul 1915 <span class="x">×</span></span>
<span class="chip-sel">✉ Brief an Elfriede · Sep 1915 <span class="x">×</span></span>
<span class="chip-in">Brief suchen …</span>
</div>
</div>
<!-- ⑥ persons -->
<div class="card">
<h3>⑥ Beteiligte Personen</h3>
<div class="hint">Treibt die Lebensweg-Ansicht &amp; Filter. <code style="font-size:9px">PersonMultiSelect</code></div>
<div class="chips">
<span class="chip-sel">Karl Raddatz <span class="x">×</span></span>
<span class="chip-sel">Elfriede Raddatz <span class="x">×</span></span>
<span class="chip-sel">Hans Raddatz <span class="x">×</span></span>
<span class="chip-in">Person suchen …</span>
</div>
</div>
</aside>
</div>
<!-- ⑦ sticky save bar -->
<div style="margin:18px -26px -26px;border-top:1px solid #e4e2d7;background:#fff;box-shadow:0 -2px 8px rgba(0,0,0,.05);padding:13px 26px;display:flex;align-items:center;justify-content:space-between">
<span style="font-size:10px;color:#9a9a96">Änderungen werden erst beim Speichern übernommen.</span>
<div style="display:flex;gap:8px">
<span class="btn danger">Löschen</span>
<span class="btn ghost">Abbrechen</span>
<span class="btn primary">Speichern</span>
</div>
</div>
</div>
</div>
<div class="annogrid">
<div class="anno"><span class="n"></span><span><b>Titel</b> — großes Serifen-Feld, wie der Geschichten-Titel. Pflichtfeld (Validierung bei Blur).</span></div>
<div class="anno"><span class="n"></span><span><b>Typ</b><code>PERSONAL</code> / <code>HISTORICAL</code> Segmented-Control. Steuert Rendering (Mint-Pille vs. Welt-Band).</span></div>
<div class="anno"><span class="n"></span><span><b>Datum + Präzision</b> — geteilte <code>DatePrecisionInput</code> (gleiche Logik wie Dokument-Datum, <code>metaDatePrecision</code>). „Zeitspanne" blendet End-Datum ein.</span></div>
<div class="anno"><span class="n"></span><span><b>Beschreibung</b> — optionales Textfeld (<code>TEXT</code>), bewusst schlicht.</span></div>
<div class="anno"><span class="n"></span><span><b>Verknüpfte Briefe</b><b>hier wird gruppiert.</b> Wiederverwendung von <code>DocumentMultiSelect</code> (Typeahead, Chips, Hidden-Inputs).</span></div>
<div class="anno"><span class="n"></span><span><b>Beteiligte Personen</b><code>PersonMultiSelect</code>. Bestimmt, in welchem „Lebensweg" das Ereignis auftaucht.</span></div>
<div class="anno"><span class="n"></span><span><b>Sticky-Save-Bar</b> — Speichern primär, Abbrechen sekundär, Löschen nur im Edit-Modus (mit Bestätigung).</span></div>
<div class="anno"><span class="n"></span><span><b>/new</b> — leeres Formular. Mit <code>?documentId=…</code> ist Feld ⑤ vorbefüllt (aus der Quick-Action, §4-D).</span></div>
</div>
<!-- ══ 3 · GROUPING / DOCUMENT PICKER ══ -->
<div class="sh">
<h2>3 · Brief-Gruppierung im Editor — der Dokument-Picker</h2>
<p>Feld ⑤ ist der unveränderte <code>DocumentMultiSelect</code>: Tippen sucht über <code>/api/documents/search?q=…</code> (debounced 300&nbsp;ms), Treffer mit ehrlichem Datums-Label, bereits gewählte werden gefiltert. Jeder Klick fügt einen Brief zum Cluster.</p>
</div>
<div class="states">
<div class="state">
<div class="sh2"><span>Suche aktiv — Dropdown</span><span style="color:#9a9a96">DocumentMultiSelect</span></div>
<div class="sb">
<div class="chips" style="margin-bottom:0">
<span class="chip-sel">✉ Westfront-Brief · Mär 1915 <span class="x">×</span></span>
<span class="chip-sel">✉ Feldpost Verdun · Jul 1915 <span class="x">×</span></span>
<span class="chip-in" style="color:#012851">Verdun▏</span>
</div>
<div class="dd">
<div class="opt hl">Feldpost aus Verdun <span class="d">· Brief · Juli 1915</span></div>
<div class="opt">Brief aus dem Verdun-Lazarett <span class="d">· Brief · August 1916</span></div>
<div class="opt">Rückkehr aus Verdun <span class="d">· Brief · ca. 1917</span></div>
</div>
<div class="cap">Label = <code style="font-size:10px">title · formatDocumentDate(precision)</code>. Bereits verknüpfte Briefe erscheinen nicht in den Treffern (Dedup).</div>
</div>
</div>
<div class="state">
<div class="sh2"><span>Inline „+ Ereignis" am Jahres-Band</span><span style="color:#9a9a96">Zeitstrahl</span></div>
<div class="sb">
<div style="position:relative;padding-left:20px">
<div style="position:absolute;left:6px;top:2px;bottom:2px;width:2px;background:linear-gradient(#a1dcd8,#012851)"></div>
<div style="font-family:'Tinos',serif;font-size:13px;font-weight:700;color:#012851;margin-bottom:8px">1915</div>
<div style="background:#FAF9F5;border:1px solid #eeede8;border-radius:4px;padding:7px 9px;margin-bottom:8px"><div style="font-size:9.5px;font-weight:700;color:#012851">✉ 24 Briefe</div><div style="font-size:8px;color:#9a9a96">Monats-Dichte ▾</div></div>
<button style="display:inline-flex;align-items:center;gap:5px;border:1px dashed #a1dcd8;background:#fff;border-radius:6px;padding:5px 11px;font-size:10px;font-weight:600;color:#012851"> Ereignis aus diesem Jahr anlegen</button>
</div>
<div class="cap">Kuratoren können auch direkt im Zeitstrahl ein Ereignis anlegen — öffnet denselben Editor, Jahr &amp; Briefe des Bandes vorbefüllt.</div>
</div>
</div>
</div>
<hr>
<!-- ══ 4 · QUICK ACTION IN DOCUMENT DETAIL ══ -->
<div class="sh">
<h2>4 · Quick-Action im Dokument-Detail — wo sie lebt</h2>
<p>Die Dokument-Detailseite ist ein <b>vollflächiger Viewer ohne Sidebar</b> (<code>fixed inset</code>). Aktions-Flächen gibt es nur zwei: die <code>DocumentTopBar</code> und den aufklappbaren <b>Details-Drawer</b>. Die Quick-Action lebt an beiden — primär als <b>„Zeitstrahl"-Spalte im Drawer</b> (spiegelt die Geschichten-Spalte), plus ein <b>Top-Bar-Button</b> für den Ein-Klick-Weg.</p>
</div>
<div class="callout navy"><strong>Warum der Details-Drawer der richtige Ort ist:</strong> Er zeigt heute schon, <i>wozu ein Brief gehört</i> — Personen, Schlagwörter und <b>Geschichten</b> (mit „Zuordnen"-Aktion, gegated über <code>canBlogWrite</code>, <code>DocumentMetadataDrawer.svelte:210</code>). Zeitstrahl-Ereignisse sind strukturell identisch („dieser Brief gehört zu diesen Ereignissen") und bekommen daher eine gleichwertige vierte/fünfte Spalte. Konsistent &amp; auffindbar dort, wo Nutzer ohnehin „Zugehörigkeit" suchen.</div>
<div class="dchrome" style="margin-bottom:16px">
<div class="dbar"><div class="ddot"></div><div class="ddot"></div><div class="ddot"></div><div class="durl"></div></div>
<!-- document topbar -->
<div style="background:#fff;border-bottom:1px solid #e4e2d7;box-shadow:0 1px 3px rgba(0,0,0,.05);display:flex;align-items:center;height:54px">
<div style="width:3px;height:100%;background:#012851"></div>
<div style="width:34px;display:flex;justify-content:center;color:#6b7280;font-size:14px"></div>
<div style="width:1px;height:22px;background:#e4e2d7;margin:0 6px"></div>
<div style="flex:0 1 auto;min-width:0;padding-right:10px">
<div style="font-family:'Tinos',serif;font-size:13px;font-weight:700;color:#012851;white-space:nowrap;overflow:hidden;text-overflow:ellipsis">Brief über die Lage an der Westfront</div>
<div style="font-size:8.5px;color:#6b7280">März 1915</div>
</div>
<div style="display:flex;gap:3px;margin-left:6px"><span style="width:20px;height:20px;border-radius:50%;background:#012851;color:#fff;font-size:7px;display:flex;align-items:center;justify-content:center">KR</span><span style="font-size:9px;color:#9a9a96;align-self:center"></span><span style="width:20px;height:20px;border-radius:50%;background:#5a8a6a;color:#fff;font-size:7px;display:flex;align-items:center;justify-content:center">ER</span></div>
<div style="margin-left:auto;display:flex;align-items:center;gap:7px;padding-right:14px">
<!-- Details toggle (active) -->
<span style="display:inline-flex;align-items:center;gap:5px;background:#012851;color:#fff;border-radius:5px;font-size:10px;font-weight:600;padding:6px 11px">Details ▴</span>
<div style="width:1px;height:22px;background:#e4e2d7"></div>
<!-- action buttons -->
<span style="display:inline-flex;align-items:center;gap:4px;border:1px solid #e4e2d7;border-radius:5px;font-size:10px;font-weight:600;color:#012851;padding:6px 11px">✎ Transkribieren</span>
<!-- NEW: Zeitstrahl quick button -->
<span style="display:inline-flex;align-items:center;gap:4px;background:#a1dcd8;color:#012851;border-radius:5px;font-size:10px;font-weight:700;padding:6px 11px">⊕ Zeitstrahl</span>
<span style="display:inline-flex;align-items:center;gap:4px;border:1px solid #e4e2d7;border-radius:5px;font-size:10px;color:#012851;padding:6px 9px"></span>
</div>
</div>
<!-- metadata drawer (opened) -->
<div style="background:#fff;border-bottom:1px solid #e4e2d7;padding:20px 24px">
<div style="display:grid;grid-template-columns:repeat(4,1fr);gap:24px">
<!-- Details -->
<div>
<div class="lbl">Details</div>
<div style="font-size:8px;font-weight:600;color:#9a9a96;margin-bottom:1px">Datum</div><div style="font-family:'Tinos',serif;font-size:12px;color:#012851;margin-bottom:8px">März 1915</div>
<div style="font-size:8px;font-weight:600;color:#9a9a96;margin-bottom:1px">Ort</div><div style="font-family:'Tinos',serif;font-size:12px;color:#012851;margin-bottom:8px">Westfront</div>
<div style="font-size:8px;font-weight:600;color:#9a9a96;margin-bottom:1px">Status</div><div style="font-family:'Tinos',serif;font-size:12px;color:#012851">Transkribiert</div>
</div>
<!-- Personen -->
<div>
<div class="lbl">Personen</div>
<div style="display:flex;align-items:center;gap:7px;margin-bottom:6px"><span style="width:26px;height:26px;border-radius:50%;background:#012851;color:#fff;font-size:8px;display:flex;align-items:center;justify-content:center">KR</span><span style="font-family:'Tinos',serif;font-size:12px;color:#012851">Karl Raddatz</span></div>
<div style="display:flex;align-items:center;gap:7px"><span style="width:26px;height:26px;border-radius:50%;background:#5a8a6a;color:#fff;font-size:8px;display:flex;align-items:center;justify-content:center">ER</span><span style="font-family:'Tinos',serif;font-size:12px;color:#012851">Elfriede Raddatz</span></div>
</div>
<!-- Schlagwörter -->
<div>
<div class="lbl">Schlagwörter</div>
<div style="display:flex;flex-wrap:wrap;gap:5px"><span style="background:#f5f4ef;border-radius:3px;font-size:8px;font-weight:700;letter-spacing:.4px;text-transform:uppercase;color:#012851;padding:3px 7px">Krieg</span><span style="background:#f5f4ef;border-radius:3px;font-size:8px;font-weight:700;letter-spacing:.4px;text-transform:uppercase;color:#012851;padding:3px 7px">Briefe von der Front</span></div>
</div>
<!-- NEW: Zeitstrahl column -->
<div style="border-left:2px solid #eef6f5;padding-left:18px;margin-left:-6px">
<div style="display:flex;align-items:center;justify-content:space-between;margin-bottom:9px">
<span class="lbl" style="margin:0;color:#012851">Zeitstrahl</span>
<span style="font-size:9px;font-weight:600;color:#6b7280">+ Zuordnen</span>
</div>
<!-- linked event -->
<div style="border:1px solid #e4e2d7;border-radius:5px;padding:7px 9px;margin-bottom:8px">
<div style="display:flex;align-items:center;justify-content:space-between"><span style="font-family:'Tinos',serif;font-size:11px;font-weight:700;color:#012851">Briefe von der Front</span><span style="color:#9a9a96;font-size:11px">×</span></div>
<div style="font-size:8px;color:#9a9a96;margin-top:1px">1915 · 24 Briefe · persönlich</div>
<span class="tagchip" style="margin-top:5px"><i></i>Krieg</span>
</div>
<!-- quick add row -->
<div style="display:flex;gap:6px">
<span style="flex:1;border:1px solid #e4e2d7;border-radius:4px;font-size:9px;color:#9a9a96;padding:5px 8px">Ereignis suchen …</span>
<span style="background:#012851;color:#fff;border-radius:4px;font-size:9px;font-weight:600;padding:5px 8px;white-space:nowrap"> Neu</span>
</div>
</div>
</div>
</div>
<div style="height:120px;background:repeating-linear-gradient(45deg,#ececec,#ececec 8px,#e4e4e4 8px,#e4e4e4 16px);display:flex;align-items:center;justify-content:center;color:#aaa;font-size:10px">↓ PDF-Viewer (Brief-Scan) …</div>
</div>
<div class="annogrid">
<div class="anno"><span class="n">A</span><span><b>Top-Bar-Button „⊕ Zeitstrahl"</b> — Mint-Akzent im Aktions-Cluster (<code>DocumentTopBarActions</code>). Öffnet ein kleines Popover zum Ein-Klick-Zuordnen, ohne den Drawer zu öffnen. Im Mobile-Menü als Eintrag.</span></div>
<div class="anno"><span class="n">B</span><span><b>„Zeitstrahl"-Spalte im Details-Drawer</b> — neue Spalte neben Geschichten. Zeigt verknüpfte Ereignisse (Titel · Datum · Tag-Chip), Unlink über <code>×</code>, plus Quick-Add-Zeile. Nur sichtbar/aktiv bei <code>canWrite</code>.</span></div>
<div class="anno"><span class="n">C</span><span><b>Quick-Add-Zeile</b> — Typeahead „Ereignis suchen …" (sofortiges Verlinken, keine Navigation) + <b>„+ Neu"</b>.</span></div>
<div class="anno"><span class="n">D</span><span><b>„+ Neu"</b><code>/zeitstrahl/events/new?documentId={id}</code> — öffnet den Editor (§2) mit diesem Brief in Feld ⑤ vorbefüllt. Spiegelt <code>/geschichten/new?documentId=</code>.</span></div>
</div>
<!-- ══ 5 · QUICK-ADD STATES ══ -->
<div class="sh"><h2>5 · Quick-Action — Zustände</h2><p>Der Typeahead in der Zeitstrahl-Spalte (oder im Top-Bar-Popover). Gleiches Muster wie <code>DocumentMultiSelect</code>, nur sucht es Ereignisse statt Dokumente.</p></div>
<div class="states" style="grid-template-columns:repeat(2,1fr)">
<div class="state">
<div class="sh2"><span>A · Nicht zugeordnet</span></div>
<div class="sb">
<div style="font-size:10px;color:#9a9a96;font-style:italic;margin-bottom:9px">Noch keinem Ereignis zugeordnet.</div>
<div style="display:flex;gap:6px"><span style="flex:1;border:1px solid #e4e2d7;border-radius:4px;font-size:10px;color:#9a9a96;padding:6px 9px">Ereignis suchen …</span><span class="btn primary" style="height:30px;font-size:10px;padding:0 11px"> Neu</span></div>
</div>
</div>
<div class="state">
<div class="sh2"><span>B · Suche — Treffer</span></div>
<div class="sb">
<div style="border:1px solid #012851;border-radius:4px;font-size:10px;color:#012851;padding:6px 9px;margin-bottom:0">Front▏</div>
<div class="dd">
<div class="opt hl">Briefe von der Front <span class="d">· 1915 · 24 Briefe</span></div>
<div class="opt">Kriegsausbruch <span class="d">· 1914 · 6 Briefe</span></div>
<div class="opt" style="color:#012851;font-weight:600"> „Front" als neues Ereignis anlegen</div>
</div>
</div>
</div>
<div class="state">
<div class="sh2"><span>C · Zugeordnet</span></div>
<div class="sb">
<div style="border:1px solid #e4e2d7;border-radius:5px;padding:7px 9px"><div style="display:flex;align-items:center;justify-content:space-between"><span style="font-family:'Tinos',serif;font-size:11px;font-weight:700;color:#012851">Briefe von der Front</span><span style="font-size:9px;color:#2e7d57">✓ verknüpft <span style="color:#9a9a96">×</span></span></div><span class="tagchip" style="margin-top:5px"><i></i>Krieg</span></div>
<div class="cap">Sofortiges Verlinken (POST). Toast „Zum Ereignis hinzugefügt", <code style="font-size:10px">aria-live</code>. Unlink über <code>×</code> (DELETE).</div>
</div>
</div>
<div class="state">
<div class="sh2"><span>D · Mehrfach zugeordnet</span></div>
<div class="sb">
<div style="display:flex;flex-direction:column;gap:5px">
<div style="border:1px solid #e4e2d7;border-radius:5px;padding:5px 9px;display:flex;justify-content:space-between"><span style="font-family:'Tinos',serif;font-size:10.5px;color:#012851">Briefe von der Front</span><span style="font-size:9px;color:#9a9a96">×</span></div>
<div style="border:1px solid #e4e2d7;border-radius:5px;padding:5px 9px;display:flex;justify-content:space-between"><span style="font-family:'Tinos',serif;font-size:10.5px;color:#012851">Weihnachten 1915</span><span style="font-size:9px;color:#9a9a96">×</span></div>
</div>
<div class="cap">Ein Brief darf zu mehreren Ereignissen gehören (ManyToMany) — alle werden gelistet.</div>
</div>
</div>
</div>
<!-- ══ 6 · TOKENS ══ -->
<div class="sh"><h2>6 · Wiederverwendete Bausteine &amp; Tokens</h2></div>
<div class="callout mint"><strong>Maximal wiederverwenden:</strong> <code>DocumentMultiSelect</code> (Brief-Gruppierung, unverändert) · <code>PersonMultiSelect</code> (Beteiligte) · <code>GeschichteEditor</code>-Layout (zwei Spalten, Sticky-Save, <code>beforeNavigate</code>) · <code>DocumentMetadataDrawer</code>-Spaltenmuster (Quick-Action) · <code>useUnsavedWarning</code> · <code>formatDocumentDate</code> / <code>DatePrecision</code>. Brand-Tokens wie im Zeitstrahl-Spec: Navy <code>#012851</code>, Mint <code>#a1dcd8</code>, Linie <code>#e4e2d7</code>, ink-3 <code>#6b7280</code>, danger <code>#c0392b</code>; Serifen-Titel (Tinos), Sans-Chrome (Montserrat).</div>
<!-- ══ 7 · IMPL-REF ══ -->
<div class="sh"><h2>7 · Implementierungs-Referenz &amp; Barrierefreiheit</h2></div>
<div class="impl-ref">
<table>
<tr><th>Baustein</th><th>Datei / Endpoint</th><th>Verantwortung</th></tr>
<tr><td>Editor-Route (neu)</td><td><code>/zeitstrahl/events/new · [id]/edit</code></td><td><code>+page.server.ts</code> (Form-Actions, <code>WRITE_ALL</code>) + <code>+page.svelte</code>; <code>?documentId=</code> vorbefüllt Feld ⑤</td></tr>
<tr><td>Editor-Komponente (neu)</td><td><code>TimelineEventEditor.svelte</code></td><td>Spiegelt <code>GeschichteEditor</code>: Titel, Typ, Datum+Präzision, Beschreibung; Sidebar-Picker; Sticky-Save; <code>beforeNavigate</code></td></tr>
<tr><td>Brief-Gruppierung (reuse)</td><td><code>DocumentMultiSelect.svelte</code></td><td>Unverändert — Typeahead <code>/api/documents/search</code>, Chips, Hidden-Inputs <code>documentIds</code></td></tr>
<tr><td>Personen (reuse)</td><td><code>PersonMultiSelect.svelte</code></td><td>Unverändert — Beteiligte Personen</td></tr>
<tr><td>Datum + Präzision</td><td><code>DatePrecisionInput</code> (geteilt)</td><td>Wie Dokument-Datum (<code>metaDatePrecision</code>); „Zeitspanne" → End-Datum; <code>formatDocumentDate</code> fürs Label</td></tr>
<tr><td>Quick-Action-Spalte (neu)</td><td><code>DocumentTimelineColumn.svelte</code></td><td>Im <code>DocumentMetadataDrawer</code> neben Geschichten; verknüpfte Ereignisse + Quick-Add; nur bei <code>canWrite</code></td></tr>
<tr><td>Quick-Add-Picker (neu)</td><td><code>DocumentTimelineEventPicker.svelte</code></td><td>Ereignis-Typeahead; sofort verlinken oder <code>?documentId=</code> zum Editor; auch im Top-Bar-Popover</td></tr>
<tr><td>Top-Bar-Button (neu)</td><td><code>DocumentTopBarActions</code> · <code>DocumentMobileMenu</code></td><td>„⊕ Zeitstrahl"-Button (canWrite); öffnet Quick-Add-Popover</td></tr>
<tr><td>Backend — CRUD</td><td><code>POST · PUT · DELETE /api/timeline/events</code></td><td><code>TimelineEventController</code>, <code>WRITE_ALL</code>; <code>TimelineEventRequest</code> mit <code>documentIds</code> / <code>personIds</code></td></tr>
<tr><td>Backend — Link/Unlink</td><td><code>PUT /api/timeline/events/{id}</code></td><td>Verlinken/Lösen läuft über das Event-Update (<code>documents</code>-Set); kein neuer ErrorCode nötig</td></tr>
<tr><td>Barrierefreiheit</td><td></td><td>Picker-Dropdowns Tastatur-navigierbar (↑↓↵), <code>aria-live</code> für „verknüpft/gelöst"; 44px-Ziele; sichtbarer Fokus-Ring; Löschen/Unlink mit Bestätigung</td></tr>
</table>
</div>
<div class="note">Offene Designentscheidung: Soll der Top-Bar-Button (A) MVP sein oder reicht zunächst die Drawer-Spalte (B)? Empfehlung: <b>B als MVP</b> (spiegelt Geschichten exakt, geringster Aufwand), A als schneller Nachzug.</div>
</div>
</body>
</html>

