grepmind-core 0.1.0-alpha
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +190 -0
- package/dist/config/types.d.ts +174 -0
- package/dist/config/types.d.ts.map +1 -0
- package/dist/config/types.js +137 -0
- package/dist/config/types.js.map +1 -0
- package/dist/git.d.ts +98 -0
- package/dist/git.d.ts.map +1 -0
- package/dist/git.js +298 -0
- package/dist/git.js.map +1 -0
- package/dist/git.test.d.ts +7 -0
- package/dist/git.test.d.ts.map +1 -0
- package/dist/git.test.js +242 -0
- package/dist/git.test.js.map +1 -0
- package/dist/index.d.ts +44 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +67 -0
- package/dist/index.js.map +1 -0
- package/dist/indexer/branch.d.ts +121 -0
- package/dist/indexer/branch.d.ts.map +1 -0
- package/dist/indexer/branch.js +451 -0
- package/dist/indexer/branch.js.map +1 -0
- package/dist/indexer/chunker.d.ts +9 -0
- package/dist/indexer/chunker.d.ts.map +1 -0
- package/dist/indexer/chunker.js +70 -0
- package/dist/indexer/chunker.js.map +1 -0
- package/dist/indexer/chunker.test.d.ts +2 -0
- package/dist/indexer/chunker.test.d.ts.map +1 -0
- package/dist/indexer/chunker.test.js +180 -0
- package/dist/indexer/chunker.test.js.map +1 -0
- package/dist/indexer/code/branch.d.ts +155 -0
- package/dist/indexer/code/branch.d.ts.map +1 -0
- package/dist/indexer/code/branch.js +550 -0
- package/dist/indexer/code/branch.js.map +1 -0
- package/dist/indexer/code/branch.test.d.ts +7 -0
- package/dist/indexer/code/branch.test.d.ts.map +1 -0
- package/dist/indexer/code/branch.test.js +241 -0
- package/dist/indexer/code/branch.test.js.map +1 -0
- package/dist/indexer/code/chunker.d.ts +61 -0
- package/dist/indexer/code/chunker.d.ts.map +1 -0
- package/dist/indexer/code/chunker.js +311 -0
- package/dist/indexer/code/chunker.js.map +1 -0
- package/dist/indexer/code/chunker.test.d.ts +2 -0
- package/dist/indexer/code/chunker.test.d.ts.map +1 -0
- package/dist/indexer/code/chunker.test.js +552 -0
- package/dist/indexer/code/chunker.test.js.map +1 -0
- package/dist/indexer/code/fts.test.d.ts +2 -0
- package/dist/indexer/code/fts.test.d.ts.map +1 -0
- package/dist/indexer/code/fts.test.js +14 -0
- package/dist/indexer/code/fts.test.js.map +1 -0
- package/dist/indexer/code/graph/embedded.d.ts +11 -0
- package/dist/indexer/code/graph/embedded.d.ts.map +1 -0
- package/dist/indexer/code/graph/embedded.js +152 -0
- package/dist/indexer/code/graph/embedded.js.map +1 -0
- package/dist/indexer/code/graph/embedded.test.d.ts +2 -0
- package/dist/indexer/code/graph/embedded.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/embedded.test.js +105 -0
- package/dist/indexer/code/graph/embedded.test.js.map +1 -0
- package/dist/indexer/code/graph/facts.d.ts +11 -0
- package/dist/indexer/code/graph/facts.d.ts.map +1 -0
- package/dist/indexer/code/graph/facts.js +456 -0
- package/dist/indexer/code/graph/facts.js.map +1 -0
- package/dist/indexer/code/graph/facts.test.d.ts +2 -0
- package/dist/indexer/code/graph/facts.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/facts.test.js +181 -0
- package/dist/indexer/code/graph/facts.test.js.map +1 -0
- package/dist/indexer/code/graph/id.d.ts +14 -0
- package/dist/indexer/code/graph/id.d.ts.map +1 -0
- package/dist/indexer/code/graph/id.js +40 -0
- package/dist/indexer/code/graph/id.js.map +1 -0
- package/dist/indexer/code/graph/id.test.d.ts +2 -0
- package/dist/indexer/code/graph/id.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/id.test.js +86 -0
- package/dist/indexer/code/graph/id.test.js.map +1 -0
- package/dist/indexer/code/graph/index.d.ts +133 -0
- package/dist/indexer/code/graph/index.d.ts.map +1 -0
- package/dist/indexer/code/graph/index.js +1876 -0
- package/dist/indexer/code/graph/index.js.map +1 -0
- package/dist/indexer/code/graph/index.test.d.ts +2 -0
- package/dist/indexer/code/graph/index.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/index.test.js +210 -0
- package/dist/indexer/code/graph/index.test.js.map +1 -0
- package/dist/indexer/code/graph/queries.d.ts +22 -0
- package/dist/indexer/code/graph/queries.d.ts.map +1 -0
- package/dist/indexer/code/graph/queries.js +79 -0
- package/dist/indexer/code/graph/queries.js.map +1 -0
- package/dist/indexer/code/graph/queries.test.d.ts +2 -0
- package/dist/indexer/code/graph/queries.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/queries.test.js +108 -0
- package/dist/indexer/code/graph/queries.test.js.map +1 -0
- package/dist/indexer/code/graph/resolver.d.ts +136 -0
- package/dist/indexer/code/graph/resolver.d.ts.map +1 -0
- package/dist/indexer/code/graph/resolver.js +839 -0
- package/dist/indexer/code/graph/resolver.js.map +1 -0
- package/dist/indexer/code/graph/resolver.test.d.ts +2 -0
- package/dist/indexer/code/graph/resolver.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/resolver.test.js +482 -0
- package/dist/indexer/code/graph/resolver.test.js.map +1 -0
- package/dist/indexer/code/graph/semantic.d.ts +33 -0
- package/dist/indexer/code/graph/semantic.d.ts.map +1 -0
- package/dist/indexer/code/graph/semantic.js +279 -0
- package/dist/indexer/code/graph/semantic.js.map +1 -0
- package/dist/indexer/code/graph/semantic.test.d.ts +2 -0
- package/dist/indexer/code/graph/semantic.test.d.ts.map +1 -0
- package/dist/indexer/code/graph/semantic.test.js +127 -0
- package/dist/indexer/code/graph/semantic.test.js.map +1 -0
- package/dist/indexer/code/index.d.ts +404 -0
- package/dist/indexer/code/index.d.ts.map +1 -0
- package/dist/indexer/code/index.js +2070 -0
- package/dist/indexer/code/index.js.map +1 -0
- package/dist/indexer/code/languages/bash.d.ts +14 -0
- package/dist/indexer/code/languages/bash.d.ts.map +1 -0
- package/dist/indexer/code/languages/bash.js +125 -0
- package/dist/indexer/code/languages/bash.js.map +1 -0
- package/dist/indexer/code/languages/css.d.ts +16 -0
- package/dist/indexer/code/languages/css.d.ts.map +1 -0
- package/dist/indexer/code/languages/css.js +204 -0
- package/dist/indexer/code/languages/css.js.map +1 -0
- package/dist/indexer/code/languages/generic.d.ts +61 -0
- package/dist/indexer/code/languages/generic.d.ts.map +1 -0
- package/dist/indexer/code/languages/generic.js +150 -0
- package/dist/indexer/code/languages/generic.js.map +1 -0
- package/dist/indexer/code/languages/graphql.d.ts +13 -0
- package/dist/indexer/code/languages/graphql.d.ts.map +1 -0
- package/dist/indexer/code/languages/graphql.js +180 -0
- package/dist/indexer/code/languages/graphql.js.map +1 -0
- package/dist/indexer/code/languages/html.d.ts +16 -0
- package/dist/indexer/code/languages/html.d.ts.map +1 -0
- package/dist/indexer/code/languages/html.js +138 -0
- package/dist/indexer/code/languages/html.js.map +1 -0
