Repository navigation

feat(embed): static (model2vec) embeddings + extra-languages gating — v0.19.0 - #7

Merged

anvanster merged 21 commits into

mainfrom

feat/static-embeddings

Jul 1, 2026

Commits on Jun 26, 2026
- feat(embed): gated split-identifier words in embed text + distillation plan
  
  anvanster
  and
  claude
  committed
  de3e45d
- refactor(embed): Embedder trait — swappable backend behind VectorEngine
  
  anvanster
  and
  claude
  committed
  eb60736
- feat(embed): StaticEmbedding backend — model2vec-format lookup, no ONNX
  
  anvanster
  and
  claude
  committed
  a30df5e
- test(embed): throughput example — static vs ONNX
  anvanster
  committed
  1da5309
- docs(embed): record progress + 103x release throughput in the plan
  
  anvanster
  and
  claude
  committed
  f883283
- test(embed): retrieval quality micro-eval — static floor vs BGE
  
  anvanster
  and
  claude
  committed
  86fa334
- feat(embed): F16 + SIF-weighted static loading + Jina-Code distill script
  
  anvanster
  and
  claude
  committed
  68b1c8e
- docs(embed): record jina-code-static A/B — 70x faster, micro-eval saturated
  
  anvanster
  and
  claude
  committed
  28e7e88
- test(embed): real 965-way retrieval eval — static ~65% of BGE (pure semantic)
  
  anvanster
  and
  claude
  committed
  4083113
- test(embed): complete static lever sweep — gap is the contextualization ceiling
  
  anvanster
  and
  claude
  committed
  56ff954
- test(embed): hybrid (BM25+semantic) eval — static is ~90% of BGE end-to-end
  
  anvanster
  and
  claude
  committed
  a651f79
Commits on Jun 27, 2026
- test(embed): embed_throughput can load a real corpus (CODEGRAPH_THROUGHPUT_CORPUS)
  
  anvanster
  and
  claude
  committed
  55fd388
Commits on Jun 28, 2026
- feat(embed): EmbeddingBackend — select static via --embedding-model static
  
  anvanster
  and
  claude
  committed
  7e5d866
- docs(embed): expose static mode in VS Code + MCP + READMEs
  
  anvanster
  and
  claude
  committed
  2db0f08
- test(embed): server-side eval driver + full-body corpus extractor
  
  anvanster
  and
  claude
  committed
  c14174c
- docs(embed): server-side eval — static ~94% of BGE through real symbol_search
  
  anvanster
  and
  claude
  committed
  49289d0
Commits on Jun 29, 2026
- perf(parser): gate 6 zero-usage grammars behind `extra-languages` (-25 MB)
  
  anvanster
  and
  claude
  committed
  a7a6c33
- docs: note COBOL/Fortran/Perl/Dart/Zig/R are extra-languages-gated
  
  anvanster
  and
  claude
  committed
  1ce7c7e
- feat(embed): static-model telemetry + release-independent model fetch
  
  anvanster
  and
  claude
  committed
  865c994
Commits on Jun 30, 2026
- chore: bump version to 0.19.0 (static embeddings feature)
  
  anvanster
  and
  claude
  committed
  f0a6dbb
- docs(embed): note static model ships bundled in VS Code extension
  
  anvanster
  and
  claude
  committed
  1391ebc

Navigation Menu

Repository navigation

feat(embed): static (model2vec) embeddings + extra-languages gating — v0.19.0 - #7

feat(embed): static (model2vec) embeddings + extra-languages gating — v0.19.0#7
anvanster merged 21 commits into
mainfrom
feat/static-embeddings

Commits on Jun 26, 2026

feat(embed): gated split-identifier words in embed text + distillation plan

refactor(embed): Embedder trait — swappable backend behind VectorEngine

feat(embed): StaticEmbedding backend — model2vec-format lookup, no ONNX

test(embed): throughput example — static vs ONNX

docs(embed): record progress + 103x release throughput in the plan

test(embed): retrieval quality micro-eval — static floor vs BGE

feat(embed): F16 + SIF-weighted static loading + Jina-Code distill script

docs(embed): record jina-code-static A/B — 70x faster, micro-eval saturated

test(embed): real 965-way retrieval eval — static ~65% of BGE (pure semantic)

test(embed): complete static lever sweep — gap is the contextualization ceiling

test(embed): hybrid (BM25+semantic) eval — static is ~90% of BGE end-to-end

Commits on Jun 27, 2026

test(embed): embed_throughput can load a real corpus (CODEGRAPH_THROUGHPUT_CORPUS)

Commits on Jun 28, 2026

feat(embed): EmbeddingBackend — select static via --embedding-model static

docs(embed): expose static mode in VS Code + MCP + READMEs

test(embed): server-side eval driver + full-body corpus extractor

docs(embed): server-side eval — static ~94% of BGE through real symbol_search

Commits on Jun 29, 2026

perf(parser): gate 6 zero-usage grammars behind `extra-languages` (-25 MB)

docs: note COBOL/Fortran/Perl/Dart/Zig/R are extra-languages-gated

feat(embed): static-model telemetry + release-independent model fetch

Commits on Jun 30, 2026

chore: bump version to 0.19.0 (static embeddings feature)

docs(embed): note static model ships bundled in VS Code extension