Merge branch 'main' into deploy

M5-B result: UniXcoder dense +4.5pp MRR over MiniLM (fair code model)

M5: embedding-experiment infra + sub-exp A null result + UniXcoder (B) setup

merge dropdown-change fix into deploy

fix(ui): re-run a column when its retriever dropdown changes

merge gold-answer UI into deploy

feat(ui): highlight gold answer + rank for eval queries

M4: README closeout — live demo URL, mark M4 done

merge dataset-name fix into deploy

fix(data): use namespaced dataset id code-search-net/code_search_net

deploy: bundle slim BM25 index for HF Space (LFS)

M4: Space deploy prep — omit gradio from requirements, LFS-track slim index

M3.2: swap in full-eval numbers, drop n=2000 caveat

M4: non-indexing Gradio app + deps refresh

M4: slim BM25 index (index-only save/load with id-order guard)

M3: Hybrid RRF + MS-MARCO cross-encoder reranker

Dense: minimal Qdrant payload + local query-vector cache

Parallel embed/upsert with .npy vector cache; Qdrant reap resilience

M3 design probe: Recall@K of candidate pools (BM25, Dense, RRF, set-union)

M2 done-criterion: qualitative BM25 vs dense comparison

M2 concept checkpoint: HNSW ef_search sweep

Surface dense-vs-BM25 deltas in README; correct stale entries

M2 complete: dense retrieval via MiniLM + Qdrant, MRR@10=0.3891 Recall@100=0.7520

Update PLAN.md and README with full-corpus BM25 results

M1 complete: BM25 baseline eval with correct code-token indexing

M0 complete: hello world end-to-end, BM25 + dense side-by-side deployed

Simplify requirements.txt for HF Spaces compatibility

Fix HF Spaces build: remove -e . from requirements, use sys.path for src/

Add requirements.txt for HF Spaces; revert sys.path hack

Merge HF Spaces README frontmatter with project docs

initial commit

M0: fix qdrant-client API (search → query_points)

Initial commit: codesearch project with BM25 and dense retrieval