Commit Graph

25 Commits

Author SHA1 Message Date
Marcel
dff203d526 feat(transcribe): wire coach into TranscriptionEditView, hide training footer when empty
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-24 21:06:50 +02:00
Marcel
ff0bb89297 refactor(autosave): rename flushViaBeacon → flushOnUnload; add void to fire-and-forget fetch
The sendBeacon name was misleading after switching to keepalive fetch.
Also adds a test to confirm flush is a no-op when pendingTexts is empty.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-23 07:12:21 +02:00
Marcel
8898863a48 refactor(transcription): extract useBlockAutoSave and useBlockDragDrop from TranscriptionEditView (#199)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-15 14:45:03 +02:00
Marcel
8f6e398af7 fix(i18n): replace hardcoded German training label chip strings with Paraglide keys
TranscriptionEditView rendered 'Kurrent-Erkennung' and 'Segmentierung'
as hardcoded German strings, breaking the en/es locales. Added
training_chip_kurrent and training_chip_segmentation keys to all three
message files and wired them up via m.training_chip_kurrent() /
m.training_chip_segmentation().

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-14 10:04:52 +02:00
Marcel
5afdc37653 feat(ui): manual-first OCR workflow — remove full-page auto-segmentation
Drawing annotations is now the primary workflow. OCR only runs on
manually drawn regions (guided mode always). Full-page layout detection
and the useExistingAnnotations checkbox are removed entirely.

- OcrTrigger: guided-only, disabled with hint when no annotations exist
- TranscriptionEditView: empty state shows draw-regions instruction,
  OCR trigger moved out of collapsible and shown inline after block list
- i18n: add ocr_trigger_no_annotations, ocr_section_heading,
  transcription_empty_draw_hint; remove ocr_use_existing_annotations keys

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 22:24:50 +02:00
Marcel
ee58b63517 feat(ocr): add guided OCR mode using existing annotation regions
When a document has manually drawn annotation boxes, the user can now
enable "Nur annotierte Bereiche" in the OCR trigger panel. The engine
skips layout detection entirely and runs recognition only within the
pre-drawn bounding boxes, preserving manual transcription blocks.

- Python: adds OcrRegion model, extend OcrRequest/OcrBlock; guided
  branch in /ocr/stream groups by page and crops each region
- Engines: add extract_region_text() to both Kraken and Surya
- Java: adds OcrBlockResult.annotationId, OcrClient.OcrRegion,
  TriggerOcrDTO.useExistingAnnotations; OcrAsyncRunner dispatches to
  upsertGuidedBlock when annotationId is present; OcrService threads
  the flag through to runSingleDocument
- TranscriptionService: adds upsertGuidedBlock (creates, updates OCR,
  or preserves MANUAL blocks)
- Frontend: guided OCR toggle in OcrTrigger shown when blocks exist;
  skips destructive-replace confirmation in guided mode

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 15:57:54 +02:00
Marcel
9b2f91ee59 feat(training): add segmentation training pipeline and complete Part 6
- Add /segtrain endpoint to OCR service (ZIP upload, ketos.segtrain,
  backup rotation, in-process model reload)
- Add segtrainModel() to OcrClient and RestClientOcrClient (10-min timeout,
  X-Training-Token header)
- Add SegmentationTrainingExportService: PAGE XML export with polygon
  de-normalization and per-page PNG rendering via PDFBox
- Add GET /api/ocr/segmentation-training-data/export endpoint
- Make TranscriptionBlock.text nullable for segmentation-only blocks
  (V31 migration)
- Add Paraglide i18n translation keys for all training UI strings (de/en/es)
- Pass source prop from TranscriptionEditView to TranscriptionBlock

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 15:15:17 +02:00
Marcel
fdf1eb92ad feat(training): add document-level training enrollment
- V29 migration: document_training_labels join table
- TrainingLabel enum: KURRENT_RECOGNITION, KURRENT_SEGMENTATION
- Document.trainingLabels @ElementCollection
- DocumentService.addTrainingLabel / removeTrainingLabel
- PATCH /api/documents/{id}/training-labels (WRITE_ALL)
- Auto-enroll on Kurrent OCR trigger (OcrService.startOcr)
- TranscriptionEditView: enrollment chips in panel footer
- JPQL queries updated to use MEMBER OF trainingLabels

