Commit History
-
[KM-556][KM-562] fix tabular catalog pipeline: trigger wiring, parquet upload, lazy Azure init, backfill script
- Wire on_db_registered in /database-clients/{id}/ingest endpoint so DB registration now builds the catalog (was silently skipped before). - CSV/XLSX no longer ingested into vector store (embedding wasted); parquet upload preserved via _upload_parquet() so TabularExecutor can still read blobs at query time. - on_document_uploaded implemented (was raise NotImplementedError); opens its own AsyncSessionLocal session, calls document_pipeline.process(). - Lazy-init Azure/PGVector clients in retrieval/document.py via @functools.cache getters; RetrievalRouter holds DocumentRetriever lazily. FastAPI no longer crashes at startup if embedding env vars are missing. - _normalize_chunks in chat_handler.py now handles RetrievalResult objects (previously silently dropped, causing empty context for unstructured path). - Cache invalidation: retrieval_router.invalidate_cache(user_id) called after document process and delete; Redis failure is logged, not raised. - Add scripts/build_initia…59b14af
-
[KM-553] migrate retrieval layer and remove obsolete rag/tools modules
- Replace src/rag/ with src/retrieval/: implement DocumentRetriever (MMR/cosine/euclidean/manhattan), simplified RetrievalRouter (unstructured-only, no schema leg, Redis cache preserved), and shared RetrievalResult/BaseRetriever base - Remove src/tools/ (orphaned LangChain @tool wrapper, never called by production code) - Update RetrievalResult imports in chat.py, query/base.py, query/executors/db_executor.py, query/executors/tabular.py, query/query_executor.py from src.rag.base to src.retrieval.base - Wire chat.py to new retrieval_router (aliased as retriever, no call-site changes) - Delete dead stubs: src/query/service.py, src/models/user_info.py, src/pipeline/document_pipeline.py (flat, shadowed by subfolder) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>df220ea -
initial structure
efc0c0a