View File

@@ -0,0 +1,391 @@
<!DOCTYPE html>
<html lang="de">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Globaler Zeitstrahl — Finale Spezifikation (Konzept A) · Milestone #14 · Familienarchiv</title>
<link href="https://fonts.googleapis.com/css2?family=Tinos:ital,wght@0,400;0,700;1,400&family=Montserrat:wght@400;500;600;700;800;900&display=swap" rel="stylesheet">
<style>
*,*::before,*::after{box-sizing:border-box;margin:0;padding:0}
body{font-family:'Montserrat',system-ui,sans-serif;background:#ECEAE4;color:#1A1A1A;line-height:1.5;font-size:13px}
.page{max-width:1320px;margin:0 auto;padding:48px 32px 120px}
/* Masthead */
.mh{padding-bottom:24px;border-bottom:3px solid #012851;margin-bottom:48px}
.mh h1{font-size:23px;font-weight:900;color:#012851;letter-spacing:-.4px}
.mh p{font-size:13px;color:#555;max-width:780px;line-height:1.75;margin-top:8px}
.mh .byline{font-size:9px;color:#999;font-weight:700;letter-spacing:1.5px;text-transform:uppercase;margin-top:10px}
.tag-row{display:flex;gap:6px;margin-top:12px;flex-wrap:wrap}
.tg{background:#012851;color:#a1dcd8;padding:2px 8px;border-radius:2px;font-size:8px;font-weight:700;letter-spacing:.8px;text-transform:uppercase}
.tg.mint{background:#a1dcd8;color:#012851}
.tg.slate{background:#607080;color:#e8edf2}
.sh{margin:60px 0 24px;padding-bottom:12px;border-bottom:2px solid #E0DDD6}
.sh h2{font-size:17px;font-weight:900;color:#012851}
.sh p{font-size:12.5px;color:#666;margin-top:5px;max-width:780px;line-height:1.65}
.callout{padding:13px 17px;border-radius:4px;font-size:12px;line-height:1.65;margin-bottom:18px}
.callout.navy{background:rgba(1,40,81,.06);border-left:3px solid #012851;color:#333}
.callout.mint{background:rgba(161,220,216,.18);border-left:3px solid #00c7b1;color:#1f3a3a}
.callout strong{font-weight:800;color:#012851}
/* legend */
.legend{display:grid;grid-template-columns:repeat(3,1fr);gap:12px}
.lg{background:#fff;border:1px solid #E0DDD6;border-radius:7px;padding:12px 14px;display:flex;gap:11px}
.lg .ico{flex-shrink:0;width:30px;height:30px;border-radius:50%;display:flex;align-items:center;justify-content:center;font-size:14px}
.lg .ttl{font-size:10.5px;font-weight:800;color:#012851;margin-bottom:2px}
.lg .body{font-size:9.5px;color:#5a5a56;line-height:1.5}
.lg .body code{font-size:8.5px;background:#F0EFE9;padding:1px 3px;border-radius:2px;color:#444}
/* precision table */
.rules{background:#fff;border:1px solid #E0DDD6;border-radius:7px;overflow:hidden;margin-top:8px}
.rules table{width:100%;border-collapse:collapse}
.rules th{background:#F4F2EC;font-size:8px;font-weight:800;text-transform:uppercase;letter-spacing:.8px;color:#888;padding:8px 12px;text-align:left;border-bottom:1px solid #E0DDD6}
.rules td{font-size:11px;color:#444;padding:7px 12px;border-bottom:1px solid #F0EEE8;vertical-align:top;line-height:1.5}
.rules tr:last-child td{border-bottom:none}
.rules td:first-child{font-weight:700;color:#012851;width:110px}
.rules td code{font-size:10px;background:#F0EFE9;padding:1px 5px;border-radius:2px;color:#444}
.rules .ex{font-family:'Tinos',serif;font-size:13px;color:#012851}
/* ── Timeline atoms (Konzept A) ── */
.tl-canvas{background:#f0efe9;border:1.5px solid #E0DDD6;border-radius:10px;padding:24px 26px 30px}
.dh{font-family:'Tinos',serif;font-size:20px;font-weight:700;color:#012851}
.dh-sub{font-size:9.5px;color:#7a7a76;margin-bottom:20px}
/* desktop centered axis */
.axis{position:relative;max-width:780px;margin:0 auto;padding:4px 0}
.axis::before{content:"";position:absolute;left:50%;top:0;bottom:0;width:2.5px;background:linear-gradient(#a1dcd8,#012851,#607080);transform:translateX(-50%)}
.ybadge{text-align:center;position:relative;margin:4px 0 14px}
.ybadge span{background:#012851;color:#fff;font-family:'Tinos',serif;font-size:13px;font-weight:700;padding:2px 15px;border-radius:12px;position:relative;z-index:2}
.pill{text-align:center;position:relative;margin-bottom:15px}
.pill .inner{display:inline-flex;align-items:center;gap:9px;background:#fff;border:1.5px solid #012851;border-radius:22px;padding:5px 16px 5px 5px;position:relative;z-index:2;box-shadow:0 2px 7px rgba(1,40,81,.12)}
.pill.curated .inner{border-color:#a1dcd8;border-width:2px}
.pill .gly{width:28px;height:28px;border-radius:50%;background:#012851;color:#a1dcd8;font-size:14px;display:flex;align-items:center;justify-content:center;flex-shrink:0}
.pill.curated .gly{background:#a1dcd8;color:#012851}
.pill .tx{text-align:left}
.pill .tx .t{font-family:'Tinos',serif;font-size:12px;font-weight:700;color:#012851;display:block;line-height:1.15}
.pill .tx .s{font-size:8.5px;color:#8a8a86}
.wband{position:relative;margin:0 -26px 15px;padding:7px 26px;background:#EEEBE2;border-top:1px solid #ddd8cc;border-bottom:1px solid #ddd8cc;text-align:center}
.wband .t{font-family:'Tinos',serif;font-style:italic;font-size:12px;color:#5a6776}
.wband .s{font-size:8.5px;color:#8a8a86;margin-left:8px}
.lrow{display:flex;align-items:flex-start;margin-bottom:12px}
.lrow .half{flex:1}
.lrow .dot{width:13px;height:13px;border-radius:50%;background:#fff;border:2.5px solid #a1dcd8;margin-top:9px;flex-shrink:0;z-index:2;position:relative}
.lrow .a{padding-right:26px;text-align:right}
.lrow .b{padding-left:26px;text-align:left}
.lcard{display:inline-block;text-align:left;background:#fff;border:1px solid #e4e2d7;border-radius:5px;padding:8px 11px;box-shadow:0 1px 3px rgba(0,0,0,.05);max-width:300px}
.lcard.ev{border-left:3px solid #a1dcd8}
.lcard .t{font-size:11px;font-weight:700;color:#1A1A1A}
.lcard.ev .t{font-family:'Tinos',serif}
.lcard .m{font-size:8.5px;color:#9a9a96;margin-top:1px}
.chip{display:inline-flex;align-items:center;gap:3px;font-size:7.5px;border-radius:9px;padding:2px 7px;margin-top:5px}
.chip i{width:6px;height:6px;border-radius:2px;display:inline-block;flex-shrink:0}
.chip.krieg{color:#a0522d;background:#f6ece6}.chip.krieg i{background:#a0522d}
.chip.weih{color:#c17a00;background:#fbf3e3}.chip.weih i{background:#c17a00}
.chip.fam{color:#5a8a6a;background:#eaf1ec}.chip.fam i{background:#5a8a6a}
/* dense aggregate strip */
.strip{max-width:440px;margin:0 auto 15px;position:relative;z-index:2;background:#fff;border:1px solid #e4e2d7;border-radius:6px;box-shadow:0 1px 3px rgba(0,0,0,.05);padding:9px 13px}
.strip .hd{display:flex;align-items:center;justify-content:space-between;margin-bottom:6px}
.strip .ct{font-size:10.5px;font-weight:700;color:#012851}
.strip .ex{font-size:8px;color:#8a8a86}
.spark{display:flex;align-items:flex-end;gap:1.5px;height:30px}
.spark div{flex:1;background:#a1dcd8;border-radius:1px;min-height:1px}
.strip .axl{display:flex;justify-content:space-between;margin-top:3px}
.strip .axl span{font-size:6px;color:#bbb}
/* compressed gap */
.gap{max-width:440px;margin:0 auto 15px;position:relative;z-index:2;display:flex;align-items:center;gap:9px;padding:5px 14px;border:1px dashed #cfccc4;border-radius:18px;background:#f0efe9;color:#9a958c;font-size:9px;font-style:italic}
.gap .ln{flex:1;height:1px;background:#ddd8cc}
.gap b{color:#7a756c;font-style:normal;font-family:'Tinos',serif;font-size:10px}
/* undated bucket */
.undated{max-width:540px;margin:20px auto 0;background:#fff;border:1px dashed #c4c0ba;border-radius:6px;padding:12px 15px}
.undated .h{font-family:'Tinos',serif;font-size:12px;font-weight:700;color:#7a756c;margin-bottom:6px}
/* case tag floating */
.casetag{display:inline-block;background:#012851;color:#a1dcd8;font-size:7px;font-weight:800;letter-spacing:.6px;text-transform:uppercase;padding:2px 7px;border-radius:3px;margin-bottom:6px}
/* narrow (phone / rail) column */
.col3{display:grid;grid-template-columns:repeat(3,1fr);gap:18px}
.modehd{font-size:9px;font-weight:800;letter-spacing:1px;text-transform:uppercase;color:#012851;margin-bottom:4px;display:flex;align-items:center;gap:6px}
.modehd .seg{background:#012851;color:#fff;font-size:7px;padding:2px 7px;border-radius:4px}
.modesub{font-size:9.5px;color:#888;font-style:italic;margin-bottom:9px;line-height:1.45;min-height:42px}
.nf{background:#fff;border:1px solid #e4e2d7;border-radius:6px;padding:14px}
.nsp{position:relative;padding-left:21px}
.nsp::before{content:"";position:absolute;left:7px;top:4px;bottom:4px;width:2px;background:linear-gradient(#a1dcd8,#012851,#607080)}
.nyr{font-family:'Tinos',serif;font-size:12px;font-weight:700;color:#012851;margin:2px 0 7px;position:relative}
.nyr::before{content:"";position:absolute;left:-14px;top:4px;width:8px;height:8px;border-radius:50%;background:#012851;border:2px solid #fff;box-shadow:0 0 0 1.5px #012851}
.nnode{position:relative;margin-bottom:9px}
.nnode .g{position:absolute;left:-18px;top:0;width:14px;height:14px;border-radius:50%;background:#012851;border:2px solid #fff;box-shadow:0 0 0 1.5px #012851;color:#a1dcd8;font-size:8px;display:flex;align-items:center;justify-content:center}
.nnode .lt{font-family:'Tinos',serif;font-size:10.5px;font-weight:700;color:#012851;line-height:1.2}
.nnode .lm{font-size:7.5px;color:#8a8a86}
.nwb{position:relative;margin:0 0 9px -9px;padding:5px 8px;background:#EEEBE2;border-left:2px solid #607080;border-radius:0 3px 3px 0}
.nwb .t{font-family:'Tinos',serif;font-style:italic;font-size:9.5px;color:#5a6776}
.nwb .s{font-size:7px;color:#8a8a86}
.nletter{position:relative;margin-bottom:7px}
.nletter .d{position:absolute;left:-15px;top:3px;width:6px;height:6px;border-radius:50%;background:#fff;border:1.5px solid #a1dcd8}
.ncard{background:#FAF9F5;border:1px solid #eeede8;border-radius:4px;padding:5px 8px}
.ncard .t{font-size:9px;font-weight:700;color:#1A1A1A}
.ncard .m{font-size:7px;color:#9a9a96}
.nstrip{background:#FAF9F5;border:1px solid #eeede8;border-radius:4px;padding:6px 8px;margin-bottom:7px}
.nbucket{border:1px solid #e4e2d7;border-radius:4px;padding:5px 8px;margin-bottom:5px;display:flex;align-items:center;justify-content:space-between}
/* impl-ref */
.impl-ref{background:#fff;border:1px solid #E0DDD6;border-radius:7px;overflow:hidden;margin-top:8px}
.impl-ref table{width:100%;border-collapse:collapse}
.impl-ref th{background:#012851;color:#fff;padding:8px 13px;text-align:left;font-size:8px;font-weight:800;letter-spacing:.6px;text-transform:uppercase}
.impl-ref td{padding:8px 13px;border-bottom:1px solid #F0EEE8;vertical-align:top;font-size:11px;color:#444;line-height:1.55}
.impl-ref tr:nth-child(even) td{background:#FAFAF7}
.impl-ref td:first-child{font-weight:700;color:#012851;white-space:nowrap;width:165px}
.impl-ref td code{font-size:9.5px;background:#F0EFE9;padding:1px 4px;border-radius:2px;font-family:'Courier New',monospace;color:#333}
.sw{display:inline-block;width:11px;height:11px;border-radius:2px;vertical-align:-1px;margin-right:5px;border:1px solid rgba(0,0,0,.12)}