- package/dist/indexer/code/languages/index.d.ts +9 -0
- package/dist/indexer/code/languages/index.d.ts.map +1 -0
- package/dist/indexer/code/languages/index.js +12 -0
- package/dist/indexer/code/languages/index.js.map +1 -0
- package/dist/indexer/code/languages/json.d.ts +12 -0
- package/dist/indexer/code/languages/json.d.ts.map +1 -0
- package/dist/indexer/code/languages/json.js +66 -0
- package/dist/indexer/code/languages/json.js.map +1 -0
- package/dist/indexer/code/languages/registry.d.ts +78 -0
- package/dist/indexer/code/languages/registry.d.ts.map +1 -0
- package/dist/indexer/code/languages/registry.js +72 -0
- package/dist/indexer/code/languages/registry.js.map +1 -0
- package/dist/indexer/code/languages/typescript.d.ts +39 -0
- package/dist/indexer/code/languages/typescript.d.ts.map +1 -0
- package/dist/indexer/code/languages/typescript.js +300 -0
- package/dist/indexer/code/languages/typescript.js.map +1 -0
- package/dist/indexer/code/languages/yaml.d.ts +13 -0
- package/dist/indexer/code/languages/yaml.d.ts.map +1 -0
- package/dist/indexer/code/languages/yaml.js +90 -0
- package/dist/indexer/code/languages/yaml.js.map +1 -0
- package/dist/indexer/code/parser.d.ts +26 -0
- package/dist/indexer/code/parser.d.ts.map +1 -0
- package/dist/indexer/code/parser.js +332 -0
- package/dist/indexer/code/parser.js.map +1 -0
- package/dist/indexer/code/retry.d.ts +58 -0
- package/dist/indexer/code/retry.d.ts.map +1 -0
- package/dist/indexer/code/retry.js +192 -0
- package/dist/indexer/code/retry.js.map +1 -0
- package/dist/indexer/code/tree/builder.d.ts +30 -0
- package/dist/indexer/code/tree/builder.d.ts.map +1 -0
- package/dist/indexer/code/tree/builder.js +132 -0
- package/dist/indexer/code/tree/builder.js.map +1 -0
- package/dist/indexer/code/tree/builder.test.d.ts +2 -0
- package/dist/indexer/code/tree/builder.test.d.ts.map +1 -0
- package/dist/indexer/code/tree/builder.test.js +31 -0
- package/dist/indexer/code/tree/builder.test.js.map +1 -0
- package/dist/indexer/code/tree/cache.d.ts +22 -0
- package/dist/indexer/code/tree/cache.d.ts.map +1 -0
- package/dist/indexer/code/tree/cache.js +85 -0
- package/dist/indexer/code/tree/cache.js.map +1 -0
- package/dist/indexer/code/tree/context.d.ts +32 -0
- package/dist/indexer/code/tree/context.d.ts.map +1 -0
- package/dist/indexer/code/tree/context.js +78 -0
- package/dist/indexer/code/tree/context.js.map +1 -0
- package/dist/indexer/code/tree/embedding.d.ts +9 -0
- package/dist/indexer/code/tree/embedding.d.ts.map +1 -0
- package/dist/indexer/code/tree/embedding.js +53 -0
- package/dist/indexer/code/tree/embedding.js.map +1 -0
- package/dist/indexer/code/tree/embedding.test.d.ts +2 -0
- package/dist/indexer/code/tree/embedding.test.d.ts.map +1 -0
- package/dist/indexer/code/tree/embedding.test.js +57 -0
- package/dist/indexer/code/tree/embedding.test.js.map +1 -0
- package/dist/indexer/code/tree/id.d.ts +3 -0
- package/dist/indexer/code/tree/id.d.ts.map +1 -0
- package/dist/indexer/code/tree/id.js +8 -0
- package/dist/indexer/code/tree/id.js.map +1 -0
- package/dist/indexer/code/tree/index.d.ts +113 -0
- package/dist/indexer/code/tree/index.d.ts.map +1 -0
- package/dist/indexer/code/tree/index.js +1146 -0
- package/dist/indexer/code/tree/index.js.map +1 -0
- package/dist/indexer/code/tree/rename.d.ts +13 -0
- package/dist/indexer/code/tree/rename.d.ts.map +1 -0
- package/dist/indexer/code/tree/rename.js +46 -0
- package/dist/indexer/code/tree/rename.js.map +1 -0
- package/dist/indexer/code/tree/repomap.d.ts +29 -0
- package/dist/indexer/code/tree/repomap.d.ts.map +1 -0
- package/dist/indexer/code/tree/repomap.js +95 -0
- package/dist/indexer/code/tree/repomap.js.map +1 -0
- package/dist/indexer/code/tree/repomap.test.d.ts +2 -0
- package/dist/indexer/code/tree/repomap.test.d.ts.map +1 -0
- package/dist/indexer/code/tree/repomap.test.js +93 -0
- package/dist/indexer/code/tree/repomap.test.js.map +1 -0
- package/dist/indexer/code/tree/stats.d.ts +26 -0
- package/dist/indexer/code/tree/stats.d.ts.map +1 -0
- package/dist/indexer/code/tree/stats.js +49 -0
- package/dist/indexer/code/tree/stats.js.map +1 -0
- package/dist/indexer/code/tree/types.d.ts +186 -0
- package/dist/indexer/code/tree/types.d.ts.map +1 -0
- package/dist/indexer/code/tree/types.js +10 -0
- package/dist/indexer/code/tree/types.js.map +1 -0
- package/dist/indexer/code/wal.d.ts +144 -0
- package/dist/indexer/code/wal.d.ts.map +1 -0
- package/dist/indexer/code/wal.js +283 -0
- package/dist/indexer/code/wal.js.map +1 -0
- package/dist/indexer/embeddings.d.ts +113 -0
- package/dist/indexer/embeddings.d.ts.map +1 -0
- package/dist/indexer/embeddings.js +477 -0
- package/dist/indexer/embeddings.js.map +1 -0
- package/dist/indexer/git-sync.d.ts +117 -0
- package/dist/indexer/git-sync.d.ts.map +1 -0
- package/dist/indexer/git-sync.js +398 -0
- package/dist/indexer/git-sync.js.map +1 -0
- package/dist/indexer/index.d.ts +175 -0
- package/dist/indexer/index.d.ts.map +1 -0
- package/dist/indexer/index.js +1096 -0
- package/dist/indexer/index.js.map +1 -0
- package/dist/indexer/mocks/mock-reranker.d.ts +12 -0
- package/dist/indexer/mocks/mock-reranker.d.ts.map +1 -0
- package/dist/indexer/mocks/mock-reranker.js +26 -0
- package/dist/indexer/mocks/mock-reranker.js.map +1 -0
- package/dist/indexer/parser.d.ts +8 -0
- package/dist/indexer/parser.d.ts.map +1 -0
- package/dist/indexer/parser.js +44 -0
- package/dist/indexer/parser.js.map +1 -0
- package/dist/indexer/parser.test.d.ts +2 -0
- package/dist/indexer/parser.test.d.ts.map +1 -0
- package/dist/indexer/parser.test.js +197 -0
- package/dist/indexer/parser.test.js.map +1 -0
- package/dist/indexer/reranking.d.ts +71 -0
- package/dist/indexer/reranking.d.ts.map +1 -0
- package/dist/indexer/reranking.integration.test.d.ts +2 -0
- package/dist/indexer/reranking.integration.test.d.ts.map +1 -0
- package/dist/indexer/reranking.integration.test.js +104 -0
- package/dist/indexer/reranking.integration.test.js.map +1 -0
- package/dist/indexer/reranking.js +256 -0
- package/dist/indexer/reranking.js.map +1 -0
- package/dist/indexer/reranking.test.d.ts +2 -0
- package/dist/indexer/reranking.test.d.ts.map +1 -0
- package/dist/indexer/reranking.test.js +130 -0
- package/dist/indexer/reranking.test.js.map +1 -0
- package/dist/indexer/wal/file-storage.d.ts +60 -0
- package/dist/indexer/wal/file-storage.d.ts.map +1 -0
- package/dist/indexer/wal/file-storage.js +277 -0
- package/dist/indexer/wal/file-storage.js.map +1 -0
- package/dist/indexer/wal/file-storage.test.d.ts +8 -0
- package/dist/indexer/wal/file-storage.test.d.ts.map +1 -0
- package/dist/indexer/wal/file-storage.test.js +444 -0