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 14:30:51 +02:00
Marcel
73229077be feat(transcription): add sticky review progress counter to TranscriptionEditView
Shows 'X / Y geprüft' with a brand-mint progress bar at the top of the
transcription panel. Derived from the blocks prop — no extra state.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 13:59:35 +02:00
Marcel
3fe6eedffb feat(ocr): allow re-running OCR when transcription blocks already exist
Add a collapsible OCR trigger below the block list in edit mode.
Uses a <details> element so it's unobtrusive — the primary workflow
is editing existing blocks, but users can expand to re-run OCR with
a confirmation dialog that warns about replacing existing blocks.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 10:37:51 +02:00
Marcel
8dc9243add feat(frontend): wire OCR trigger + review toggle into transcription panel
Some checks failed
CI / Unit & Component Tests (push) Failing after 1s
CI / Backend Unit Tests (push) Failing after 1s
CI / Unit & Component Tests (pull_request) Failing after 1s
CI / Backend Unit Tests (pull_request) Failing after 1s
- OcrTrigger component rendered in the transcription empty state when
  the document has a file and user has write permission
- Review checkmark toggle on each TranscriptionBlock (turquoise when
  reviewed, muted outline when not). Calls PUT .../review to toggle.
- TranscriptionBlockData type: added source + reviewed fields
- +page.svelte: triggerOcr() and reviewToggle() functions wired up
- Paraglide translations (de/en/es) for review toggle + reviewed count

All 687 frontend tests pass.

Refs #226, #230

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-12 22:02:56 +02:00
Marcel
ef11cbee4e feat(transcription): clicking annotation focuses corresponding block
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
Pass activeAnnotationId to TranscriptionEditView. An $effect watches
it and sets activeBlockId to the block matching the annotation,
activating its turquoise focus border.

2 new tests (RED/GREEN):
- activates block matching activeAnnotationId (turquoise border)
- no block activated when activeAnnotationId is null

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 23:36:06 +02:00
Marcel
c22f2e41b1 fix(transcription): replace broken HTML5 drag with pointer-based drag
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
HTML5 drag-and-drop didn't work — the grip handle couldn't initiate
drag properly. Replace with pointer event-based drag:

- Grip handle pointerdown starts drag, captures pointer
- Pointermove tracks offset, shows floaty style (shadow, scale, ring)
- Turquoise drop indicator line appears between blocks at cursor position
- Pointerup finalizes: reorders array and calls PUT /reorder endpoint

Visual feedback:
- Dragged block: shadow-xl, ring-2 ring-turquoise/40, scale 1.02, opacity 0.9
- Drop indicator: turquoise h-1 rounded bar between blocks

6 new TranscriptionEditView tests:
- renders blocks in sort order
- shows next-block CTA
- shows empty state
- move-up disabled on first block
- move-down disabled on last block
- drag handle present on each block

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 23:07:42 +02:00
Marcel
7d2d615e0c feat(transcription): add drag-and-drop + arrow button reordering
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
TranscriptionBlock:
- Desktop: grip handle (⠿) on left side, serves as drag handle
- Mobile (<768px): ▲/▼ arrow buttons (44px tap targets) replace grip
- isFirst/isLast disable boundary arrows
- onMoveUp/onMoveDown callbacks for arrow button clicks

TranscriptionEditView:
- HTML5 drag-and-drop on block wrappers (only initiates from grip handle)
- Dragged block shows 40% opacity
- On drop: reorder array and call PUT /reorder endpoint
- Arrow handlers: swap adjacent blocks and call reorder endpoint

5 new tests:
- drag handle element present
- move-up disabled when isFirst
- move-down disabled when isLast
- onMoveUp fires on click
- onMoveDown fires on click

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 23:00:52 +02:00
Marcel
4a88b3ba82 feat(transcription): add dashed next-block CTA below block list
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
Shows a muted dashed-outline box after the last block:
"Markiere eine weitere Passage im Scan, um Block N anzulegen"
Guides new users on how to create additional blocks.