hr{border:none;border-top:2px dashed #C8C4BE;margin:50px 0}
.note{font-size:11px;color:#888;font-style:italic;margin-top:10px;line-height:1.6}
</style>
</head>
<body>
<div class="page">
<!-- ══ MASTHEAD ══ -->
<div class="mh">
<h1>Globaler Zeitstrahl — Finale Spezifikation</h1>
<p>Kanonische Spezifikation für <code style="font-family:monospace;font-size:12px">/zeitstrahl</code> auf Basis von <strong>Konzept A „Der Lebensfaden"</strong>: eine durchgehende vertikale Achse, die Personen-Lebensereignisse, kuratierte Ereignisse und Briefe zu einer Erzählung in der Zeit verwebt. Dieselbe Komponente betreibt den globalen Zeitstrahl und den per-Person „Lebensweg". Enthält die vollständige Fall-Abdeckung (leere Jahre, wenige Briefe, hunderte Briefe, undatiert) und die drei Gruppierungs-Modi.</p>
<div class="tag-row">
<span class="tg">Milestone #14 · Zeitstrahl</span>
<span class="tg mint">Konzept A — final</span>
<span class="tg slate">Phone-first · honest DatePrecision</span>
</div>
<div class="byline">Familienarchiv · 2026-06-08 · Leonie Voss, UX Lead · ersetzt die A/B/C-Exploration zeitstrahl-global-concepts.html</div>
</div>
<!-- ══ 1 · ANATOMIE ══ -->
<div class="sh">
<h2>1 · Anatomie von Konzept A</h2>
<p>Eine Achse, sieben Bausteine. Die Zeit ist die Achse — Lebensereignisse &amp; Jahre als zentrierte Pillen <i>unterbrechen</i> den Faden (Text wird nie von der Linie gekreuzt), Welt-Ereignisse legen sich als Bänder quer, Briefe verdichten sich adaptiv.</p>
</div>
<div class="legend" style="margin-bottom:16px">
<div class="lg"><div class="ico" style="background:#012851;color:#a1dcd8"></div><div><div class="ttl">Lebensereignis-Pille</div><div class="body">Geburt <b>*</b> · Tod <b></b> · Heirat <b></b>. Abgeleitet aus <code>Person</code>-Daten. Zentriert, gefüllt — unterbricht die Achse. Glyphen aus <code>personLifeDates.ts</code>.</div></div></div>
<div class="lg"><div class="ico" style="background:#fff;border:2px solid #a1dcd8;color:#012851"></div><div><div class="ttl">Kuratierte Ereignis-Pille</div><div class="body"><code>PERSONAL</code> — Umzug, Auswanderung. Mint-Rand. Editierbar im Kurator-Editor.</div></div></div>
<div class="lg"><div class="ico" style="background:#EEEBE2;color:#607080;border:1px solid #607080"></div><div><div class="ttl">Welt-Band</div><div class="body"><code>HISTORICAL</code> — Krieg, Inflation. Gedämpftes Band quer über die Achse als Kontext.</div></div></div>
<div class="lg"><div class="ico" style="background:#fff;border:2px solid #a1dcd8"></div><div><div class="ttl">Einzel-Brief</div><div class="body">Kleiner Punkt + Karte, alternierend links/rechts. Wurzel-Tag-Farbchip. Link zu <code>/documents/[id]</code>.</div></div></div>
<div class="lg"><div class="ico" style="background:#a1dcd8;color:#012851;font-size:11px"></div><div><div class="ttl">Jahres-Strip</div><div class="body">Verdichtung dichter Jahre: Anzahl + 12-Monats-Sparkline. <code>MonthBucket</code> / <code>aggregateToYears</code>.</div></div></div>
<div class="lg"><div class="ico" style="background:#f0efe9;border:1px dashed #c4c0ba;color:#9a958c;font-size:11px"></div><div><div class="ttl">Lücke &amp; Ohne-Datum</div><div class="body">Ruhige/leere Jahre als dünne Span-Zeile gefaltet; <code>UNKNOWN</code>-Briefe im „Ohne Datum"-Eimer am Ende.</div></div></div>
</div>
<div class="callout navy"><strong>Gruppierungs-Umschalter</strong> (oben rechts im Zeitstrahl): <span style="display:inline-flex;border:1.5px solid #012851;border-radius:5px;overflow:hidden;vertical-align:middle;margin:0 4px"><span style="background:#012851;color:#fff;font-size:9px;font-weight:700;padding:3px 11px">Datum</span><span style="color:#012851;font-size:9px;font-weight:700;padding:3px 11px;border-left:1px solid #012851">Ereignis</span><span style="color:#012851;font-size:9px;font-weight:700;padding:3px 11px;border-left:1px solid #012851">Thema</span></span> steuert <b>nur, wie lose Briefe gebündelt werden</b>. Lebensereignisse, kuratierte Ereignisse und Welt-Bänder bleiben in allen Modi gleich auf der Achse. Standard = <b>Datum</b>.</div>
<!-- ══ 2 · GROUPING MODES ══ -->
<div class="sh">
<h2>2 · Die drei Gruppierungs-Modi</h2>
<p>Gleicher Ausschnitt (19141915), dreimal gerendert. Nur die <b>losen Briefe</b> ordnen sich um — die Achse bleibt stabil. Schmale Spaltenbreite = Phone-/Lebensweg-Form derselben Komponente.</p>
</div>
<div class="col3">
<!-- MODE: DATUM -->
<div>
<div class="modehd"><span class="seg">Datum</span> Chronologisch</div>
<div class="modesub">Standard. Briefe nach Datum; dichte Jahre verdichten zum Strip. Reine Zeit-Reihung.</div>
<div class="nf">
<div class="nsp">
<div class="nyr">1914</div>
<div class="nnode"><span class="g"></span><div class="lt">Heirat: Karl &amp; Elfriede</div><div class="lm">1914 · abgeleitet</div></div>
<div class="nwb"><div class="t">◍ Erster Weltkrieg</div><div class="s">19141918</div></div>
<div class="nletter"><span class="d"></span><div class="ncard"><div class="t">✉ Kriegsausbruch — Brief an die Familie</div><div class="m">Karl → Elfriede · 4. Aug 1914</div></div></div>
<div class="nyr">1915</div>
<div class="nnode"><span class="g">*</span><div class="lt">Geburt: Hans Raddatz</div><div class="lm">Sommer 1915 · abgeleitet</div></div>
<div class="nstrip">
<div style="display:flex;justify-content:space-between;align-items:center;margin-bottom:5px"><span style="font-size:9.5px;font-weight:700;color:#012851">✉ 24 Briefe</span><span style="font-size:7px;color:#8a8a86">Monate ▾</span></div>
<div style="display:flex;align-items:flex-end;gap:1.5px;height:20px"><div style="flex:1;height:20%;background:#a1dcd8"></div><div style="flex:1;height:35%;background:#a1dcd8"></div><div style="flex:1;height:55%;background:#a1dcd8"></div><div style="flex:1;height:70%;background:#a1dcd8"></div><div style="flex:1;height:60%;background:#a1dcd8"></div><div style="flex:1;height:85%;background:#a1dcd8"></div><div style="flex:1;height:100%;background:#a1dcd8"></div><div style="flex:1;height:88%;background:#a1dcd8"></div><div style="flex:1;height:72%;background:#a1dcd8"></div><div style="flex:1;height:48%;background:#a1dcd8"></div><div style="flex:1;height:55%;background:#a1dcd8"></div><div style="flex:1;height:40%;background:#a1dcd8"></div></div>
</div>
</div>
</div>
</div>
<!-- MODE: EREIGNIS -->
<div>
<div class="modehd"><span class="seg">Ereignis</span> Kuratiert</div>
<div class="modesub">Briefe bündeln unter kuratierte Ereignisse (<code>TimelineEvent.documents</code>). Erzählende Cluster statt Listen.</div>
<div class="nf">
<div class="nsp">
<div class="nyr">1914</div>
<div class="nnode"><span class="g"></span><div class="lt">Heirat: Karl &amp; Elfriede</div><div class="lm">1914 · abgeleitet</div></div>
<div class="nwb"><div class="t">◍ Erster Weltkrieg</div><div class="s">19141918</div></div>
<div class="nyr">1915</div>
<div class="nnode"><span class="g">*</span><div class="lt">Geburt: Hans Raddatz</div><div class="lm">Sommer 1915 · abgeleitet</div></div>
<div class="nletter"><span class="d"></span><div class="ncard" style="border-left:2px solid #a1dcd8"><div class="t">✉ Briefe von der Front · 24</div><div class="m">Karl ⇄ Elfriede &amp; Hans · 1915 ▾</div><span class="chip krieg" style="margin-top:4px"><i></i>Krieg</span></div></div>
<div class="nletter"><span class="d"></span><div class="ncard" style="border-left:2px solid #a1dcd8"><div class="t">✉ Weihnachten 1915 · 3</div><div class="m">kuratiertes Ereignis ▾</div><span class="chip weih" style="margin-top:4px"><i></i>Weihnachten</span></div></div>
</div>
</div>
</div>
<!-- MODE: THEMA -->
<div>
<div class="modehd"><span class="seg">Thema</span> Nach Wurzel-Tag</div>
<div class="modesub">Optional (Post-MVP). Lose Briefe je Jahr in Wurzel-Tag-Eimer; Mehrfach-Tags dedupliziert auf den primären.</div>
<div class="nf">
<div class="nsp">
<div class="nyr">1914</div>
<div class="nnode"><span class="g"></span><div class="lt">Heirat: Karl &amp; Elfriede</div><div class="lm">1914 · abgeleitet</div></div>
<div class="nwb"><div class="t">◍ Erster Weltkrieg</div><div class="s">19141918</div></div>
<div class="nbucket"><span class="chip krieg" style="margin:0"><i></i>Krieg</span><span style="font-size:8px;color:#8a8a86">6 ▾</span></div>
<div class="nyr">1915</div>
<div class="nnode"><span class="g">*</span><div class="lt">Geburt: Hans Raddatz</div><div class="lm">Sommer 1915 · abgeleitet</div></div>
<div class="nbucket"><span class="chip krieg" style="margin:0"><i></i>Krieg Briefe von der Front</span><span style="font-size:8px;color:#8a8a86">24 ▾</span></div>
<div class="nbucket"><span class="chip weih" style="margin:0"><i></i>Weihnachten</span><span style="font-size:8px;color:#8a8a86">3 ▾</span></div>
<div class="nbucket"><span class="chip fam" style="margin:0"><i></i>Familie</span><span style="font-size:8px;color:#8a8a86">2 ▾</span></div>
</div>
</div>
<div class="note" style="margin-top:8px">Hinweis im UI: „Brief mit mehreren Tags erscheint unter seinem primären Tag."</div>
</div>
</div>
<!-- ══ 3 · ALL CASES PREVIEW ══ -->
<div class="sh">
<h2>3 · Vollständige Vorschau — alle Dichte-Fälle</h2>
<p>Ein durchgehender Zeitstrahl (Desktop, zentrale Achse) von 1899 bis „Ohne Datum". Jeder Dichte-Fall kommt genau einmal vor — von leeren Jahren bis zu hunderten Briefen.</p>
</div>
<div class="callout navy" style="margin-bottom:18px">
<strong>Abgedeckte Fälle:</strong>
① leere Jahre (gefaltete Lücke) ·
② wenige Briefe ≤ 3 (einzelne Karten) ·
③ hunderte Briefe (Jahres-Strip + Sparkline) ·
④ kuratiertes Ereignis &amp; Welt-Band ·
⑤ ungenaue Präzision (<code>Sommer</code>, <code>ca.</code>) ·
⑥ undatierte Briefe (Ohne-Datum-Eimer).
</div>
<div class="tl-canvas">
<div class="dh">Zeitstrahl</div>
<div class="dh-sub">Die Familie Raddatz · 18991950 · 412 Briefe · 38 Ereignisse &nbsp;·&nbsp; <span style="color:#012851">Gruppierung: Datum</span></div>
<div style="text-align:center;margin-bottom:8px"><span class="casetag">① Leere Jahre → gefaltet</span></div>
<div class="axis">
<!-- ① empty years gap -->
<div class="gap"><span class="ln"></span><b>1899 1908</b> · keine Einträge<span class="ln"></span></div>
<!-- ② 3 letters -->
<div class="ybadge"><span>1909</span></div>
<div style="text-align:center;margin-bottom:4px"><span class="casetag">② Wenige Briefe → einzeln</span></div>
<div class="lrow"><div class="half a"><div class="lcard"><div class="t">✉ Brief aus Stettin</div><div class="m">Elfriede → Karl · Mai 1909</div><span class="chip fam"><i></i>Familie</span></div></div><div class="dot"></div><div class="half b"></div></div>
<div class="lrow"><div class="half a"></div><div class="dot"></div><div class="half b"><div class="lcard"><div class="t">✉ Geburtstagsgruß</div><div class="m">Karl → Hans · Sep 1909</div></div></div></div>