- package/dist/indexer/wal/file-storage.test.js.map +1 -0
- package/dist/indexer/wal/index.d.ts +41 -0
- package/dist/indexer/wal/index.d.ts.map +1 -0
- package/dist/indexer/wal/index.js +61 -0
- package/dist/indexer/wal/index.js.map +1 -0
- package/dist/indexer/wal/integration.test.d.ts +11 -0
- package/dist/indexer/wal/integration.test.d.ts.map +1 -0
- package/dist/indexer/wal/integration.test.js +378 -0
- package/dist/indexer/wal/integration.test.js.map +1 -0
- package/dist/indexer/wal/lancedb-storage.d.ts +72 -0
- package/dist/indexer/wal/lancedb-storage.d.ts.map +1 -0
- package/dist/indexer/wal/lancedb-storage.js +462 -0
- package/dist/indexer/wal/lancedb-storage.js.map +1 -0
- package/dist/indexer/wal/lancedb-storage.test.d.ts +8 -0
- package/dist/indexer/wal/lancedb-storage.test.d.ts.map +1 -0
- package/dist/indexer/wal/lancedb-storage.test.js +415 -0
- package/dist/indexer/wal/lancedb-storage.test.js.map +1 -0
- package/dist/indexer/wal/sync-wal.d.ts +144 -0
- package/dist/indexer/wal/sync-wal.d.ts.map +1 -0
- package/dist/indexer/wal/sync-wal.js +863 -0
- package/dist/indexer/wal/sync-wal.js.map +1 -0
- package/dist/indexer/wal/sync-wal.test.d.ts +8 -0
- package/dist/indexer/wal/sync-wal.test.d.ts.map +1 -0
- package/dist/indexer/wal/sync-wal.test.js +752 -0
- package/dist/indexer/wal/sync-wal.test.js.map +1 -0
- package/dist/indexer/wal/types.d.ts +167 -0
- package/dist/indexer/wal/types.d.ts.map +1 -0
- package/dist/indexer/wal/types.js +12 -0
- package/dist/indexer/wal/types.js.map +1 -0
- package/dist/indexer/watcher.d.ts +36 -0
- package/dist/indexer/watcher.d.ts.map +1 -0
- package/dist/indexer/watcher.js +110 -0
- package/dist/indexer/watcher.js.map +1 -0
- package/dist/search/explore.d.ts +62 -0
- package/dist/search/explore.d.ts.map +1 -0
- package/dist/search/explore.js +111 -0
- package/dist/search/explore.js.map +1 -0
- package/dist/search/fts.d.ts +23 -0
- package/dist/search/fts.d.ts.map +1 -0
- package/dist/search/fts.js +64 -0
- package/dist/search/fts.js.map +1 -0
- package/dist/search/fts.test.d.ts +2 -0
- package/dist/search/fts.test.d.ts.map +1 -0
- package/dist/search/fts.test.js +27 -0
- package/dist/search/fts.test.js.map +1 -0
- package/dist/search/grep.d.ts +75 -0
- package/dist/search/grep.d.ts.map +1 -0
- package/dist/search/grep.js +96 -0
- package/dist/search/grep.js.map +1 -0
- package/dist/search/grep.test.d.ts +2 -0
- package/dist/search/grep.test.d.ts.map +1 -0
- package/dist/search/grep.test.js +178 -0
- package/dist/search/grep.test.js.map +1 -0
- package/dist/search/hybrid-grep.d.ts +43 -0
- package/dist/search/hybrid-grep.d.ts.map +1 -0
- package/dist/search/hybrid-grep.js +130 -0
- package/dist/search/hybrid-grep.js.map +1 -0
- package/dist/search/hybrid-grep.test.d.ts +2 -0
- package/dist/search/hybrid-grep.test.d.ts.map +1 -0
- package/dist/search/hybrid-grep.test.js +133 -0
- package/dist/search/hybrid-grep.test.js.map +1 -0
- package/dist/search/rg-executor.d.ts +63 -0
- package/dist/search/rg-executor.d.ts.map +1 -0
- package/dist/search/rg-executor.js +146 -0
- package/dist/search/rg-executor.js.map +1 -0
- package/dist/search/rg-executor.test.d.ts +2 -0
- package/dist/search/rg-executor.test.d.ts.map +1 -0
- package/dist/search/rg-executor.test.js +104 -0
- package/dist/search/rg-executor.test.js.map +1 -0
- package/dist/search/rg-parser/extractor.d.ts +14 -0
- package/dist/search/rg-parser/extractor.d.ts.map +1 -0
- package/dist/search/rg-parser/extractor.js +82 -0
- package/dist/search/rg-parser/extractor.js.map +1 -0
- package/dist/search/rg-parser/extractor.test.d.ts +2 -0
- package/dist/search/rg-parser/extractor.test.d.ts.map +1 -0
- package/dist/search/rg-parser/extractor.test.js +35 -0
- package/dist/search/rg-parser/extractor.test.js.map +1 -0
- package/dist/search/rg-parser/fts-builder.d.ts +7 -0
- package/dist/search/rg-parser/fts-builder.d.ts.map +1 -0
- package/dist/search/rg-parser/fts-builder.js +18 -0
- package/dist/search/rg-parser/fts-builder.js.map +1 -0
- package/dist/search/rg-parser/fts-builder.test.d.ts +2 -0
- package/dist/search/rg-parser/fts-builder.test.d.ts.map +1 -0
- package/dist/search/rg-parser/fts-builder.test.js +26 -0
- package/dist/search/rg-parser/fts-builder.test.js.map +1 -0
- package/dist/search/rg-parser/index.d.ts +36 -0
- package/dist/search/rg-parser/index.d.ts.map +1 -0
- package/dist/search/rg-parser/index.js +83 -0
- package/dist/search/rg-parser/index.js.map +1 -0
- package/dist/search/rg-parser/index.test.d.ts +2 -0
- package/dist/search/rg-parser/index.test.d.ts.map +1 -0
- package/dist/search/rg-parser/index.test.js +34 -0
- package/dist/search/rg-parser/index.test.js.map +1 -0
- package/dist/search/rg-parser/strategy.d.ts +14 -0
- package/dist/search/rg-parser/strategy.d.ts.map +1 -0
- package/dist/search/rg-parser/strategy.js +31 -0
- package/dist/search/rg-parser/strategy.js.map +1 -0
- package/dist/search/rg-parser/strategy.test.d.ts +2 -0
- package/dist/search/rg-parser/strategy.test.d.ts.map +1 -0
- package/dist/search/rg-parser/strategy.test.js +29 -0
- package/dist/search/rg-parser/strategy.test.js.map +1 -0
- package/dist/types.d.ts +345 -0
- package/dist/types.d.ts.map +1 -0
- package/dist/types.js +7 -0
- package/dist/types.js.map +1 -0
- package/dist/utils/vault.d.ts +84 -0
- package/dist/utils/vault.d.ts.map +1 -0
- package/dist/utils/vault.js +138 -0
- package/dist/utils/vault.js.map +1 -0
- package/dist/utils/vault.test.d.ts +2 -0
- package/dist/utils/vault.test.d.ts.map +1 -0
- package/dist/utils/vault.test.js +153 -0
- package/dist/utils/vault.test.js.map +1 -0
- package/package.json +69 -0
|
@@ -0,0 +1,256 @@
|
|
|
1
|
+
import { pipeline } from '@xenova/transformers';
|
|
2
|
+
import { createError } from '../types.js';
|
|
3
|
+
/**
|
|
4
|
+
* Normalize scores to [0, 1] range.
|
|
5
|
+
*/
|
|
6
|
+
export function normalizeScores(scores) {
|
|
7
|
+
if (scores.length === 0)
|
|
8
|
+
return [];
|
|
9
|
+
const min = Math.min(...scores);
|
|
10
|
+
const max = Math.max(...scores);
|
|
11
|
+
if (max === min)
|
|
12
|
+
return scores.map(() => 1);
|
|
13
|
+
return scores.map((score) => (score - min) / (max - min));
|
|
14
|
+
}
|
|
15
|
+
/**
|
|
16
|
+
* Truncate document by character length, preferring word boundary.
|
|
17
|
+
*/
|
|
18
|
+
export function truncateDocument(doc, maxLength) {
|
|
19
|
+
if (doc.length <= maxLength)
|
|
20
|
+
return doc;
|
|
21
|
+
const truncated = doc.slice(0, maxLength);
|
|
22
|
+
const lastSpace = truncated.lastIndexOf(' ');
|
|
23
|
+
return lastSpace > maxLength * 0.8 ? truncated.slice(0, lastSpace) : truncated;
|
|
24
|
+
}
|
|
25
|
+
/**
|
|
26
|
+
* Split an array into fixed-size batches.