Matches the spec's empty block CTA design (S1, bottom of block list).
i18n key transcription_next_block_cta added for de/en/es.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 22:52:15 +02:00
Marcel
351f31b183 feat(comments): inline edit on click + trash icon for own comments
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
Own comments:
- Click the text to open inline edit (textarea replaces text)
- Enter saves, Escape cancels
- Small trash icon always visible in bottom-right corner
- Hover on text shows cursor-text + subtle bg highlight

Other people's comments: read-only, no edit/delete affordances.

Re-add currentUserId prop chain for ownership check.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 22:42:24 +02:00
Marcel
f09b605752 refactor(comments): flat compact comment thread matching spec design
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
Rework CommentThread.svelte to match the annotation-transcription spec:
- Flat message list (no nested reply threading)
- Compact inline style: orange left border, tinted background
- Chat bubble icon (💬) with comment count header
- Avatar circles with author initials
- Quoted text extracted and rendered as italic left-bordered snippet
- Simple MentionEditor input at bottom (keeps @mention support)
- Removed: reply-to-specific threading, edit/delete buttons, nesting

Remove dead components no longer used after annotate mode removal:
- AnnotationCommentPanel, AnnotationSidePanel, AnnotateHintStrip
- PanelDiscussion, PanelHistory, PanelMetadata, PanelTranscription
- Associated spec files

Simplify prop chain: remove currentUserId, canAdmin, targetCommentId
from CommentThread, TranscriptionBlock, TranscriptionEditView.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 22:18:24 +02:00
Marcel
be4f1ed73b fix(transcription): always show comment list, compose box on demand
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Has been cancelled
CI / Backend Unit Tests (pull_request) Has been cancelled
CI / E2E Tests (pull_request) Has been cancelled
Comments were only visible after clicking "Kommentieren". Now:
- Comment list always renders (CommentThread with loadOnMount=true)
- Compose box hidden by default (showCompose prop on CommentThread)
- Clicking "Kommentieren" sets commentOpen=true → shows compose box
- Closing hides compose box but comments remain visible

This separates "viewing comments" (always) from "writing a comment"
(on demand via Kommentieren button).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 22:02:15 +02:00
Marcel
6475ebcc60 fix(transcription): auto-expand comment thread when block has comments
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Failing after 1m20s
CI / Backend Unit Tests (pull_request) Failing after 2m38s
CI / E2E Tests (pull_request) Has been cancelled
Comments were only shown after clicking "Kommentieren". Now:
- Load comment counts per block when blocks are loaded
- Pass commentCount prop to TranscriptionBlock
- If commentCount > 0, the comment thread is expanded by default
- If commentCount is 0, thread stays collapsed behind the button

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 21:50:37 +02:00
Marcel
8c26876345 feat(transcription): add block-level comment threads with quote support
Some checks failed
CI / Unit & Component Tests (push) Has been cancelled
CI / Backend Unit Tests (push) Has been cancelled
CI / E2E Tests (push) Has been cancelled
CI / Unit & Component Tests (pull_request) Failing after 1m33s
CI / Backend Unit Tests (pull_request) Failing after 2m47s
CI / E2E Tests (pull_request) Failing after 19m44s
TranscriptionBlock.svelte:
- "Kommentieren" button opens expandable comment thread per block
- Text selection in textarea captured as quoted text (> "...") prefix
- Quote hint "Text markieren für Zitat" shown when block is active/focused
- Comment thread uses existing CommentThread with blockId prop

CommentThread.svelte:
- Add blockId prop for block-level comments URL routing
- Add quotedText prop — pre-fills comment input with markdown blockquote
- commentsBase now supports 3 URL patterns: document, annotation, block

TranscriptionEditView.svelte:
- Pass canComment + currentUserId through to block components

3 new frontend tests:
- Kommentieren button present
- Quote hint shown when active
- Quote hint hidden when inactive

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 21:05:39 +02:00
Marcel
052f70e871 fix(transcription): use navigator.sendBeacon for beforeunload save
Replace async executeSave in beforeunload handler with
navigator.sendBeacon — synchronous and reliable for page unload.
Sends pending text as JSON blob to the block update endpoint.