<div class="lrow"><div class="half a"><div class="lcard"><div class="t">✉ Brief zum Jahresende</div><div class="m">Karl → Elfriede · Dez 1909</div><span class="chip weih"><i></i>Weihnachten</span></div></div><div class="dot"></div><div class="half b"></div></div>
<!-- ④ life event + world band -->
<div class="ybadge"><span>1914</span></div>
<div class="pill"><span class="inner"><span class="gly"></span><span class="tx"><span class="t">Heirat: Karl &amp; Elfriede Raddatz</span><span class="s">1914 · abgeleitet aus Beziehung</span></span></span></div>
<div style="text-align:center;margin-bottom:6px"><span class="casetag">④ Welt-Band (RANGE 19141918)</span></div>
<div class="wband"><span class="t">◍ Erster Weltkrieg</span><span class="s">19141918 · historisch · 187 Briefe in dieser Zeit</span></div>
<!-- ③ hundreds of letters strip -->
<div class="ybadge"><span>1915</span></div>
<div style="text-align:center;margin-bottom:6px"><span class="casetag">③ Hunderte Briefe → Jahres-Strip</span> &nbsp; <span class="casetag" style="background:#607080">⑤ „Sommer 1915"</span></div>
<div class="pill"><span class="inner"><span class="gly">*</span><span class="tx"><span class="t">Geburt: Hans Raddatz</span><span class="s">Sommer 1915 · abgeleitet · SEASON</span></span></span></div>
<div class="strip">
<div class="hd"><span class="ct">✉ 187 Briefe</span><span class="ex">Monats-Dichte · antippen → Monate → Briefe ▾</span></div>
<div class="spark"><div style="height:22%"></div><div style="height:30%"></div><div style="height:48%"></div><div style="height:66%"></div><div style="height:58%"></div><div style="height:80%"></div><div style="height:100%"></div><div style="height:92%"></div><div style="height:74%"></div><div style="height:50%"></div><div style="height:44%"></div><div style="height:34%"></div></div>
<div class="axl"><span>Jan 1915</span><span>Dez 1915</span></div>
</div>
<!-- empty-ish span with some letters collapsed -->
<div class="gap"><span class="ln"></span><b>1916 1922</b> · Nachkriegsjahre · 96 Briefe<span class="ln"></span></div>
<!-- ⑤ approx precision world band -->
<div style="text-align:center;margin-bottom:6px"><span class="casetag" style="background:#607080">⑤ „ca. 1923" → APPROX</span></div>
<div class="wband"><span class="t">◍ Hyperinflation</span><span class="s">ca. 1923 · historisch</span></div>
<!-- curated personal event -->
<div class="ybadge"><span>1924</span></div>
<div class="pill curated"><span class="inner"><span class="gly"></span><span class="tx"><span class="t">Auswanderung nach Argentinien</span><span class="s">Frühjahr 1924 · persönlich · kuratiert</span></span></span></div>
<!-- tail empty -->
<div class="gap"><span class="ln"></span><b>1925 1950</b> · keine Einträge<span class="ln"></span></div>
</div>
<!-- ⑥ undated bucket -->
<div style="text-align:center;margin:6px 0"><span class="casetag" style="background:#7a756c">⑥ Undatiert → eigener Eimer am Ende</span></div>
<div class="undated">
<div class="h">Ohne Datum · 11 Briefe</div>
<div style="display:flex;align-items:center;gap:8px;padding:4px 0;border-top:1px solid #f0eee8"><span style="width:5px;height:5px;border-radius:50%;background:#c4c0ba;flex-shrink:0"></span><span style="font-size:9.5px;color:#3a3a36;flex:1">Brief ohne Jahresangabe</span><span style="font-size:8px;color:#aaa">Präzision UNKNOWN</span></div>
<div style="display:flex;align-items:center;gap:8px;padding:4px 0"><span style="width:5px;height:5px;border-radius:50%;background:#c4c0ba;flex-shrink:0"></span><span style="font-size:9.5px;color:#3a3a36;flex:1">Fragment, Absender unklar</span><span style="font-size:8px;color:#aaa">+ 9 weitere ▾</span></div>
</div>
</div>
<div class="note">Kein erfundenes Datum: undatierte Briefe wandern nie spekulativ in ein Jahr, sondern bleiben sichtbar im Eimer. <code>RANGE</code>-Einträge (Krieg) erscheinen einmal im Start-Jahr mit Spannen-Marker, nicht in jedem überspannten Jahr.</div>
<!-- ══ 4 · PRECISION ══ -->
<div class="sh"><h2>4 · Datums-Präzision (geteilt von Ereignissen &amp; Briefen)</h2><p>Eine Render-Logik für alle datierten Einträge — <code>dateLabel.ts</code>, gespeist von <code>DatePrecision</code>.</p></div>
<div class="rules">
<table>
<tr><th>DatePrecision</th><th>Darstellung</th><th>Beispiel</th><th>Wirkung auf der Achse</th></tr>
<tr><td>DAY</td><td>vollständiges Datum</td><td class="ex">28. Juli 1914</td><td>exakte Sortierung im Jahres-Band</td></tr>
<tr><td>MONTH</td><td>Monat + Jahr</td><td class="ex">Juli 1914</td><td>Monats-Sortierung</td></tr>
<tr><td>SEASON</td><td>Jahreszeit + Jahr</td><td class="ex">Sommer 1914</td><td>grobe Reihung</td></tr>
<tr><td>YEAR</td><td>nur Jahr</td><td class="ex">1914</td><td>ans Band-Ende</td></tr>
<tr><td>APPROX</td><td>„ca." + Jahr</td><td class="ex">ca. 1914</td><td>mit „ca."-Marker</td></tr>
<tr><td>RANGE</td><td>StartEnde</td><td class="ex">19141918</td><td>Start-Jahr, Spannen-Marker, nicht dupliziert</td></tr>
<tr><td>UNKNOWN</td><td>undatiert</td><td class="ex">Ohne Datum</td><td>eigener Eimer am Ende</td></tr>
</table>
</div>
<!-- ══ 5 · RESPONSIVE ══ -->
<div class="sh"><h2>5 · Responsiv — eine Komponente, drei Breiten</h2><p>Identisches Markup &amp; identische Daten. Nur die Achs-Position wechselt per Container-Query.</p></div>
<div class="legend" style="grid-template-columns:repeat(3,1fr)">
<div class="lg"><div class="ico" style="background:#012851;color:#a1dcd8;font-size:10px"></div><div><div class="ttl">≥ 1024px · Desktop</div><div class="body"><b>Zentrale Achse</b>, Briefe alternierend links/rechts, Welt-Bänder über volle Breite. Pillen unterbrechen die Linie.</div></div></div>
<div class="lg"><div class="ico" style="background:#012851;color:#a1dcd8;font-size:10px"></div><div><div class="ttl">&lt; 1024px · Phone</div><div class="body"><b>Linke Achse</b>, alles einseitig rechts. DOM-Reihenfolge bleibt streng chronologisch (<code>&lt;ol&gt;</code>) — Screenreader liest linear.</div></div></div>
<div class="lg"><div class="ico" style="background:#012851;color:#a1dcd8;font-size:10px"></div><div><div class="ttl">Lebensweg-Rail · 35%</div><div class="body">Gleiche linke Achse in der Personenseite (<code>&lt;TimelineView personId&gt;</code>), gefiltert auf eine Person. Rail-Tauglichkeit = Stärke von A.</div></div></div>
</div>
<!-- ══ 6 · TOKENS ══ -->
<div class="sh"><h2>6 · Design-Tokens (echte, ausgelieferte Werte)</h2><p>Aus <code>frontend/src/routes/layout.css</code>. Keine Hardcodes in der Komponente.</p></div>
<div class="impl-ref">
<table>
<tr><th>Rolle</th><th>Token</th><th>Wert</th><th>Einsatz</th></tr>
<tr><td>Achse / Knoten / Header</td><td><code>brand-navy</code></td><td><span class="sw" style="background:#012851"></span>#012851</td><td>Spine, Lebensereignis-Pillen, Jahres-Badges, Titel</td></tr>
<tr><td>Akzent / Brief-Punkt</td><td><code>brand-mint</code></td><td><span class="sw" style="background:#a1dcd8"></span>#a1dcd8</td><td>Brief-Punkte, kuratierte Pillen-Ränder, Sparkline, Dark-Mode-Auswahl</td></tr>
<tr><td>Historisch / Welt</td><td><code>tag-slate</code></td><td><span class="sw" style="background:#607080"></span>#607080</td><td>Welt-Bänder &amp; Glyphe ◍ — gedämpft</td></tr>
<tr><td>Tag-Chip-Farben</td><td><code>--c-tag-*</code> (Wurzel)</td><td><span class="sw" style="background:#5a8a6a"></span><span class="sw" style="background:#a0522d"></span><span class="sw" style="background:#c17a00"></span><span class="sw" style="background:#7a4f9a"></span></td><td>sage · sienna · amber · violet — Farbe vom Wurzel-Tag, Punkt + Label</td></tr>
<tr><td>Seite / Karte / Linie</td><td><code>canvas · surface · line</code></td><td><span class="sw" style="background:#f0efe9"></span><span class="sw" style="background:#fff"></span><span class="sw" style="background:#e4e2d7"></span></td><td>#f0efe9 · #ffffff · #e4e2d7</td></tr>
<tr><td>Text sekundär</td><td><code>text-ink-3</code></td><td><span class="sw" style="background:#6b7280"></span>#6b7280</td><td>Meta-Zeilen (4,8:1 auf weiß — AA ✓)</td></tr>
<tr><td>Schrift</td><td><code>font-serif · font-sans</code></td><td>Tinos · Montserrat</td><td>Namen/Titel serif · Labels/Chrome sans</td></tr>
<tr><td>Lebensdaten-Glyphen</td><td><code>personLifeDates.ts</code></td><td>* † ⚭</td><td>Geburt · Tod · Heirat — konsistent mit Personenkarten</td></tr>
</table>
</div>
<!-- ══ 7 · IMPL-REF ══ -->
<div class="sh"><h2>7 · Implementierungs-Referenz &amp; Barrierefreiheit</h2><p>Domain-Ordner <code>frontend/src/lib/timeline/</code>; Route <code>/zeitstrahl</code>; Backend <code>GET /api/timeline</code>.</p></div>
<div class="impl-ref">
<table>
<tr><th>Baustein</th><th>Komponente / Datei</th><th>Verantwortung</th></tr>
<tr><td>Orchestrator</td><td><code>TimelineView.svelte</code></td><td>Lädt <code>/api/timeline</code>; optionaler <code>personId</code> für globalen vs. Lebensweg-Modus; hält den Gruppierungs-Modus</td></tr>
<tr><td>Jahres-Band</td><td><code>YearBand.svelte</code></td><td>Jahres-Badge + Einträge; Lücken-Faltung ruhiger Spannen</td></tr>
<tr><td>Ereignis-Pille</td><td><code>EventCard.svelte</code></td><td>PERSONAL / HISTORICAL / abgeleitet; zentrierte Pille bzw. Welt-Band; präzisions-bewusstes Label</td></tr>
<tr><td>Brief-Karte</td><td><code>LetterCard.svelte</code></td><td>Einzel-Brief, alternierende Seite, Wurzel-Tag-Chip, Link <code>/documents/[id]</code></td></tr>
<tr><td>Jahres-Strip</td><td><code>YearLetterStrip.svelte</code></td><td>Adaptive Verdichtung ab Schwellwert; 12-Monats-Sparkline aus <code>MonthBucket</code> / <code>aggregateToYears</code> (<code>lib/document/timeline.ts</code>)</td></tr>
<tr><td>Datums-Helfer</td><td><code>dateLabel.ts</code></td><td><code>DatePrecision</code> → deutsches Label; geteilt von Ereignissen &amp; Briefen</td></tr>
<tr><td>Kurator-Editor</td><td><code>/zeitstrahl/events/new · [id]/edit</code></td><td>Ereignis anlegen/bearbeiten; Personen- + Dokument-Mehrfach-Picker (Bulk-Linking); <code>WRITE_ALL</code></td></tr>
<tr><td>Quick-Add</td><td><code>DocumentTimelineEventPicker.svelte</code></td><td>Auf <code>/documents/[id]</code>: Ereignis wählen/neu anlegen; verlinkt einen Brief</td></tr>
<tr><td>Daten-API</td><td><code>GET /api/timeline</code></td><td>Verschmilzt kuratierte + abgeleitete Ereignisse + Briefe in <code>TimelineDTO</code> (Jahres-Eimer + Ohne-Datum); Filter <code>personId · type · fromYear · toYear</code></td></tr>
<tr><td>Barrierefreiheit</td><td></td><td>Achse = <code>&lt;ol&gt;</code>, chronologische DOM-Reihenfolge; ◍ ✉ * nie nur Farbe — Glyphe + Label; 44px-Tap-Ziele; <code>prefers-reduced-motion</code>; axe in Light &amp; Dark</td></tr>
</table>
</div>
</div>
</body>
</html>

