refactor(search): remove NLP/smart-search feature entirely (#772)

## 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

docs/GLOSSARY.md

@@ -165,23 +165,7 @@ _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/`.
 
----
-
-## NL Search Terms
-
-**NlSearch** — the natural-language document search feature. Users type a plain-German query (e.g. "Was hat Walter im Krieg an Emma geschrieben?"); the backend parses it via Ollama, resolves person names to database UUIDs, and delegates to the standard `DocumentService.searchDocuments()` path. Endpoint: `POST /api/search/nl`.
-
-**NlQueryInterpretation** — the structured result of parsing a natural-language query. Contains: `resolvedPersons` (persons whose names unambiguously matched one DB record), `ambiguousPersons` (all candidates when a name matched more than one person), `keywords` (LLM-extracted search terms), `dateFrom`/`dateTo` (extracted date range), `rawQuery` (the original user input), `keywordsApplied` (whether keyword FTS was used), `resolvedTags` (tags matched by keyword→tag resolution), and `tagsApplied` (whether the OR-union tag filter was applied).
-
-**keyword→tag resolution** — the post-Ollama step in `NlQueryParserService` where each LLM-extracted keyword is substring-matched against the tag taxonomy via `TagService.findByNameContaining()`. Keywords that hit one or more tags are removed from the FTS text list and become an OR-union tag filter; keywords with no match remain as FTS text. Matching is case-insensitive and traverses the tag hierarchy via the recursive CTE `findDescendantIdsByName`. See ADR-033.
-
-**PersonHint** — a lightweight `{id, displayName}` pair used in `NlQueryInterpretation` to describe a resolved or ambiguous person without exposing the full `Person` entity to the frontend.
-
-**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"). The vocabulary is deliberately match strength, not the search layer's resolved/ambiguous buckets — `NlQueryParserService` maps one direct → resolved (auto-select), ≥2 direct → ambiguous, partial-only → ambiguous suggestions ("Meintest du …?"), and no candidates → folded into full-text search.
-
-**TagHint** — a lightweight `{id, name, color?}` triple used in `NlQueryInterpretation.resolvedTags` to describe a tag matched by keyword→tag resolution. `color` is the tag's effective color (one-level inheritance from parent when the tag has no own color), or null if neither tag nor parent has a color.
-
-**theme chip** `[frontend]` — a removable chip rendered in `InterpretationChipRow` for each entry in `NlQueryInterpretation.resolvedTags` when `tagsApplied` is `true`. Displays "Thema: {tag.name}" (prefix varies by locale). Clicking × removes the tag from the OR-union filter and navigates to `/documents?tag=…&tagOp=OR` with remaining tag and person parameters preserved.
+**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").
 
 ---