|
|
27
|
+
*/
|
|
28
|
+
export function batchArray(array, batchSize) {
|
|
29
|
+
const batches = [];
|
|
30
|
+
for (let i = 0; i < array.length; i += batchSize) {
|
|
31
|
+
batches.push(array.slice(i, i + batchSize));
|
|
32
|
+
}
|
|
33
|
+
return batches;
|
|
34
|
+
}
|
|
35
|
+
function sortAndSlice(results, topK) {
|
|
36
|
+
const sorted = [...results].sort((a, b) => b.score - a.score);
|
|
37
|
+
return sorted.slice(0, topK ?? sorted.length);
|
|
38
|
+
}
|
|
39
|
+
function extractClassificationScore(result) {
|
|
40
|
+
if (Array.isArray(result)) {
|
|
41
|
+
const first = result[0];
|
|
42
|
+
return typeof first?.score === 'number' ? first.score : 0;
|
|
43
|
+
}
|
|
44
|
+
if (typeof result === 'object' && result !== null) {
|
|
45
|
+
const maybeScore = result.score;
|
|
46
|
+
return typeof maybeScore === 'number' ? maybeScore : 0;
|
|
47
|
+
}
|
|
48
|
+
return 0;
|
|
49
|
+
}
|
|
50
|
+
export class WasmRerankerProvider {
|
|
51
|
+
classifier = null;
|
|
52
|
+
initPromise = null;
|
|
53
|
+
modelName;
|
|
54
|
+
maxDocumentLength;
|
|
55
|
+
constructor(modelName, maxDocumentLength) {
|
|
56
|
+
this.modelName = modelName;
|
|
57
|
+
this.maxDocumentLength = maxDocumentLength;
|
|
58
|
+
}
|
|
59
|
+
isInitialized() {
|
|
60
|
+
return this.classifier !== null;
|
|
61
|
+
}
|
|
62
|
+
getModelName() {
|
|
63
|
+
return this.modelName;
|
|
64
|
+
}
|
|
65
|
+
async getClassifier() {
|
|
66
|
+
if (this.classifier)
|
|
67
|
+
return this.classifier;
|
|
68
|
+
if (!this.initPromise) {
|
|
69
|
+
this.initPromise = pipeline('text-classification', this.modelName, {
|
|
70
|
+
quantized: true,
|
|
71
|
+
});
|
|
72
|
+
}
|
|
73
|
+
this.classifier = await this.initPromise;
|
|
74
|
+
return this.classifier;
|
|
75
|
+
}
|
|
76
|
+
async initialize() {
|
|
77
|
+
await this.getClassifier();
|
|
78
|
+
}
|
|
79
|
+
async rerank(query, documents, options) {
|
|
80
|
+
if (documents.length === 0)
|
|
81
|
+
return [];
|
|
82
|
+
try {
|
|
83
|
+
const classifier = await this.getClassifier();
|
|
84
|
+
const rawScores = [];
|
|
85
|
+
for (const doc of documents) {
|
|
86
|
+
const truncatedDoc = truncateDocument(doc, this.maxDocumentLength);
|
|
87
|
+
const result = await classifier(`${query} [SEP] ${truncatedDoc}`);
|
|
88
|
+
rawScores.push(extractClassificationScore(result));
|
|
89
|
+
}
|
|
90
|
+
const normalized = normalizeScores(rawScores);
|
|
91
|
+
const ranked = normalized.map((score, index) => ({
|
|
92
|
+
index,
|
|
93
|
+
score,
|
|
94
|
+
...(options?.returnDocuments ? { document: documents[index] } : {}),
|
|
95
|
+
}));
|
|
96
|
+
return sortAndSlice(ranked, options?.topK);
|
|
97
|
+
}
|
|
98
|
+
catch (error) {
|
|
99
|
+
return createError('RERANK_FAILED', `WASM reranking failed: ${error.message}`, { error: error.message });
|
|
100
|
+
}
|
|
101
|
+
}
|
|
102
|
+
}
|
|
103
|
+
export class VoyageRerankerProvider {
|
|
104
|
+
initialized = false;
|
|
105
|
+
modelName;
|
|
106
|
+
apiKey;
|
|
107
|
+
batchSize;
|
|
108
|
+
maxDocumentLength;
|
|
109
|
+
constructor(modelName, apiKey, batchSize, maxDocumentLength) {
|
|
110
|
+
this.modelName = modelName;
|
|
111
|
+
this.apiKey = apiKey;
|
|
112
|
+
this.batchSize = batchSize;
|
|
113
|
+
this.maxDocumentLength = maxDocumentLength;
|
|
114
|
+
}
|
|
115
|
+
async initialize() {
|
|
116
|
+
this.initialized = true;
|
|
117
|
+
}
|
|
118
|
+
isInitialized() {
|
|
119
|
+
return this.initialized;
|
|
120
|
+
}
|
|
121
|
+
getModelName() {
|
|
122
|
+
return `voyage:${this.modelName}`;
|
|
123
|
+
}
|
|
124
|
+
async rerank(query, documents, options) {
|
|
125
|
+
if (documents.length === 0)
|
|
126
|
+
return [];
|
|
127
|
+
try {
|
|
128
|
+
const truncatedDocuments = documents.map((doc) => truncateDocument(doc, this.maxDocumentLength));
|
|
129
|
+
const results = [];
|
|
130
|
+
for (let offset = 0; offset < truncatedDocuments.length; offset += this.batchSize) {
|
|
131
|
+
const batch = truncatedDocuments.slice(offset, offset + this.batchSize);
|
|
132
|
+
const response = await fetch('https://api.voyageai.com/v1/rerank', {
|
|
133
|
+
method: 'POST',
|
|
134
|
+
headers: {
|
|
135
|
+
Authorization: `Bearer ${this.apiKey}`,
|
|
136
|
+
'Content-Type': 'application/json',
|
|
137
|
+
},
|
|
138
|
+
body: JSON.stringify({
|
|
139
|
+
query,
|
|
140
|
+
documents: batch,
|
|
141
|
+
model: this.modelName,
|
|
142
|
+
top_k: batch.length,
|
|
143
|
+
}),
|
|
144
|
+
});
|
|
145
|
+
if (!response.ok) {
|
|
146
|
+
const body = await response.text();
|
|
147
|
+
return createError('RERANK_FAILED', `Voyage rerank error ${response.status}: ${body}`);
|
|
148
|
+
}
|
|
149
|
+
const data = (await response.json());
|
|
150
|
+
for (const item of data.data ?? []) {
|
|
151
|
+
const globalIndex = offset + item.index;
|
|
152
|
+
results.push({
|
|
153
|
+
index: globalIndex,
|
|
154
|
+
score: item.relevance_score,
|
|
155
|
+
...(options?.returnDocuments ? { document: documents[globalIndex] } : {}),
|
|
156
|
+
});
|
|
157
|
+
}
|
|
158
|
+
}
|
|
159
|
+
return sortAndSlice(results, options?.topK);
|
|
160
|
+
}
|
|
161
|
+
catch (error) {
|
|
162
|
+
return createError('RERANK_FAILED', `Voyage reranking failed: ${error.message}`, { error: error.message });
|
|
163
|
+
}
|
|
164
|
+
}
|
|
165
|
+
}
|
|
166
|
+
export class LocalAIRerankerProvider {
|
|
167
|
+
initialized = false;
|
|
168
|
+
baseUrl;
|
|
169
|
+
modelName;
|
|
170
|
+
apiKey;
|
|
171
|
+
batchSize;
|
|
172
|
+
maxDocumentLength;
|
|
173
|
+
constructor(baseUrl, modelName, batchSize, maxDocumentLength, apiKey) {
|
|
174
|
+
this.baseUrl = baseUrl.replace(/\/$/, '');
|
|
175
|
+
this.modelName = modelName;
|
|
176
|
+
this.batchSize = batchSize;
|
|
177
|
+
this.maxDocumentLength = maxDocumentLength;
|
|
178
|
+
this.apiKey = apiKey;
|
|
179
|
+
}
|
|
180
|
+
async initialize() {
|
|
181
|
+
this.initialized = true;
|
|
182
|
+
}
|
|
183
|
+
isInitialized() {
|
|
184
|
+
return this.initialized;
|
|
185
|
+
}
|
|
186
|
+
getModelName() {
|
|
187
|
+
return `localai:${this.modelName}`;
|
|
188
|
+
}
|
|
189
|
+
async rerank(query, documents, options) {
|
|
190
|
+
if (documents.length === 0)
|
|
191
|
+
return [];
|
|
192
|
+
try {
|
|
193
|
+
const truncatedDocuments = documents.map((doc) => truncateDocument(doc, this.maxDocumentLength));
|
|
194
|
+
const rawScores = new Array(documents.length).fill(0);
|
|
195
|
+
for (let offset = 0; offset < truncatedDocuments.length; offset += this.batchSize) {
|
|
196
|
+
const batch = truncatedDocuments.slice(offset, offset + this.batchSize);
|
|
197
|
+
const headers = {
|
|
198
|
+
'Content-Type': 'application/json',
|
|
199
|
+
};
|
|
200
|
+
if (this.apiKey) {
|
|
201
|
+
headers.Authorization = `Bearer ${this.apiKey}`;
|
|
202
|
+
}
|
|
203
|
+
const response = await fetch(`${this.baseUrl}/v1/rerank`, {