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,182 @@
# Family Timeline (Zeitstrahl) — Design Spec
**Date:** 2026-06-07
**Status:** Approved — pending implementation plan
## Problem
The archive can capture, transcribe, organize, and browse letters, but the transcribed material does not yet add up to a *story in time*. Readers (younger, phone-first) have no way to feel the family's history unfold; transcribers don't see their work become something larger. A previous attempt to derive meaning automatically (LLM search) was slow and low-quality, so the family is wary of auto-extraction from handwriting.
## Goal
A **hand-curated, year-banded vertical timeline** — the "Zeitstrahl" — that weaves three layers into one chronological view:
1. **Person life-events** derived from already-curated structured data (`Person` birth/death dates, marriage years from `PersonRelationship.fromYear`). Trusted, free, no extra entry. (Requires the Person birth/death fields to move from year-integers to date + precision — see foundational issue 1.)
2. **Hand-curated events** the family writes — both **personal** (a move, an illness, emigration) and **historical** (a war, hyperinflation). Editorially controlled, always correct.
3. **Letters**, auto-placed by their existing `documentDate`, optionally hand-linked to an event to cluster them.
Two surfaces, one component:
- **Global timeline** at `/zeitstrahl`.
- **Per-person "Lebensweg"** — the same view filtered to one person, embedded on the Person detail page.
Built for phones (vertical scroll), honest about date precision, with no fabricated dates.
### Non-goals (YAGNI)
- ❌ Auto-extracting events from transcription text — explicitly avoided; this is what makes the feature trustworthy.
- ❌ Importing an external historical-events dataset — historical events are hand-entered too.
- ❌ A map / geographic view — that is a separate future feature (B2).
- ❌ Per-derived-event hide/override toggle — deferred refinement; MVP shows all derived events.
- ❌ Day-resolution timeline axis — the axis is the **year**; finer dates only affect within-band ordering and label text.
## Core principle: the year is the axis
Most dates in the archive are year-only (birth/death/marriage years are years by nature; many letters carry `YEAR`/`APPROX` precision). Therefore:
- The timeline spine is a **sequence of year bands**. Everything for a given year lives in that band.
- **Finer ordering only when we have it.** A `DAY`-precision letter (`1923-04-12`) sorts above a `YEAR`-precision one (`1923`) *within* the 1923 band; we never invent a day we don't have.
- **An "Ohne Datum" bucket** at the end holds items with `UNKNOWN` precision.
- **Honest precision rendering** reuses the existing `DatePrecision` enum for every dated item (events and letters share one rendering path).
### Date rendering (shared by events and letters)
| `DatePrecision` | German render | Example |
|---|---|---|
| `DAY` | full date | `28. Juli 1914` |
| `MONTH` | month + year | `Juli 1914` |
| `SEASON` | season + year | `Sommer 1914` |
| `YEAR` | year only | `1914` |
| `APPROX` | "ca." + year | `ca. 1914` |
| `RANGE` | startend year | `19141918` |
| `UNKNOWN` | undated bucket | `Ohne Datum` |
A `RANGE` item is shown in its **start year's band** with a span marker; it is not duplicated across every year it covers.
## Data model
A new `timeline/` domain package on the backend (kept deliberately separate from the in-flight Lesereisen/`Geschichte` work in #750753).
### `TimelineEvent` entity
Mirrors the `Document` date model for consistency, so events and letters use one date-handling code path.
| Field | Type | Notes |
|---|---|---|
| `id` | `UUID` | `@GeneratedValue(UUID)` |
| `title` | `String` | required |
| `type` | `EventType` enum | `PERSONAL`, `HISTORICAL` |
| `eventDate` | `LocalDate` | required — most precise date known (WW1 → `1914-07-28`; vague year → `1920-01-01`) |
| `precision` | `DatePrecision` | reuse existing enum; default `YEAR` — governs rendering & whether the day matters |
| `eventDateEnd` | `LocalDate` (nullable) | only set when `precision == RANGE` |
| `description` | `TEXT` (nullable) | free-text narrative for the event |
| `persons` | ManyToMany `Person` | who the event involves; drives the per-person view & filtering |
| `documents` | ManyToMany `Document` | optional hand-linked supporting letters (the "cluster letters to an event" feature) |
| `createdBy` / `createdAt` / `updatedBy` / `updatedAt` / `version` | audit | standard entity conventions |
- `@Schema(requiredMode = REQUIRED)` on every always-populated field (`id`, `title`, `type`, `eventDate`, `precision`).
- Collections use `@Builder.Default new HashSet<>()`.
- New Flyway migration adds `timeline_events`, `timeline_event_persons`, `timeline_event_documents` join tables.
### `EventType` enum
`PERSONAL` | `HISTORICAL`. Personal events render with a person/family accent; historical events with a "world" accent and muted styling so the two layers are visually separable.
### Prerequisite: migrate `Person` birth/death to date + precision
Today `Person` stores `birthYear`/`deathYear` as `Integer`, so a known exact birthday (e.g. `1901-03-14`) has nowhere to live and derived events are stuck at year precision. This is fixed by a **foundational Person-domain migration** that the timeline depends on (and which delivers value on its own — precise dates then render on person cards, hover cards, and the Stammbaum).
**Change:** replace `birthYear`/`deathYear` (`Integer`) with:
| Field | Type | Notes |
|---|---|---|
| `birthDate` | `LocalDate` (nullable) | most precise date known |
| `birthDatePrecision` | `DatePrecision` (nullable) | `YEAR` for year-only, `DAY` for exact birthdays, etc. |
| `deathDate` | `LocalDate` (nullable) | |
| `deathDatePrecision` | `DatePrecision` (nullable) | |
**Flyway data migration:** existing `birth_year``birth_date = '{year}-01-01'`, `birth_date_precision = 'YEAR'` (same for death); then drop the year columns.
**Re-import preservation (ADR-025):** the canonical importer (`PersonRegisterImporter` / `tools/import-normalizer/persons_tree.py`) only carries the *year*. On re-import it must **not** clobber a hand-entered finer-than-`YEAR` date — if the existing precision is `DAY`/`MONTH`/`SEASON`, preserve it; only refresh from the spreadsheet year when the field is empty or still `YEAR`-from-import.
**Bounding the blast radius:** `PersonNodeDTO` keeps exposing an `Integer birthYear`/`deathYear` *derived* from the new date (`birthDate.getYear()`), so the Stammbaum layout (`familyForest.ts` et al.) is untouched. Display surfaces (person card, hover card) move to a shared precision-aware formatter — extend the existing `frontend/src/lib/person/personLifeDates.ts`. The person edit/new forms gain date inputs with a precision selector.
**Scope note:** `PersonRelationship.fromYear` (marriage year) stays `Integer`/`YEAR` for MVP — precise marriage dates are a later, parallel extension if wanted.
### Derived person-events (not persisted)
Assembled on read from the migrated `Person` data; never stored:
| Source | Derived event | `eventDate` | precision |
|---|---|---|---|
| `Person.birthDate` | *Geburt: {name}* | `Person.birthDate` | `Person.birthDatePrecision` |
| `Person.deathDate` | *Tod: {name}* | `Person.deathDate` | `Person.deathDatePrecision` |
| `PersonRelationship` `SPOUSE_OF.fromYear` | *Heirat: {A} & {B}* | `{fromYear}-01-01` | `YEAR` |
Emitted in the same DTO shape as a curated event, flagged `derived: true`, `type = PERSONAL`. They cannot be edited from the timeline (they are edited at their source: Person record / relationship). A marriage is derived once per `SPOUSE_OF` edge (symmetric edges are stored once — see existing relationship rules).
### Letters
Placed by `Document.documentDate`:
- Band = `documentDate.getYear()`; `UNKNOWN` precision → "Ohne Datum" bucket.
- Sub-ordered within a band by full date when precision allows.
- A letter may also appear under an event it's linked to (via `TimelineEvent.documents`) as a cluster, in addition to its own band placement.
## Assembly & API
A `TimelineService` merges the three layers into a year-bucketed DTO for the requested scope and filters. Layering rules apply: the service owns `TimelineEventRepository` and reaches Person/Document/Relationship data through their **services**, never their repositories.
### DTOs
- `TimelineEntryDTO` — one renderable item: `kind` (`EVENT` | `LETTER`), `eventDate`, `precision`, `eventDateEnd`, `title`, `type` (for events), `derived` flag, plus the source id (eventId / documentId) and minimal display fields (sender/receiver names for letters, linked person ids for events).
- `TimelineYearDTO``{ year: int, entries: TimelineEntryDTO[] }`.
- `TimelineDTO``{ years: TimelineYearDTO[], undated: TimelineEntryDTO[] }`.
### Endpoints
- `GET /api/timeline` — global timeline. Query params (all optional): `personId`, `generation`, `type` (`PERSONAL`/`HISTORICAL`), `fromYear`, `toYear`. The per-person "Lebensweg" is just `GET /api/timeline?personId=…` — no separate endpoint. Requires `READ_ALL`.
- `POST /api/timeline/events` — create a curated event. `@RequirePermission(Permission.WRITE_ALL)`.
- `PUT /api/timeline/events/{id}` — update. `@RequirePermission(Permission.WRITE_ALL)`.
- `DELETE /api/timeline/events/{id}` — delete. `@RequirePermission(Permission.WRITE_ALL)`.
- `GET /api/timeline/events/{id}` — fetch a single event for the edit form. Requires `READ_ALL`.
Input DTO `TimelineEventRequest` lives flat in the `timeline/` package. Errors use `DomainException.notFound/...`; **no new `ErrorCode`** is required. Run `npm run generate:api` after backend model/endpoint changes.
## Frontend
- New domain dir `frontend/src/lib/timeline/`:
- `TimelineView.svelte` — orchestrator; accepts an optional `personId` prop so the same component powers both global and per-person views.
- `YearBand.svelte` — one year section header + its entries.
- `EventCard.svelte` — renders a `PERSONAL`/`HISTORICAL`/derived event with precision-aware date label.
- `LetterCard.svelte` — compact letter row (sender → receiver, snippet/title, date), links to `/documents/[id]`.
- `TimelineFilters.svelte` — person, generation, layer toggles, year range.
- `dateLabel.ts` — the shared precision→label helper (reuse/extend `lib/document/timeline.ts` helpers like `formatTickLabel` where they fit).
- Routes:
- `/zeitstrahl` — global timeline (`+page.server.ts` loads `/api/timeline`).
- `/zeitstrahl/events/new` and `/zeitstrahl/events/[id]/edit` — curator forms, gated to `WRITE_ALL`, using the form-actions pattern.
- Person detail page gains a **Lebensweg** card section embedding `<TimelineView personId={person.id} />`.
- Styling per project conventions (card pattern, brand tokens, `font-serif` for names/titles, `BackButton`, mobile-first at 375px, dark-mode tokens).
- i18n keys added to `messages/{de,en,es}.json` (German primary).
## Testing
- Backend: `TimelineService` assembly/merge/sort/precision-bucketing (unit + `@DataJpaTest` against Postgres via Testcontainers); controller permission gating; derived-event assembly (birth/death/marriage, symmetric marriage dedup).
- Frontend: `dateLabel.ts` precision rendering; `TimelineView` global vs `personId` modes (`*.svelte.spec.ts`); filter behavior.
- Follow project test discipline: targeted single-file runs locally only; full sweep left to CI.
## Proposed issue breakdown (milestone "Zeitstrahl / Family Timeline")
Ordered so each issue is independently shippable and reviewable; later issues depend on earlier ones. Issue 1 is a standalone Person-domain improvement and a hard prerequisite for the timeline's derived events.
1. **Person birth/death → date + precision (foundational)** — replace `birthYear`/`deathYear` with `birthDate`/`deathDate` + precision on `Person`; Flyway data migration (year → `YYYY-01-01`, `YEAR`); update importer with re-import preservation rule; derive year in `PersonNodeDTO` (Stammbaum untouched); move person card / hover card to a precision-aware `personLifeDates.ts`; add date+precision inputs to person new/edit forms. Ships value on its own.
2. **Backend: `TimelineEvent` entity + migration** — entity, `EventType`, Flyway migration + join tables, repository.
3. **Backend: TimelineEvent CRUD API**`TimelineEventController` + `TimelineService` write methods, `TimelineEventRequest` DTO, permission gating, `GET /events/{id}`.
4. **Backend: derived person-events** — assemble Geburt/Tod/Heirat from migrated Person + relationship data via their services; unit-tested dedup.
5. **Backend: timeline assembly endpoint**`GET /api/timeline` merging events + derived events + letters into `TimelineDTO`; year-bucketing, precision sort, undated bucket, filters.
6. **Frontend: shared date-label helper + types**`dateLabel.ts`, regen API types.
7. **Frontend: global `/zeitstrahl` view**`TimelineView`, `YearBand`, `EventCard`, `LetterCard`, server load.
8. **Frontend: filters**`TimelineFilters` (person / generation / layer / year range).
9. **Frontend: curator event forms**`/zeitstrahl/events/new` + `/[id]/edit`, gated, with document & person pickers.
10. **Frontend: per-person Lebensweg** — embed `<TimelineView personId>` on Person detail.
11. **Polish & a11y** — mobile layout at 375px, dark mode, axe checks, i18n completeness (de/en/es).
> An ADR may be warranted for the new `timeline/` domain + entity (per `docs/CLAUDE.md`, significant data-model change). Add as the next sequential ADR number when implementation starts.

