mirror of
https://github.com/lllin000/PaperForge.git
synced 2026-07-22 06:50:53 +00:00
Phase 0 (#35): _needs_derived_rebuild returns (False, 'legacy_ocr') for papers without structured blocks and derived_version. _can_rebuild adds has_structured check. No misleading rebuild suggestions for legacy users. Phase 1 (#36): migrate_chroma_to_vec0() copies existing ChromaDB vectors to vec0 tables with zero API calls. Idempotent (skips by paper_id). delete_paper_vectors() cleans both ChromaDB and vec0. CLI: embed migrate. 5 new tests. Phase 3a (#39): Search bar in collection panel — two-mode input (M metadata / @ deep) with card results display. JSON output parsed from search/retrieve CLI. 289 lines TypeScript, 175 lines CSS. Build passes clean. Co-authored-by: Phase0LegacyOcr <agent> Co-authored-by: ChromaMigration <agent> Co-authored-by: SearchBarUI <agent> |
||
|---|---|---|
| .. | ||
| exports | ||
| ocr-complete/TSTONE001 | ||
| ocrtest_vault | ||
| TestZoteroData | ||
| batch_check.py | ||
| deep_inspect.py | ||
| generate_ocr_fixture.py | ||
| generate_sandbox.py | ||
| precise_batch.py | ||