|
|
204
|
+
method: 'POST',
|
|
205
|
+
headers,
|
|
206
|
+
body: JSON.stringify({
|
|
207
|
+
model: this.modelName,
|
|
208
|
+
query,
|
|
209
|
+
documents: batch,
|
|
210
|
+
top_n: batch.length,
|
|
211
|
+
}),
|
|
212
|
+
});
|
|
213
|
+
if (!response.ok) {
|
|
214
|
+
const body = await response.text();
|
|
215
|
+
return createError('RERANK_FAILED', `LocalAI rerank error ${response.status}: ${body}`);
|
|
216
|
+
}
|
|
217
|
+
const data = (await response.json());
|
|
218
|
+
const items = data.data ?? data.results ?? [];
|
|
219
|
+
for (const item of items) {
|
|
220
|
+
const globalIndex = offset + item.index;
|
|
221
|
+
const score = item.relevance_score ?? item.score ?? 0;
|
|
222
|
+
rawScores[globalIndex] = score;
|
|
223
|
+
}
|
|
224
|
+
}
|
|
225
|
+
const normalizedScores = normalizeScores(rawScores);
|
|
226
|
+
const results = normalizedScores.map((score, index) => ({
|
|
227
|
+
index,
|
|
228
|
+
score,
|
|
229
|
+
...(options?.returnDocuments ? { document: documents[index] } : {}),
|
|
230
|
+
}));
|
|
231
|
+
return sortAndSlice(results, options?.topK);
|
|
232
|
+
}
|
|
233
|
+
catch (error) {
|
|
234
|
+
return createError('RERANK_FAILED', `LocalAI reranking failed: ${error.message}`, { error: error.message });
|
|
235
|
+
}
|
|
236
|
+
}
|
|
237
|
+
}
|
|
238
|
+
export function createRerankerProvider(config, secrets) {
|
|
239
|
+
if (!config.enabled || config.provider === 'none') {
|
|
240
|
+
return null;
|
|
241
|
+
}
|
|
242
|
+
switch (config.provider) {
|
|
243
|
+
case 'wasm':
|
|
244
|
+
return new WasmRerankerProvider(config.wasmModel, config.maxDocumentLength);
|
|
245
|
+
case 'voyage':
|
|
246
|
+
if (!secrets?.voyageApiKey) {
|
|
247
|
+
throw new Error('Voyage API key required for reranking');
|
|
248
|
+
}
|
|
249
|
+
return new VoyageRerankerProvider(config.voyageModel, secrets.voyageApiKey, config.batchSize, config.maxDocumentLength);
|
|
250
|
+
case 'localai':
|
|
251
|
+
return new LocalAIRerankerProvider(config.localaiBaseUrl, config.localaiRerankModel, config.batchSize, config.maxDocumentLength, secrets?.localaiApiKey);
|
|
252
|
+
default:
|
|
253
|
+
return null;
|
|
254
|
+
}
|
|
255
|
+
}
|
|
256
|
+
//# sourceMappingURL=reranking.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"reranking.js","sourceRoot":"","sources":["../../src/indexer/reranking.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAmC,MAAM,sBAAsB,CAAC;AAEjF,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAqC1C;;GAEG;AACH,MAAM,UAAU,eAAe,CAAC,MAAgB;IAC9C,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACnC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC;IAChC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC;IAChC,IAAI,GAAG,KAAK,GAAG;QAAE,OAAO,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;IAC5C,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC;AAC5D,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,gBAAgB,CAAC,GAAW,EAAE,SAAiB;IAC7D,IAAI,GAAG,CAAC,MAAM,IAAI,SAAS;QAAE,OAAO,GAAG,CAAC;IACxC,MAAM,SAAS,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;IAC1C,MAAM,SAAS,GAAG,SAAS,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC7C,OAAO,SAAS,GAAG,SAAS,GAAG,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AACjF,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAI,KAAU,EAAE,SAAiB;IACzD,MAAM,OAAO,GAAU,EAAE,CAAC;IAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QACjD,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,YAAY,CAAC,OAAuB,EAAE,IAAa;IAC1D,MAAM,MAAM,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAC9D,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,IAAI,MAAM,CAAC,MAAM,CAAC,CAAC;AAChD,CAAC;AAED,SAAS,0BAA0B,CAAC,MAAe;IACjD,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QAC1B,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAoC,CAAC;QAC3D,OAAO,OAAO,KAAK,EAAE,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5D,CAAC;IAED,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;QAClD,MAAM,UAAU,GAAI,MAA8B,CAAC,KAAK,CAAC;QACzD,OAAO,OAAO,UAAU,KAAK,QAAQ,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;IACzD,CAAC;IAED,OAAO,CAAC,CAAC;AACX,CAAC;AAED,MAAM,OAAO,oBAAoB;IACvB,UAAU,GAAsC,IAAI,CAAC;IACrD,WAAW,GAA+C,IAAI,CAAC;IACtD,SAAS,CAAS;IAClB,iBAAiB,CAAS;IAE3C,YAAY,SAAiB,EAAE,iBAAyB;QACtD,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,IAAI,CAAC,iBAAiB,GAAG,iBAAiB,CAAC;IAC7C,CAAC;IAED,aAAa;QACX,OAAO,IAAI,CAAC,UAAU,KAAK,IAAI,CAAC;IAClC,CAAC;IAED,YAAY;QACV,OAAO,IAAI,CAAC,SAAS,CAAC;IACxB,CAAC;IAEO,KAAK,CAAC,aAAa;QACzB,IAAI,IAAI,CAAC,UAAU;YAAE,OAAO,IAAI,CAAC,UAAU,CAAC;QAE5C,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;YACtB,IAAI,CAAC,WAAW,GAAG,QAAQ,CAAC,qBAAqB,EAAE,IAAI,CAAC,SAAS,EAAE;gBACjE,SAAS,EAAE,IAAI;aAChB,CAAC,CAAC;QACL,CAAC;QAED,IAAI,CAAC,UAAU,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC;QACzC,OAAO,IAAI,CAAC,UAAU,CAAC;IACzB,CAAC;IAED,KAAK,CAAC,UAAU;QACd,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;IAC7B,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,SAAmB,EAAE,OAAuB;QACtE,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEtC,IAAI,CAAC;YACH,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;YAC9C,MAAM,SAAS,GAAa,EAAE,CAAC;YAE/B,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC5B,MAAM,YAAY,GAAG,gBAAgB,CAAC,GAAG,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC;gBACnE,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,GAAG,KAAK,UAAU,YAAY,EAAE,CAAC,CAAC;gBAClE,SAAS,CAAC,IAAI,CAAC,0BAA0B,CAAC,MAAM,CAAC,CAAC,CAAC;YACrD,CAAC;YAED,MAAM,UAAU,GAAG,eAAe,CAAC,SAAS,CAAC,CAAC;YAC9C,MAAM,MAAM,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;gBAC/C,KAAK;gBACL,KAAK;gBACL,GAAG,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACpE,CAAC,CAAC,CAAC;YAEJ,OAAO,YAAY,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;QAC7C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,WAAW,CAChB,eAAe,EACf,0BAA2B,KAAe,CAAC,OAAO,EAAE,EACpD,EAAE,KAAK,EAAG,KAAe,CAAC,OAAO,EAAE,CACpC,CAAC;QACJ,CAAC;IACH,CAAC;CACF;AAED,MAAM,OAAO,sBAAsB;IACzB,WAAW,GAAG,KAAK,CAAC;IACX,SAAS,CAAS;IAClB,MAAM,CAAS;IACf,SAAS,CAAS;IAClB,iBAAiB,CAAS;IAE3C,YAAY,SAAiB,EAAE,MAAc,EAAE,SAAiB,EAAE,iBAAyB;QACzF,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,IAAI,CAAC,iBAAiB,GAAG,iBAAiB,CAAC;IAC7C,CAAC;IAED,KAAK,CAAC,UAAU;QACd,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC;IAC1B,CAAC;IAED,aAAa;QACX,OAAO,IAAI,CAAC,WAAW,CAAC;IAC1B,CAAC;IAED,YAAY;QACV,OAAO,UAAU,IAAI,CAAC,SAAS,EAAE,CAAC;IACpC,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,SAAmB,EAAE,OAAuB;QACtE,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEtC,IAAI,CAAC;YACH,MAAM,kBAAkB,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,gBAAgB,CAAC,GAAG,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC,CAAC;YACjG,MAAM,OAAO,GAAmB,EAAE,CAAC;YAEnC,