View File

@@ -0,0 +1,188 @@
# spaCy NLP Service — Design Spec
**Date:** 2026-06-07
**Status:** Prototype
## Problem
The current NL search uses Ollama (`qwen2.5:7b-instruct-q4_K_M`) to parse free-text queries into structured extractions (person names, dates, role, keywords). Inference takes 515 seconds per query, making the feature too slow to be useful compared to filling in the filter UI manually.
## Goal
Build a standalone `nlp-service/` prototype that replaces Ollama with spaCy for query parsing. The prototype is scoped to **extraction quality evaluation** — run it locally, curl it with real archive queries, and measure whether spaCy extracts names/dates/keywords well enough to justify a full migration. No Java-side changes in this iteration.
## Extraction Contract
The service must produce an output compatible with the existing `OllamaExtraction` Java record:
| Field | Type | Description |
|---|---|---|
| `personNames` | `string[]` | Names of persons mentioned, left-to-right order |
| `personRole` | `"sender"` \| `"receiver"` \| `"any"` | Role of the person(s) in the document |
| `dateFrom` | `string \| null` | ISO 8601 date `YYYY-MM-DD` or null |
| `dateTo` | `string \| null` | ISO 8601 date `YYYY-MM-DD` or null |
| `keywords` | `string[]` | Content words — fuzzy-matched against tags by Java |
| `rawQuery` | `string` | Echo of the input query |
**Two-person ordering:** `personNames` must be in left-to-right span order. Java maps `[0]` → sender, `[1]` → receiver.
**`rawQuery` note:** In the current Java code `rawQuery` is set by the caller, not parsed from Ollama. The service echoes the input for convenience; the eventual `RestClientSpacyClient` will set it from the input directly, same as today.
## Architecture
```
nlp-service/
├── main.py # FastAPI app — /parse and /health endpoints
├── extractor.py # NLP pipeline: NER → role → dates → keywords
├── models.py # Pydantic request/response types
├── requirements.txt
├── Dockerfile
└── CLAUDE.md
```
Sits alongside `ocr-service/` in the repo. For the prototype it runs standalone (no docker-compose wiring).
## Extraction Pipeline (`extractor.py`)
Five steps run in sequence on each query.
### Step 1 — NER pass
Run spaCy on the query using the model for the requested language. Collect:
- All `PER` spans → candidates for `personNames`
- All `DATE` spans → raw text strings for step 3
### Step 2 — Role detection
Only relevant when exactly **one** PER entity is found. Walk the dependency tree of the PER span's root token; check if a governing `case` or `prep` token matches the sender or receiver preposition set for the language:
| Language | Sender prepositions | Receiver prepositions |
|---|---|---|
| `de` | von, vom | an, nach, für |
| `en` | from, by | to, for |
| `es` | de, por | para, a |
- One person + sender preposition → `personRole = "sender"`
- One person + receiver preposition → `personRole = "receiver"`
- One person + no match / two or more persons → `personRole = "any"`
Two-person queries always return `"any"` — Java derives direction from position.
### Step 3 — Date parsing
For each DATE span, inspect the token immediately before the span to detect range direction:
| Direction token | Effect |
|---|---|
| vor / before / antes de | Span → `dateTo` |
| nach / after / después de | Span → `dateFrom` |
| zwischen…und / between…and / entre…y | Earlier span → `dateFrom`, later → `dateTo` |
| No direction token (bare year/date) | Span → both `dateFrom` and `dateTo` set to that year (year-range, Jan 1Dec 31) |
`dateparser.parse()` with `PREFER_DAY_OF_MONTH=first` converts the span text to a Python `date`. For `dateTo` results that resolve to a year boundary, set to Dec 31 of that year (mirrors `RestClientOllamaClient.parseDate()` behaviour).
Output as ISO strings (`YYYY-MM-DD`) or `null`.
### Step 4 — Keyword extraction
Collect tokens that satisfy all of:
- POS tag is `NOUN` or `PROPN`
- Not a stopword
- Not inside any NER span (PER or DATE)
- Lemma length ≥ 3
Output as lowercased lemmas. These are fuzzy-matched against the tags table by `NlQueryParserService.resolveTags()` on the Java side — no tag lookup in the Python service.
Examples:
- "Briefe aus dem Krieg" → `keywords: ["brief", "krieg"]`
- "Texte über Weihnachten" → `keywords: ["text", "weihnachten"]`
### Step 5 — Assembly
```json
{
"personNames": ["Opa Hermann", "Marie"],
"personRole": "any",
"dateFrom": null,
"dateTo": "1920-12-31",
"keywords": ["brief"],
"rawQuery": "Briefe von Opa Hermann an Marie vor 1920"
}
```
## API
### `POST /parse`
**Request:**
```json
{ "query": "Briefe von Opa Hermann an Marie vor 1920", "lang": "de" }
```
`lang` is a required enum: `"de"` | `"en"` | `"es"`. Unknown values → HTTP 422 (FastAPI validation).
**Response:** extraction object as above, HTTP 200.
**Error:** pipeline crash → HTTP 500 `{"detail": "..."}`.
### `GET /health`
Returns HTTP 200 `{"status": "ok"}` when all three models are loaded.
## Language Models
| `lang` | spaCy model |
|---|---|
| `de` | `de_core_news_sm` |
| `en` | `en_core_web_sm` |
| `es` | `es_core_news_sm` |
All three models are loaded at startup and held in memory. Routing is by the `lang` field on the request.
## Dockerfile
Mirrors `ocr-service/``python:3.11-slim`, non-root user, models baked into the image:
```dockerfile
FROM python:3.11-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
RUN python -m spacy download de_core_news_sm \
&& python -m spacy download en_core_web_sm \
&& python -m spacy download es_core_news_sm
COPY . .
RUN useradd --no-create-home --shell /usr/sbin/nologin --uid 1001 nlp \
&& chown -R nlp:nlp /app
USER nlp
EXPOSE 8001
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8001"]
```
Image size: ~350 MB. No volume needed — models live in the image layer.
## Local Dev
```bash
cd nlp-service
pip install -r requirements.txt
python -m spacy download de_core_news_sm en_core_web_sm es_core_news_sm
uvicorn main:app --reload --port 8001
curl -X POST http://localhost:8001/parse \
-H "Content-Type: application/json" \
-d '{"query": "Briefe von Opa Hermann an Marie vor 1920", "lang": "de"}'
```
## Known Limitations
- **Historical names:** spaCy models are trained on modern news corpora. Unusual 18991950 German names may not score as `PER`. Mitigation: the Java `resolveNames()` already does fuzzy matching against the persons table, so partial name extraction is recoverable.
- **Role detection:** the preposition sets are a fixed enumeration (~12 tokens across 3 languages). Sentences that express direction without one of these prepositions will fall through to `personRole = "any"`. This is acceptable — `"any"` is the safe default and searches both sender and receiver positions.
- **"über Oma" ambiguity:** if spaCy recognises "Oma" as a PER entity it lands in `personNames` (person search); if not, it lands in `keywords` (tag search via Java). Both paths return relevant results. The prototype evaluation will reveal which path dominates for real archive queries.
## Out of Scope (prototype)
- docker-compose integration (Ollama replacement)
- Java-side changes (`RestClientSpacyClient`, rename `OllamaClient``NlParserClient`)
- Tag lookup inside the Python service
- Automated test suite (pytest fixtures) — evaluation is done by curling the running service