Fixes @Sara: "beforeunload handlers cannot reliably await async"

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 20:28:37 +02:00
Marcel
a3fbcf346b fix(ui): semantic turquoise tokens, badge styling, saved fade animation
- Add turquoise/turquoise-fg semantic color tokens to layout.css
  (light + dark mode), replacing all hardcoded #00C7B1 in components
- Bump Details toggle from text-xs to text-sm for visual hierarchy
- Block badge: navy → turquoise, overlapping top-left card border
  with absolute positioning to visually link PDF annotation badges
- Saved indicator: smooth 300ms opacity fade before removal
  (new 'fading' state in SaveState type)
- Transcribe buttons: use border-turquoise/bg-turquoise/text-turquoise-fg

Fixes @Leonie concerns: toggle visual weight, semantic tokens,
badge styling, saved fade animation

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 20:26:41 +02:00
Marcel
b21778b3d1 refactor(types): extract TranscriptionBlockData to shared types
Move duplicated type definition from TranscriptionEditView.svelte
and +page.svelte into $lib/types.ts for single source of truth.

Fixes @Felix: "Consider extracting the TranscriptionBlockData type"

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 20:22:35 +02:00
Marcel
6463a32dfc fix: address PR review feedback — security, architecture, dead code
Fixes from PR #178 review:

Migration fixes:
- V18/V19: fix FK references from app_users to users (correct table name)
- V18: change annotation_id FK from ON DELETE CASCADE to ON DELETE RESTRICT
  (block is aggregate root, cascade flows from block, not annotation)

Backend fixes:
- TranscriptionService.deleteBlock(): remove userId param, delete block first
  then annotation directly via repository (no ownership check — block owns annotation)
- TranscriptionService.sanitizeText(): remove flawed regex HTML stripping,
  textarea content is plain text by design — just enforce max length
- TranscriptionBlockController.requireUserId(): throw DomainException.unauthorized()
  instead of silently returning null on auth failure
- CreateTranscriptionBlockDTO: add @Min/@Positive validation on coordinates
- Add @Slf4j logging to TranscriptionService for create/delete operations

Frontend fixes:
- Delete DocumentBottomPanel.svelte entirely (issue #175 requirement)
- Remove redundant mode exclusivity $effect (handled at toggle call sites)
- Remove dead handleCommentClick + onCommentClick prop (comments are future work)
- Remove quote hint UI (depends on comment feature)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 11:43:35 +02:00
Marcel
1efd3d8e23 feat(transcription): add frontend transcription editing UI (#176)
Some checks failed
CI / Unit & Component Tests (push) Failing after 1m27s
CI / Backend Unit Tests (push) Failing after 2m40s
CI / E2E Tests (push) Failing after 4m44s
CI / Unit & Component Tests (pull_request) Failing after 1m21s
CI / Backend Unit Tests (pull_request) Failing after 2m27s
CI / E2E Tests (pull_request) Failing after 4m47s
TranscriptionBlock.svelte: editable block card with auto-resize textarea,
  per-block save indicator, turquoise focus border, delete with confirmation
TranscriptionEditView.svelte: right panel with sorted block list,
  debounced auto-save (1.5s), beforeunload flush, empty state CTA
DocumentTopBar: add Transcribe/Done toggle with turquoise styling,
  mode exclusivity (transcribe and annotate mutually exclusive)
Document detail page: split view in transcribe mode (PDF left, blocks right),
  load/save/delete blocks via fetch, block focus syncs to annotation

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-05 11:34:01 +02:00