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,kBAAkB,CAAC,MAAM,EAAE,MAAM,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;gBAClF,MAAM,KAAK,GAAG,kBAAkB,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC;gBACxE,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,oCAAoC,EAAE;oBACjE,MAAM,EAAE,MAAM;oBACd,OAAO,EAAE;wBACP,aAAa,EAAE,UAAU,IAAI,CAAC,MAAM,EAAE;wBACtC,cAAc,EAAE,kBAAkB;qBACnC;oBACD,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;wBACnB,KAAK;wBACL,SAAS,EAAE,KAAK;wBAChB,KAAK,EAAE,IAAI,CAAC,SAAS;wBACrB,KAAK,EAAE,KAAK,CAAC,MAAM;qBACpB,CAAC;iBACH,CAAC,CAAC;gBAEH,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;oBACjB,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;oBACnC,OAAO,WAAW,CAChB,eAAe,EACf,uBAAuB,QAAQ,CAAC,MAAM,KAAK,IAAI,EAAE,CAClD,CAAC;gBACJ,CAAC;gBAED,MAAM,IAAI,GAAG,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAkC,CAAC;gBACtE,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,IAAI,IAAI,EAAE,EAAE,CAAC;oBACnC,MAAM,WAAW,GAAG,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC;oBACxC,OAAO,CAAC,IAAI,CAAC;wBACX,KAAK,EAAE,WAAW;wBAClB,KAAK,EAAE,IAAI,CAAC,eAAe;wBAC3B,GAAG,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,SAAS,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;qBAC1E,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,OAAO,YAAY,CAAC,OAAO,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;QAC9C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,WAAW,CAChB,eAAe,EACf,4BAA6B,KAAe,CAAC,OAAO,EAAE,EACtD,EAAE,KAAK,EAAG,KAAe,CAAC,OAAO,EAAE,CACpC,CAAC;QACJ,CAAC;IACH,CAAC;CACF;AAED,MAAM,OAAO,uBAAuB;IAC1B,WAAW,GAAG,KAAK,CAAC;IACX,OAAO,CAAS;IAChB,SAAS,CAAS;IAClB,MAAM,CAAU;IAChB,SAAS,CAAS;IAClB,iBAAiB,CAAS;IAE3C,YACE,OAAe,EACf,SAAiB,EACjB,SAAiB,EACjB,iBAAyB,EACzB,MAAe;QAEf,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC1C,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,IAAI,CAAC,iBAAiB,GAAG,iBAAiB,CAAC;QAC3C,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED,KAAK,CAAC,UAAU;QACd,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC;IAC1B,CAAC;IAED,aAAa;QACX,OAAO,IAAI,CAAC,WAAW,CAAC;IAC1B,CAAC;IAED,YAAY;QACV,OAAO,WAAW,IAAI,CAAC,SAAS,EAAE,CAAC;IACrC,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,SAAmB,EAAE,OAAuB;QACtE,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEtC,IAAI,CAAC;YACH,MAAM,kBAAkB,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,gBAAgB,CAAC,GAAG,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC,CAAC;YACjG,MAAM,SAAS,GAAG,IAAI,KAAK,CAAS,SAAS,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAE9D,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,kBAAkB,CAAC,MAAM,EAAE,MAAM,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;gBAClF,MAAM,KAAK,GAAG,kBAAkB,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC;gBACxE,MAAM,OAAO,GAA2B;oBACtC,cAAc,EAAE,kBAAkB;iBACnC,CAAC;gBACF,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;oBAChB,OAAO,CAAC,aAAa,GAAG,UAAU,IAAI,CAAC,MAAM,EAAE,CAAC;gBAClD,CAAC;gBAED,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,GAAG,IAAI,CAAC,OAAO,YAAY,EAAE;oBACxD,MAAM,EAAE,MAAM;oBACd,OAAO;oBACP,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;wBACnB,KAAK,EAAE,IAAI,CAAC,SAAS;wBACrB,KAAK;wBACL,SAAS,EAAE,KAAK;wBAChB,KAAK,EAAE,KAAK,CAAC,MAAM;qBACpB,CAAC;iBACH,CAAC,CAAC;gBAEH,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;oBACjB,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;oBACnC,OAAO,WAAW,CAChB,eAAe,EACf,wBAAwB,QAAQ,CAAC,MAAM,KAAK,IAAI,EAAE,CACnD,CAAC;gBACJ,CAAC;gBAED,MAAM,IAAI,GAAG,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAA0B,CAAC;gBAC9D,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,OAAO,IAAI,EAAE,CAAC;gBAC9C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBACzB,MAAM,WAAW,GAAG,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC;oBACxC,MAAM,KAAK,GAAG,IAAI,CAAC,eAAe,IAAI,IAAI,CAAC,KAAK,IAAI,CAAC,CAAC;oBACtD,SAAS,CAAC,WAAW,CAAC,GAAG,KAAK,CAAC;gBACjC,CAAC;YACH,CAAC;YAED,MAAM,gBAAgB,GAAG,eAAe,CAAC,SAAS,CAAC,CAAC;YACpD,MAAM,OAAO,GAAmB,gBAAgB,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;gBACtE,KAAK;gBACL,KAAK;gBACL,GAAG,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACpE,CAAC,CAAC,CAAC;YAEJ,OAAO,YAAY,CAAC,OAAO,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;QAC9C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,WAAW,CAChB,eAAe,EACf,6BAA8B,KAAe,CAAC,OAAO,EAAE,EACvD,EAAE,KAAK,EAAG,KAAe,CAAC,OAAO,EAAE,CACpC,CAAC;QACJ,CAAC;IACH,CAAC;CACF;AAED,MAAM,UAAU,sBAAsB,CACpC,MAAuB,EACvB,OAA2D;IAE3D,IAAI,CAAC,MAAM,CAAC,OAAO,IAAI,MAAM,CAAC,QAAQ,KAAK,MAAM,EAAE,CAAC;QAClD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,QAAQ,MAAM,CAAC,QAAQ,EAAE,CAAC;QACxB,KAAK,MAAM;YACT,OAAO,IAAI,oBAAoB,CAAC,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,iBAAiB,CAAC,CAAC;QAC9E,KAAK,QAAQ;YACX,IAAI,CAAC,OAAO,EAAE,YAAY,EAAE,CAAC;gBAC3B,MAAM,IAAI,KAAK,CAAC,uCAAuC,CAAC,CAAC;YAC3D,CAAC;YACD,OAAO,IAAI,sBAAsB,CAC/B,MAAM,CAAC,WAAW,EAClB,OAAO,CAAC,YAAY,EACpB,MAAM,CAAC,SAAS,EAChB,MAAM,CAAC,iBAAiB,CACzB,CAAC;QACJ,KAAK,SAAS;YACZ,OAAO,IAAI,uBAAuB,CAChC,MAAM,CAAC,cAAc,EACrB,MAAM,CAAC,kBAAkB,EACzB,MAAM,CAAC,SAAS,EAChB,MAAM,CAAC,iBAAiB,EACxB,OAAO,EAAE,aAAa,CACvB,CAAC;QACJ;YACE,OAAO,IAAI,CAAC;IAChB,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"reranking.test.d.ts","sourceRoot":"","sources":["../../src/indexer/reranking.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,130 @@
|
|
|
1
|
+
import { beforeEach, describe, expect, it, vi } from 'vitest';
|
|
2
|
+
import { WasmRerankerProvider, VoyageRerankerProvider, LocalAIRerankerProvider, normalizeScores, truncateDocument, batchArray, createRerankerProvider, } from './reranking.js';
|
|
3
|
+
import { isError } from '../types.js';
|
|
4
|
+
const { pipelineMock, classifierMock } = vi.hoisted(() => ({
|
|
5
|
+
pipelineMock: vi.fn(),
|
|
6
|
+
classifierMock: vi.fn(),
|
|
7
|
+
}));
|
|
8
|
+
vi.mock('@xenova/transformers', () => ({
|
|
9
|
+
pipeline: pipelineMock.mockImplementation(async () => classifierMock),
|
|
10
|
+
}));
|
|
11
|
+
describe('RerankerProvider', () => {
|
|
12
|
+
beforeEach(() => {
|
|
13
|
+
vi.clearAllMocks();
|
|
14
|
+
classifierMock.mockImplementation(async (input) => {
|
|
15
|
+
if (input.includes('doc-high'))
|
|
16
|
+
return [{ label: 'relevant', score: 0.9 }];
|
|
17
|
+
if (input.includes('doc-mid'))
|
|
18
|
+
return [{ label: 'relevant', score: 0.5 }];
|
|
19
|
+
return [{ label: 'relevant', score: 0.1 }];
|
|
20
|
+
});
|
|
21
|
+
});
|
|
22
|
+
describe('normalizeScores', () => {
|
|
23
|
+
it('normalizes scores to [0, 1] range', () => {
|
|
24
|
+
expect(normalizeScores([1, 2, 3])).toEqual([0, 0.5, 1]);
|
|
25
|
+
expect(normalizeScores([-1, 0, 1])).toEqual([0, 0.5, 1]);
|
|
26
|
+
});
|
|
27
|
+
it('handles equal scores', () => {
|
|
28
|
+
expect(normalizeScores([5, 5, 5])).toEqual([1, 1, 1]);
|
|
29
|
+
});
|
|
30
|
+