View File

@@ -0,0 +1,36 @@
import { expect, test } from '@playwright/test';
/**
* Auto-title sync, full-stack happy path (#726). A document whose stored title equals its
* machine-generated auto-title must follow a date correction forward on save; a hand-edit would
* be kept. The exhaustive permutations live in the backend unit/integration suites — this is the
* single end-to-end pass, and it also asserts the FR-005 helper line is present on the edit form.
*/
test.describe('Document auto-title sync (#726)', () => {
test('editing the date rebuilds the auto-title, and the edit form explains it', async ({
page
}) => {
// 1. Create a document with no date/location, so its stored title == its auto-title
// (originalFilename only). createDocument derives originalFilename from the title.
await page.goto('/documents/new');
await page.waitForSelector('[data-hydrated]');
await page.getByLabel('Titel').fill('E2E Auto-Titel Sync');
await page.getByRole('button', { name: 'Speichern', exact: true }).click();
await expect(page).toHaveURL(/\/documents\/[^/]+$/);
const detailUrl = page.url();
// 2. The edit form carries the FR-005 helper explaining the auto-generated title.
await page.goto(`${detailUrl}/edit`);
await page.waitForSelector('[data-hydrated]');
await expect(page.locator('#title-help')).toBeVisible();
// 3. Add a YEAR-precision date WITHOUT touching the title, then save.
await page.locator('#documentDate').fill('15.01.1928');
await page.locator('#metaDatePrecision').selectOption('YEAR');
await page.getByRole('button', { name: 'Speichern', exact: true }).click();
// 4. The detail page shows the regenerated title carrying the new year.
await expect(page).toHaveURL(/\/documents\/[^/]+$/);
await expect(page.getByRole('heading', { name: /E2E Auto-Titel Sync.*1928/ })).toBeVisible();
});
});

View File

@@ -56,6 +56,7 @@
"form_label_sender": "Absender",
"form_label_receivers": "Empfänger",
"form_label_title": "Titel",
"form_helper_title_autogenerated": "Wird automatisch aus Datum und Ort gebildet — sobald du den Titel änderst, bleibt deine Version erhalten.",
"form_label_tags": "Schlagworte",
"form_label_content": "Inhalt",
"form_placeholder_content": "Kurze Beschreibung des Inhalts…",

View File

@@ -56,6 +56,7 @@
"form_label_sender": "Sender",
"form_label_receivers": "Recipients",
"form_label_title": "Title",
"form_helper_title_autogenerated": "Generated automatically from the date and place — as soon as you edit the title, your version is kept.",
"form_label_tags": "Tags",
"form_label_content": "Content",
"form_placeholder_content": "Brief description of the content…",

View File

@@ -56,6 +56,7 @@
"form_label_sender": "Remitente",
"form_label_receivers": "Destinatarios",
"form_label_title": "Título",
"form_helper_title_autogenerated": "Se genera automáticamente a partir de la fecha y el lugar; en cuanto edites el título, se conservará tu versión.",
"form_label_tags": "Etiquetas",
"form_label_content": "Contenido",
"form_placeholder_content": "Breve descripción del contenido…",

View File

@@ -17,6 +17,7 @@ let {
titleRequired = false,
suggestedTitle = '',
hideTitle = false,
showTitleHelp = false,
editMode = false
}: {
tags?: Tag[];
@@ -31,6 +32,7 @@ let {
titleRequired?: boolean;
suggestedTitle?: string;
hideTitle?: boolean;
showTitleHelp?: boolean;
editMode?: boolean;
} = $props();
@@ -72,8 +74,14 @@ const titleValue = $derived(titleDirty ? currentTitle : suggestedTitle || curren
titleDirty = true;
}}
required={titleRequired}
aria-describedby={showTitleHelp ? 'title-help' : undefined}
class="block w-full rounded border border-line p-2 text-sm shadow-sm focus:outline-none focus-visible:ring-2 focus-visible:ring-focus-ring"
/>
{#if showTitleHelp}
<p id="title-help" class="mt-1 text-sm text-ink-3">
{m.form_helper_title_autogenerated()}
</p>
{/if}
</div>
{/if}

View File

@@ -1,6 +1,7 @@
import { afterEach, describe, expect, it } from 'vitest';
import { cleanup, render } from 'vitest-browser-svelte';
import DescriptionSection from './DescriptionSection.svelte';
import { m } from '$lib/paraglide/messages.js';
afterEach(() => cleanup());
@@ -55,3 +56,28 @@ describe('DescriptionSection — onMount seeding (Felix B1/B2 fix regression fen
expect(input.value).toBe('Parent Value');
});
});
describe('DescriptionSection — auto-generated title helper (FR-TITLE-005)', () => {
it('shows the helper with the localized text and wires aria-describedby when showTitleHelp is set', async () => {
render(DescriptionSection, { showTitleHelp: true });
const help = document.querySelector('#title-help') as HTMLElement;
expect(help).not.toBeNull();
expect(help.textContent?.trim()).toBe(m.form_helper_title_autogenerated());
// ≥14px for the 60+ audience (FR-005 prefers a larger size than the 12px field hints).
expect(help.classList.contains('text-sm')).toBe(true);
const titleInput = document.querySelector('input#title') as HTMLInputElement;
expect(titleInput.getAttribute('aria-describedby')).toBe('title-help');
});
it('omits the helper by default (e.g. the new-document form)', async () => {
render(DescriptionSection, {});
expect(document.querySelector('#title-help')).toBeNull();
const titleInput = document.querySelector('input#title') as HTMLInputElement;
expect(titleInput.getAttribute('aria-describedby')).toBeNull();
});
it('omits the helper when the title field is hidden (bulk edit)', async () => {
render(DescriptionSection, { showTitleHelp: true, hideTitle: true });
expect(document.querySelector('#title-help')).toBeNull();
});
});

View File

@@ -221,6 +221,7 @@ async function handleReplaceFile(e: Event) {
initialArchiveFolder={doc.archiveFolder ?? ''}
initialSummary={doc.summary ?? ''}
titleRequired={true}
showTitleHelp={true}
/>
</form>

View File

@@ -0,0 +1,14 @@
import { describe, it, expect } from 'vitest';
import { canArmDraw, shouldDisarmDraw } from './drawCue';
describe('draw cue policy', () => {
it('arms only in edit mode', () => {
expect(canArmDraw('edit')).toBe(true);
expect(canArmDraw('read')).toBe(false);
});
it('disarms in every mode except edit', () => {
expect(shouldDisarmDraw('read')).toBe(true);
expect(shouldDisarmDraw('edit')).toBe(false);
});
});

View File

@@ -0,0 +1,20 @@
/**
* Policy for the "draw a new region" keyboard cue (the `n` shortcut) in the
* transcribe panel — issue #327.
*
* The cue is only valid while editing: `n` arms it in edit mode, and it must
* clear when a region is drawn or when the panel leaves edit mode. Pure so both
* rules are testable without mounting the page.
*/
type PanelMode = 'read' | 'edit';
/** The draw cue may only be armed while in edit mode. */
export function canArmDraw(panelMode: PanelMode): boolean {
return panelMode === 'edit';
}
/** Leaving edit mode must disarm the draw cue. */
export function shouldDisarmDraw(panelMode: PanelMode): boolean {
return !canArmDraw(panelMode);
}

View File

@@ -0,0 +1,30 @@
import { describe, it, expect } from 'vitest';
import { resolveTrainingMark, RECOGNITION_TRAINING_LABEL } from './trainingMark';
describe('resolveTrainingMark', () => {
it('is a no-op (null) when no region is active', () => {
expect(resolveTrainingMark(null, [])).toBe(null);
expect(resolveTrainingMark(null, [RECOGNITION_TRAINING_LABEL])).toBe(null);
});
it('enrols recognition training when a region is active and not yet enrolled', () => {
expect(resolveTrainingMark('ann-1', [])).toEqual({
label: RECOGNITION_TRAINING_LABEL,
enrolled: true
});
});
it('un-enrols when recognition training is already enrolled', () => {
expect(resolveTrainingMark('ann-1', [RECOGNITION_TRAINING_LABEL])).toEqual({
label: RECOGNITION_TRAINING_LABEL,
enrolled: false
});
});
it('ignores unrelated document training labels', () => {
expect(resolveTrainingMark('ann-1', ['KURRENT_SEGMENTATION'])).toEqual({
label: RECOGNITION_TRAINING_LABEL,
enrolled: true
});
});
});

View File

@@ -0,0 +1,31 @@
/**
* "Mark for training" (the `t` shortcut) decision logic — issue #327.
*
* Training enrollment is document-level — two fixed script-type chips
* (KURRENT_RECOGNITION / KURRENT_SEGMENTATION); there is no per-region training
* flag yet (that arrives with #321). `t` toggles the primary recognition
* enrollment and is a silent no-op unless a region is active, so it reads as an
* action on the region the transcriber is working on.
*
* Pure so the no-op-when-no-region guard and the enrolled flip are testable
* without mounting the page.
*/
export const RECOGNITION_TRAINING_LABEL = 'KURRENT_RECOGNITION';
export type TrainingMarkToggle = { label: string; enrolled: boolean };
/**
* Decide the recognition-training toggle for the active region, or null when no
* region is active (the `t` shortcut is then a silent no-op).
*
* @param activeAnnotationId the currently active region, or null
* @param currentLabels the document's currently enrolled training labels
*/
export function resolveTrainingMark(
activeAnnotationId: string | null,
currentLabels: readonly string[]
): TrainingMarkToggle | null {
if (!activeAnnotationId) return null;
const enrolled = !currentLabels.includes(RECOGNITION_TRAINING_LABEL);
return { label: RECOGNITION_TRAINING_LABEL, enrolled };
}

