eb9997326f
Shell script run_retrieve.sh for non-LLM
...
gneration queries (returns only chunks), track
development notes and README.
2026-03-01 07:39:28 -05:00
d034fdaad0
Fix deploy script to only stage listed public files
...
Prevents untracked files (output_test.txt, etc.) from being included.
2026-02-27 05:58:52 -05:00
d3113a9987
Add deploy_public.sh and LICENSE to main
...
deploy_public.sh automates publishing to the Forgejo public branch:
copies public files from main, strips private sections from README,
commits, and force-pushes.
2026-02-27 05:58:15 -05:00
afdd6ae7e0
Move clippings store into clippings_search/store_clippings/
...
Rename storage_clippings/ to clippings_search/store_clippings/ to keep
the experimental clippings search self-contained in its subdirectory.
2026-02-26 16:53:36 -05:00
13ec5b45d8
Remove vector store size from README and CLAUDE.md
2026-02-26 16:43:43 -05:00
13785d667a
Rename storage_exp/ to store/, remove unused storage/
...
Update all active scripts, .gitignore, CLAUDE.md, and README.md.
Also fix stale filename references in script header comments.
2026-02-26 16:36:57 -05:00
5a3294f74c
Reorganize project: rename scripts, archive superseded, add clippings_search/
...
- Rename build_exp_claude.py → build_store.py
- Rename query_hybrid_bm25_v4.py → query_hybrid.py
- Rename retrieve_hybrid_raw.py → retrieve.py
- Archive query_topk_prompt_engine_v3.py (superseded by hybrid)
- Archive retrieve_raw.py (superseded by hybrid)
- Move build_clippings.py, retrieve_clippings.py → clippings_search/
- Update run_query.sh, README.md, CLAUDE.md for new names
2026-02-26 16:24:32 -05:00
Eric
b4bf89ce4b
Built semantic search over clippings files.
...
Embedded text as a ChromaDB to learn that.
Updated requirements to include new depenendencies
2026-02-22 07:48:48 -05:00
Eric
c47c5e5c4f
Fix path metadata consistency in incremental vector store updates
...
Prepend "./" to file paths passed to SimpleDirectoryReader during
incremental updates so source document metadata matches full rebuilds.
Also add API_key_temp to .gitignore.
2026-02-20 07:32:06 -05:00
Eric
e9fc99ddc6
Initial commit: RAG pipeline for semantic search over personal journal archive
...
Vector search with cross-encoder re-ranking, hybrid BM25+vector retrieval,
incremental index updates, and multiple LLM backends (Ollama local, OpenAI API).
2026-02-20 06:02:28 -05:00