it('handles empty array', () => {
|
|
31
|
+
expect(normalizeScores([])).toEqual([]);
|
|
32
|
+
});
|
|
33
|
+
});
|
|
34
|
+
describe('helpers', () => {
|
|
35
|
+
it('truncates long documents before scoring', () => {
|
|
36
|
+
const longDoc = `${'a '.repeat(4000)}end`;
|
|
37
|
+
const truncated = truncateDocument(longDoc, 100);
|
|
38
|
+
expect(truncated.length).toBeLessThanOrEqual(100);
|
|
39
|
+
});
|
|
40
|
+
it('batches arrays', () => {
|
|
41
|
+
const batches = batchArray([1, 2, 3, 4, 5], 2);
|
|
42
|
+
expect(batches).toEqual([[1, 2], [3, 4], [5]]);
|
|
43
|
+
});
|
|
44
|
+
});
|
|
45
|
+
describe('WasmRerankerProvider', () => {
|
|
46
|
+
it('initializes lazily', async () => {
|
|
47
|
+
const provider = new WasmRerankerProvider('Xenova/ms-marco-MiniLM-L-6-v2', 2000);
|
|
48
|
+
expect(provider.isInitialized()).toBe(false);
|
|
49
|
+
await provider.rerank('query', ['doc-low']);
|
|
50
|
+
expect(provider.isInitialized()).toBe(true);
|
|
51
|
+
expect(pipelineMock).toHaveBeenCalledTimes(1);
|
|
52
|
+
});
|
|
53
|
+
it('returns normalized scores', async () => {
|
|
54
|
+
const provider = new WasmRerankerProvider('Xenova/ms-marco-MiniLM-L-6-v2', 2000);
|
|
55
|
+
const results = await provider.rerank('query', ['doc-low', 'doc-mid', 'doc-high']);
|
|
56
|
+
expect(isError(results)).toBe(false);
|
|
57
|
+
if (!isError(results)) {
|
|
58
|
+
results.forEach((r) => {
|
|
59
|
+
expect(r.score).toBeGreaterThanOrEqual(0);
|
|
60
|
+
expect(r.score).toBeLessThanOrEqual(1);
|
|
61
|
+
});
|
|
62
|
+
expect(results[0].score).toBeGreaterThan(results[1].score);
|
|
63
|
+
}
|
|
64
|
+
});
|
|
65
|
+
});
|
|
66
|
+
describe('VoyageRerankerProvider', () => {
|
|
67
|
+
it('batches large document sets for API providers', async () => {
|
|
68
|
+
const fetchSpy = vi.fn(async (_url, init) => {
|
|
69
|
+
const body = JSON.parse(init?.body || '{}');
|
|
70
|
+
return new Response(JSON.stringify({
|
|
71
|
+
data: body.documents.map((_, index) => ({
|
|
72
|
+
index,
|
|
73
|
+
relevance_score: 1 - index / 100,
|
|
74
|
+
})),
|
|
75
|
+
}), { status: 200, headers: { 'Content-Type': 'application/json' } });
|
|
76
|
+
});
|
|
77
|
+
vi.stubGlobal('fetch', fetchSpy);
|
|
78
|
+
const provider = new VoyageRerankerProvider('rerank-2.5-lite', 'test-key', 10, 2000);
|
|
79
|
+
const docs = Array.from({ length: 25 }, (_, i) => `doc-${i}`);
|
|
80
|
+
const results = await provider.rerank('query', docs);
|
|
81
|
+
expect(isError(results)).toBe(false);
|
|
82
|
+
expect(fetchSpy).toHaveBeenCalledTimes(3);
|
|
83
|
+
vi.unstubAllGlobals();
|
|
84
|
+
});
|
|
85
|
+
});
|
|
86
|
+
describe('LocalAIRerankerProvider', () => {
|
|
87
|
+
it('normalizes scores from LocalAI responses', async () => {
|
|
88
|
+
const fetchSpy = vi.fn(async (_url, init) => {
|
|
89
|
+
const body = JSON.parse(init?.body || '{}');
|
|
90
|
+
return new Response(JSON.stringify({
|
|
91
|
+
results: body.documents.map((_, index) => ({
|
|
92
|
+
index,
|
|
93
|
+
score: (index + 1) * 10,
|
|
94
|
+
})),
|
|
95
|
+
}), { status: 200, headers: { 'Content-Type': 'application/json' } });
|
|
96
|
+
});
|
|
97
|
+
vi.stubGlobal('fetch', fetchSpy);
|
|
98
|
+
const provider = new LocalAIRerankerProvider('http://localhost:8080', 'jina-reranker-v1-base-en', 10, 2000);
|
|
99
|
+
const results = await provider.rerank('query', ['a', 'b']);
|
|
100
|
+
expect(isError(results)).toBe(false);
|
|
101
|
+
if (!isError(results)) {
|
|
102
|
+
expect(results[0].score).toBe(1);
|
|
103
|
+
expect(results[1].score).toBe(0);
|
|
104
|
+
}
|
|
105
|
+
vi.unstubAllGlobals();
|
|
106
|
+
});
|
|
107
|
+
});
|
|
108
|
+
describe('createRerankerProvider', () => {
|
|
109
|
+
const baseConfig = {
|
|
110
|
+
enabled: true,
|
|
111
|
+
provider: 'wasm',
|
|
112
|
+
topK: 50,
|
|
113
|
+
prerankThreshold: 0.3,
|
|
114
|
+
maxDocumentLength: 2000,
|
|
115
|
+
batchSize: 100,
|
|
116
|
+
wasmModel: 'Xenova/ms-marco-MiniLM-L-6-v2',
|
|
117
|
+
voyageModel: 'rerank-2.5-lite',
|
|
118
|
+
localaiBaseUrl: 'http://localhost:8080',
|
|
119
|
+
localaiRerankModel: 'jina-reranker-v1-base-en',
|
|
120
|
+
};
|
|
121
|
+
it('returns null when disabled', () => {
|
|
122
|
+
const provider = createRerankerProvider({ ...baseConfig, enabled: false, provider: 'none' });
|
|
123
|
+
expect(provider).toBeNull();
|
|
124
|
+
});
|
|
125
|
+
it('throws when voyage key missing', () => {
|
|
126
|
+
expect(() => createRerankerProvider({ ...baseConfig, provider: 'voyage' })).toThrow('Voyage API key required for reranking');
|
|
127
|
+
});
|
|
128
|
+
});
|
|
129
|
+
});
|
|
130
|
+
//# sourceMappingURL=reranking.test.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"reranking.test.js","sourceRoot":"","sources":["../../src/indexer/reranking.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAC;AAC9D,OAAO,EACL,oBAAoB,EACpB,sBAAsB,EACtB,uBAAuB,EACvB,eAAe,EACf,gBAAgB,EAChB,UAAU,EACV,sBAAsB,GACvB,MAAM,gBAAgB,CAAC;AACxB,OAAO,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AAGtC,MAAM,EAAE,YAAY,EAAE,cAAc,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE,CAAC,CAAC;IACzD,YAAY,EAAE,EAAE,CAAC,EAAE,EAAE;IACrB,cAAc,EAAE,EAAE,CAAC,EAAE,EAAE;CACxB,CAAC,CAAC,CAAC;AAEJ,EAAE,CAAC,IAAI,CAAC,sBAAsB,EAAE,GAAG,EAAE,CAAC,CAAC;IACrC,QAAQ,EAAE,YAAY,CAAC,kBAAkB,CAAC,KAAK,IAAI,EAAE,CAAC,cAAc,CAAC;CACtE,CAAC,CAAC,CAAC;AAEJ,QAAQ,CAAC,kBAAkB,EAAE,GAAG,EAAE;IAChC,UAAU,CAAC,GAAG,EAAE;QACd,EAAE,CAAC,aAAa,EAAE,CAAC;QACnB,cAAc,CAAC,kBAAkB,CAAC,KAAK,EAAE,KAAa,EAAE,EAAE;YACxD,IAAI,KAAK,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,OAAO,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YAC3E,IAAI,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC;gBAAE,OAAO,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YAC1E,OAAO,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;QAC7C,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,iBAAiB,EAAE,GAAG,EAAE;QAC/B,EAAE,CAAC,mCAAmC,EAAE,GAAG,EAAE;YAC3C,MAAM,CAAC,eAAe,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;YACxD,MAAM,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAC3D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,sBAAsB,EAAE,GAAG,EAAE;YAC9B,MAAM,CAAC,eAAe,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QACxD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qBAAqB,EAAE,GAAG,EAAE;YAC7B,MAAM,CAAC,eAAe,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC1C,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,SAAS,EAAE,GAAG,EAAE;QACvB,EAAE,CAAC,yCAAyC,EAAE,GAAG,EAAE;YACjD,MAAM,OAAO,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC;YAC1C,MAAM,SAAS,GAAG,gBAAgB,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;YACjD,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,mBAAmB,CAAC,GAAG,CAAC,CAAC;QACpD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,gBAAgB,EAAE,GAAG,EAAE;YACxB,MAAM,