View File

@@ -84,22 +84,6 @@ export interface paths {
patch?: never;
trace?: never;
};
"/api/persons/{id}/confirm": {
parameters: {
query?: never;
header?: never;
path?: never;
cookie?: never;
};
get?: never;
put?: never;
post?: never;
delete?: never;
options?: never;
head?: never;
patch: operations["confirmPerson"];
trace?: never;
};
"/api/documents/{id}": {
parameters: {
query?: never;
@@ -708,6 +692,22 @@ export interface paths {
patch?: never;
trace?: never;
};
"/api/admin/backfill-titles": {
parameters: {
query?: never;
header?: never;
path?: never;
cookie?: never;
};
get?: never;
put?: never;
post: operations["backfillTitles"];
delete?: never;
options?: never;
head?: never;
patch?: never;
trace?: never;
};
"/api/admin/backfill-file-hashes": {
parameters: {
query?: never;
@@ -740,6 +740,22 @@ export interface paths {
patch: operations["patchFamilyMember"];
trace?: never;
};
"/api/persons/{id}/confirm": {
parameters: {
query?: never;
header?: never;
path?: never;
cookie?: never;
};
get?: never;
put?: never;
post?: never;
delete?: never;
options?: never;
head?: never;
patch: operations["confirmPerson"];
trace?: never;
};
"/api/notifications/{id}/read": {
parameters: {
query?: never;
@@ -859,7 +875,7 @@ export interface paths {
path?: never;
cookie?: never;
};
get: operations["search"];
get: operations["search_1"];
put?: never;
post?: never;
delete?: never;
@@ -1323,7 +1339,7 @@ export interface paths {
path?: never;
cookie?: never;
};
get: operations["search_1"];
get: operations["search_2"];
put?: never;
post?: never;
delete?: never;
@@ -1651,7 +1667,7 @@ export interface components {
/** Format: int32 */
deathYear?: number;
/** Format: int32 */
generation?: number | null;
generation?: number;
};
Person: {
/** Format: uuid */
@@ -1668,7 +1684,7 @@ export interface components {
/** Format: int32 */
deathYear?: number;
/** Format: int32 */
generation?: number | null;
generation?: number;
familyMember: boolean;
sourceRef?: string;
provisional: boolean;
@@ -1803,6 +1819,75 @@ export interface components {
/** Format: uuid */
targetId: string;
};
Pageable: {
/** Format: int32 */
page?: number;
/** Format: int32 */
size?: number;
sort?: string[];
};
ActivityActorDTO: {
initials: string;
color: string;
name?: string;
};
DocumentListItem: {
/** Format: uuid */
id: string;
title: string;
originalFilename: string;
thumbnailUrl?: string;
/** Format: date */
documentDate?: string;
/** @enum {string} */
metaDatePrecision: "DAY" | "MONTH" | "SEASON" | "YEAR" | "RANGE" | "APPROX" | "UNKNOWN";
/** Format: date */
metaDateEnd?: string;
sender?: components["schemas"]["Person"];
receivers: components["schemas"]["Person"][];
tags: components["schemas"]["Tag"][];
archiveBox?: string;
archiveFolder?: string;
location?: string;
summary?: string;
/** Format: int32 */
completionPercentage: number;
contributors: components["schemas"]["ActivityActorDTO"][];
matchData: components["schemas"]["SearchMatchData"];
/** Format: date-time */
createdAt: string;
/** Format: date-time */
updatedAt: string;
};
DocumentSearchResult: {
items: components["schemas"]["DocumentListItem"][];
/** Format: int64 */
totalElements: number;
/** Format: int32 */
pageNumber: number;
/** Format: int32 */
pageSize: number;
/** Format: int32 */
totalPages: number;
/** Format: int64 */
undatedCount: number;
};
MatchOffset: {
/** Format: int32 */
start: number;
/** Format: int32 */
length: number;
};
SearchMatchData: {
transcriptionSnippet?: string;
titleOffsets: components["schemas"]["MatchOffset"][];
senderMatched: boolean;
matchedReceiverIds: string[];
matchedTagIds: string[];
snippetOffsets: components["schemas"]["MatchOffset"][];
summarySnippet?: string;
summaryOffsets: components["schemas"]["MatchOffset"][];
};
CreateRelationshipRequest: {
/** Format: uuid */
relatedPersonId: string;
@@ -2188,11 +2273,6 @@ export interface components {
/** Format: int64 */
transcriptionCount: number;
};
ActivityActorDTO: {
initials: string;
color: string;
name?: string;
};
TranscriptionQueueItemDTO: {
/** Format: uuid */
id: string;
@@ -2235,25 +2315,6 @@ export interface components {
/** Format: int64 */
totalStories: number;
};
PersonSummaryDTO: {
title?: string;
/** Format: uuid */
id?: string;
displayName?: string;
firstName?: string;
lastName?: string;
/** Format: int64 */
documentCount?: number;
/** Format: int32 */
birthYear?: number;
/** Format: int32 */
deathYear?: number;
alias?: string;
notes?: string;
personType?: string;
familyMember?: boolean;
provisional?: boolean;
};
PersonSearchResult: {
items: components["schemas"]["PersonSummaryDTO"][];
/** Format: int64 */
@@ -2265,6 +2326,25 @@ export interface components {
/** Format: int32 */
totalPages: number;
};
PersonSummaryDTO: {
title?: string;
/** Format: uuid */
id?: string;
displayName?: string;
firstName?: string;
lastName?: string;
/** Format: int64 */
documentCount?: number;
notes?: string;
/** Format: int32 */
birthYear?: number;
/** Format: int32 */
deathYear?: number;
provisional?: boolean;
alias?: string;
personType?: string;
familyMember?: boolean;
};
InferredRelationshipWithPersonDTO: {
person: components["schemas"]["PersonNodeDTO"];
label: string;
@@ -2280,7 +2360,7 @@ export interface components {
/** Format: int32 */
deathYear?: number;
/** Format: int32 */
generation?: number | null;
generation?: number;
familyMember: boolean;
};
InferredRelationshipDTO: {
@@ -2433,63 +2513,6 @@ export interface components {
/** Format: int32 */
totalPages?: number;
};
DocumentListItem: {
/** Format: uuid */
id: string;
title: string;
originalFilename: string;
thumbnailUrl?: string;
/** Format: date */
documentDate?: string;
/** @enum {string} */
metaDatePrecision: "DAY" | "MONTH" | "SEASON" | "YEAR" | "RANGE" | "APPROX" | "UNKNOWN";
/** Format: date */
metaDateEnd?: string;
sender?: components["schemas"]["Person"];
receivers: components["schemas"]["Person"][];
tags: components["schemas"]["Tag"][];
archiveBox?: string;
archiveFolder?: string;
location?: string;
summary?: string;
/** Format: int32 */
completionPercentage: number;
contributors: components["schemas"]["ActivityActorDTO"][];
matchData: components["schemas"]["SearchMatchData"];
/** Format: date-time */
createdAt: string;
/** Format: date-time */
updatedAt: string;
};
DocumentSearchResult: {
items: components["schemas"]["DocumentListItem"][];
/** Format: int64 */
totalElements: number;
/** Format: int32 */
pageNumber: number;
/** Format: int32 */
pageSize: number;
/** Format: int32 */
totalPages: number;
/** Format: int64 */
undatedCount: number;
};
MatchOffset: {
/** Format: int32 */
start: number;
/** Format: int32 */
length: number;
};
SearchMatchData: {
transcriptionSnippet?: string;
titleOffsets: components["schemas"]["MatchOffset"][];
senderMatched: boolean;
matchedReceiverIds: string[];
matchedTagIds: string[];
snippetOffsets: components["schemas"]["MatchOffset"][];
summarySnippet?: string;
summaryOffsets: components["schemas"]["MatchOffset"][];
};
IncompleteDocumentDTO: {
/** Format: uuid */
id: string;
@@ -2828,6 +2851,26 @@ export interface operations {
};
};
};
deletePerson: {
parameters: {
query?: never;
header?: never;
path: {
id: string;
};
cookie?: never;
};
requestBody?: never;
responses: {
/** @description No Content */
204: {
headers: {
[name: string]: unknown;
};
content?: never;
};
};
};
getDocument: {
parameters: {
query?: never;
@@ -3184,48 +3227,6 @@ export interface operations {
};
};
};
confirmPerson: {
parameters: {
query?: never;
header?: never;
path: {
id: string;
};
cookie?: never;
};
requestBody?: never;
responses: {
/** @description OK */
200: {
headers: {
[name: string]: unknown;
};
content: {
"*/*": components["schemas"]["Person"];
};
};
};
};
deletePerson: {
parameters: {
query?: never;
header?: never;
path: {
id: string;
};
cookie?: never;
};
requestBody?: never;
responses: {
/** @description No Content */
204: {
headers: {
[name: string]: unknown;
};
content?: never;
};
};
};
createPerson: {
parameters: {
query?: never;
@@ -4117,6 +4118,26 @@ export interface operations {
};
};
};
backfillTitles: {
parameters: {
query?: never;
header?: never;
path?: never;
cookie?: never;
};
requestBody?: never;
responses: {
/** @description OK */
200: {
headers: {
[name: string]: unknown;
};
content: {
"*/*": components["schemas"]["BackfillResult"];
};
};
};
};
backfillFileHashes: {
parameters: {
query?: never;
@@ -4163,6 +4184,28 @@ export interface operations {
};
};
};
confirmPerson: {
parameters: {
query?: never;
header?: never;
path: {
id: string;
};
cookie?: never;
};
requestBody?: never;
responses: {
/** @description OK */
200: {
headers: {
[name: string]: unknown;
};
content: {
"*/*": components["schemas"]["Person"];
};
};
};
};
markOneRead: {
parameters: {
query?: never;
@@ -4443,7 +4486,7 @@ export interface operations {
};
};
};
search: {
search_1: {
parameters: {
query?: {
q?: string;
@@ -5067,7 +5110,7 @@ export interface operations {
};
};
};
search_1: {
search_2: {
parameters: {
query?: {
q?: string;

View File

@@ -1099,9 +1099,11 @@ describe('StammbaumTree keyboard tab order (#718)', () => {
const CLARA = '00000000-0000-0000-0000-0000000000a3';
const HANS = '00000000-0000-0000-0000-0000000000a4';
// Walter ↔ Eugenie (gen 0); their children Clara + Hans (gen 1). buildLayout
// sorts each generation alphabetically, so the deterministic visual order is
// Eugenie, Walter (top row) then Clara, Hans (next row).
// Walter ↔ Eugenie (gen 0); their children Clara + Hans (gen 1). The tidy-tree
// layout (#724) orders a couple's run by structural ownership (earliest birth
// year, then a deterministic id tie-break), not alphabetically — with no birth
// years here Walter (id …a1) owns the run and Eugenie sits to his right. So the
// deterministic visual order is Walter, Eugenie (top row) then Clara, Hans.
const FAMILY_EDGES = [
{
id: 'sp',
@@ -1171,7 +1173,7 @@ describe('StammbaumTree keyboard tab order (#718)', () => {
});
// Top generation left-to-right, then next generation left-to-right.
expect(nodeLabelsInDomOrder()).toEqual(['Eugenie', 'Walter', 'Clara', 'Hans']);
expect(nodeLabelsInDomOrder()).toEqual(['Walter', 'Eugenie', 'Clara', 'Hans']);
});
it('orders tab stops by rendered position regardless of input order', () => {

View File

@@ -307,9 +307,21 @@ describe('renderTranscriptionBody', () => {
expect(result).not.toMatch(/>O"Brien<\/a>/);
});
it('renders nothing when mentionedPersons is undefined-empty and no @ triggers', () => {
const result = renderTranscriptionBody('Plain old transcription text.', []);
expect(result).toBe('Plain old transcription text.');
it('renders a deleted-person @mention as plain text with no dead link (graceful degradation)', () => {
// AC-6 (#684): when a mentioned person is deleted, V71's ON DELETE CASCADE removes the
// sidecar row, so the displayName reaches the renderer with an empty mentionedPersons
// array. The reader must still see the name as plain text — never a dead <a>, a
// person-mention class, a data-person-id, or an href. This locks the degradation
// contract so a future renderer refactor cannot silently reintroduce a dead link.
const result = renderTranscriptionBody('Brief an @Auguste Raddatz vom Mai', []);
// (a) the reader still sees the readable name and the surrounding sentence verbatim
expect(result).toContain('Auguste Raddatz');
expect(result).toBe('Brief an @Auguste Raddatz vom Mai');
// (b) none of the anchor artifacts leak through
expect(result).not.toContain('<a');
expect(result).not.toContain('person-mention');
expect(result).not.toContain('data-person-id');
expect(result).not.toContain('href');
});
it('skips substitution when personId is not a UUID (defense in depth)', () => {

View File

@@ -81,9 +81,9 @@ $effect(() => {
onblur={onblur}
aria-label={m.docs_search_placeholder()}
placeholder={m.docs_search_placeholder()}
class="block w-full border-line py-2.5 pr-10 pl-3 placeholder-ink-3 shadow-sm focus:outline-none focus-visible:ring-2 focus-visible:ring-focus-ring"
class="block w-full border-line py-2.5 pr-4 pl-10 placeholder-ink-3 shadow-sm focus:outline-none focus-visible:ring-2 focus-visible:ring-focus-ring"
/>
<div class="pointer-events-none absolute inset-y-0 right-0 flex items-center pr-3">
<div class="pointer-events-none absolute inset-y-0 left-0 flex items-center pl-3">
{#if isLoading}
<svg
role="status"

View File

@@ -289,19 +289,19 @@ $effect(() => {
from={from}
to={to}
onchange={(event) => {
from = event.from;
to = event.to;
// Drag commits filter + zoom atomically (Graylog-style range selector).
// Single click and clear omit zoomFrom/zoomTo so existing zoom is preserved.
if ('zoomFrom' in event) {
triggerSearchWithZoom(event.zoomFrom ?? null, event.zoomTo ?? null);
} else {
triggerSearchKeepZoom();
}
}}
from = event.from;
to = event.to;
// Drag commits filter + zoom atomically (Graylog-style range selector).
// Single click and clear omit zoomFrom/zoomTo so existing zoom is preserved.
if ('zoomFrom' in event) {
triggerSearchWithZoom(event.zoomFrom ?? null, event.zoomTo ?? null);
} else {
triggerSearchKeepZoom();
}
}}
onzoomchange={(event) => {
triggerSearchWithZoom(event?.zoomFrom ?? null, event?.zoomTo ?? null);
}}
triggerSearchWithZoom(event?.zoomFrom ?? null, event?.zoomTo ?? null);
}}
/>
</div>

View File

@@ -13,6 +13,8 @@ import { transcribeShortcuts } from '$lib/shared/actions/transcribeShortcuts';
import { createOcrJob } from '$lib/ocr/useOcrJob.svelte';
import { createTranscriptionBlocks } from '$lib/document/transcription/useTranscriptionBlocks.svelte';
import { stepRegion } from '$lib/document/transcription/regionNavigation';
import { resolveTrainingMark } from '$lib/document/transcription/trainingMark';
import { canArmDraw, shouldDisarmDraw } from '$lib/document/transcription/drawCue';
import { createFileLoader } from '$lib/document/viewer/useFileLoader.svelte';
import { scrollToCommentFromQuery } from '$lib/shared/utils/deepLinkScroll';
import { getConfirmService } from '$lib/shared/services/confirm.svelte';
@@ -112,17 +114,9 @@ function toggleMode() {
if (canWrite) panelMode = panelMode === 'read' ? 'edit' : 'read';
}
// Training enrollment is document-level — two fixed script-type chips
// (KURRENT_RECOGNITION / KURRENT_SEGMENTATION); there is no per-region training
// flag (that would arrive with #321). "t" toggles the primary recognition
// enrollment and stays a no-op unless a region is active, so it reads as an
// action on the region the transcriber is working on.
const RECOGNITION_TRAINING_LABEL = 'KURRENT_RECOGNITION';
function toggleTrainingMark() {
if (!activeAnnotationId) return;
const enrolled = !(doc.trainingLabels ?? []).includes(RECOGNITION_TRAINING_LABEL);
transcription.toggleTrainingLabel(RECOGNITION_TRAINING_LABEL, enrolled);
const toggle = resolveTrainingMark(activeAnnotationId, doc.trainingLabels ?? []);
if (toggle) transcription.toggleTrainingLabel(toggle.label, toggle.enrolled);
}
function deleteCurrentRegion() {
@@ -131,7 +125,7 @@ function deleteCurrentRegion() {
// Disarm the draw cue whenever we leave edit mode.
$effect(() => {
if (panelMode !== 'edit') drawArmed = false;
if (shouldDisarmDraw(panelMode)) drawArmed = false;
});
const shortcutOptions = {
@@ -142,7 +136,9 @@ const shortcutOptions = {
goToPrevRegion: () => goToRegion(-1),
toggleMode,
closePanel: () => (transcribeMode = false),
startDrawMode: () => (drawArmed = true),
startDrawMode: () => {
if (canArmDraw(panelMode)) drawArmed = true;
},
toggleTrainingMark,
deleteCurrentRegion,
openCheatsheet: () => (cheatsheetOpen = true)

View File

@@ -20,4 +20,5 @@ scrape_configs:
- job_name: ocr-service
metrics_path: /metrics
static_configs:
- targets: ['ocr:8000']
- targets: ['ocr-service:8000']