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YAC/C,MAAM,CAAC,OAAO,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACjD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,sBAAsB,EAAE,GAAG,EAAE;QACpC,EAAE,CAAC,oBAAoB,EAAE,KAAK,IAAI,EAAE;YAClC,MAAM,QAAQ,GAAG,IAAI,oBAAoB,CAAC,+BAA+B,EAAE,IAAI,CAAC,CAAC;YACjF,MAAM,CAAC,QAAQ,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC7C,MAAM,QAAQ,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC;YAC5C,MAAM,CAAC,QAAQ,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC5C,MAAM,CAAC,YAAY,CAAC,CAAC,qBAAqB,CAAC,CAAC,CAAC,CAAC;QAChD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,2BAA2B,EAAE,KAAK,IAAI,EAAE;YACzC,MAAM,QAAQ,GAAG,IAAI,oBAAoB,CAAC,+BAA+B,EAAE,IAAI,CAAC,CAAC;YACjF,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,SAAS,EAAE,SAAS,EAAE,UAAU,CAAC,CAAC,CAAC;YACnF,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACrC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;gBACtB,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE;oBACpB,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC;oBAC1C,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,CAAC;gBACzC,CAAC,CAAC,CAAC;gBACH,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;YAC7D,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,wBAAwB,EAAE,GAAG,EAAE;QACtC,EAAE,CAAC,+CAA+C,EAAE,KAAK,IAAI,EAAE;YAC7D,MAAM,QAAQ,GAAG,EAAE,CAAC,EAAE,CAAC,KAAK,EAAE,IAAY,EAAE,IAAkB,EAAE,EAAE;gBAChE,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAE,IAAI,EAAE,IAAe,IAAI,IAAI,CAA4B,CAAC;gBACnF,OAAO,IAAI,QAAQ,CACjB,IAAI,CAAC,SAAS,CAAC;oBACb,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;wBACtC,KAAK;wBACL,eAAe,EAAE,CAAC,GAAG,KAAK,GAAG,GAAG;qBACjC,CAAC,CAAC;iBACJ,CAAC,EACF,EAAE,MAAM,EAAE,GAAG,EAAE,OAAO,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,EAAE,CACjE,CAAC;YACJ,CAAC,CAAC,CAAC;YACH,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;YAEjC,MAAM,QAAQ,GAAG,IAAI,sBAAsB,CAAC,iBAAiB,EAAE,UAAU,EAAE,EAAE,EAAE,IAAI,CAAC,CAAC;YACrF,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAC9D,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,MAAM,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;YAErD,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACrC,MAAM,CAAC,QAAQ,CAAC,CAAC,qBAAqB,CAAC,CAAC,CAAC,CAAC;YAC1C,EAAE,CAAC,gBAAgB,EAAE,CAAC;QACxB,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,yBAAyB,EAAE,GAAG,EAAE;QACvC,EAAE,CAAC,0CAA0C,EAAE,KAAK,IAAI,EAAE;YACxD,MAAM,QAAQ,GAAG,EAAE,CAAC,EAAE,CAAC,KAAK,EAAE,IAAY,EAAE,IAAkB,EAAE,EAAE;gBAChE,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAE,IAAI,EAAE,IAAe,IAAI,IAAI,CAA4B,CAAC;gBACnF,OAAO,IAAI,QAAQ,CACjB,IAAI,CAAC,SAAS,CAAC;oBACb,OAAO,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;wBACzC,KAAK;wBACL,KAAK,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,EAAE;qBACxB,CAAC,CAAC;iBACJ,CAAC,EACF,EAAE,MAAM,EAAE,GAAG,EAAE,OAAO,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,EAAE,CACjE,CAAC;YACJ,CAAC,CAAC,CAAC;YACH,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;YAEjC,MAAM,QAAQ,GAAG,IAAI,uBAAuB,CAC1C,uBAAuB,EACvB,0BAA0B,EAC1B,EAAE,EACF,IAAI,CACL,CAAC;YAEF,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;YAC3D,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACrC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;gBACtB,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBACjC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACnC,CAAC;YAED,EAAE,CAAC,gBAAgB,EAAE,CAAC;QACxB,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,wBAAwB,EAAE,GAAG,EAAE;QACtC,MAAM,UAAU,GAAoB;YAClC,OAAO,EAAE,IAAI;YACb,QAAQ,EAAE,MAAM;YAChB,IAAI,EAAE,EAAE;YACR,gBAAgB,EAAE,GAAG;YACrB,iBAAiB,EAAE,IAAI;YACvB,SAAS,EAAE,GAAG;YACd,SAAS,EAAE,+BAA+B;YAC1C,WAAW,EAAE,iBAAiB;YAC9B,cAAc,EAAE,uBAAuB;YACvC,kBAAkB,EAAE,0BAA0B;SAC/C,CAAC;QAEF,EAAE,CAAC,4BAA4B,EAAE,GAAG,EAAE;YACpC,MAAM,QAAQ,GAAG,sBAAsB,CAAC,EAAE,GAAG,UAAU,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC;YAC7F,MAAM,CAAC,QAAQ,CAAC,CAAC,QAAQ,EAAE,CAAC;QAC9B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,gCAAgC,EAAE,GAAG,EAAE;YACxC,MAAM,CAAC,GAAG,EAAE,CACV,sBAAsB,CAAC,EAAE,GAAG,UAAU,EAAE,QAAQ,EAAE,QAAQ,EAAE,CAAC,CAC9D,CAAC,OAAO,CAAC,uCAAuC,CAAC,CAAC;QACrD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,60 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* FileWalStorage - File-based implementation of WalStorage interface
|
|
3
|
+
*
|
|
4
|
+
* File structure:
|
|
5
|
+
* .lancedb/
|
|
6
|
+
* ├── sync.lock # Global lock with PID and heartbeat
|
|
7
|
+
* ├── sync.wal # Append-only log (bounded size via truncation)
|
|
8
|
+
* ├── sync.checkpoint # Last checkpoint snapshot (recovery point)
|
|
9
|
+
* └── sync.manifest # Current state + last completed task
|
|
10
|
+
*/
|
|
11
|
+
import type { WalStorage, LockData, WalEntry, CheckpointData, ManifestData } from './types.js';
|
|
12
|
+
export declare class FileWalStorage implements WalStorage {
|
|
13
|
+
private lockPath;
|
|
14
|
+
private walPath;
|
|
15
|
+
private checkpointPath;
|
|
16
|
+
private manifestPath;
|
|
17
|
+
constructor(dir: string);
|
|
18
|
+
acquireLock(lock: LockData): Promise<boolean>;
|
|
19
|
+
releaseLock(lockId: string): Promise<void>;
|
|
20
|
+
getLock(): Promise<LockData | null>;
|
|
21
|
+
updateHeartbeat(lockId: string): Promise<void>;
|
|
22
|
+
/**
|
|
23
|
+
* Force clear the lock file.
|
|
24
|
+
* Use with caution - only for stale lock recovery.
|
|
25
|
+
*/
|
|
26
|
+
forceClearLock(): Promise<void>;
|
|
27
|
+
/**
|
|
28
|
+
* Atomically clear lock only if it matches expected state.
|
|
29
|
+
* Prevents race condition where two processes detect stale lock simultaneously.
|
|
30
|
+
*/
|
|
31
|
+
compareAndClearLock(expectedLock: LockData): Promise<boolean>;
|
|
32
|
+
appendEntry(entry: WalEntry): Promise<void>;
|
|
33
|
+
getEntriesFromLsn(lsn: number): Promise<WalEntry[]>;
|
|
34
|
+
getAllEntries(): Promise<WalEntry[]>;
|
|
35
|
+
truncateBeforeLsn(lsn: number): Promise<void>;
|
|
36
|
+
clearWal(): Promise<void>;
|
|
37
|
+
writeCheckpoint(data: CheckpointData): Promise<void>;
|
|
38
|
+
getCheckpoint(): Promise<CheckpointData | null>;
|
|
39
|
+
clearCheckpoint(): Promise<void>;
|
|
40
|
+
writeManifest(data: ManifestData): Promise<void>;
|
|
41
|
+
getManifest(): Promise<ManifestData | null>;
|
|
42
|
+
clearManifest(): Promise<void>;
|
|
43
|
+
clear(): Promise<void>;
|
|
44
|
+
private atomicWrite;
|
|
45
|
+
private readJsonFile;
|
|
46
|
+
/**
|
|
47
|
+
* Check if PID is still running (for stale lock detection)
|
|
48
|
+
*/
|
|
49
|
+
static isPidRunning(pid: number): boolean;
|
|
50
|
+
/**
|
|
51
|
+
* Get paths for debugging/testing
|
|
52
|
+
*/
|
|
53
|
+
getPaths(): {
|
|
54
|
+
lock: string;
|
|
55
|
+
wal: string;
|
|
56
|
+
checkpoint: string;
|
|
57
|
+
manifest: string;
|
|
58
|
+
};
|
|
59
|
+
}
|
|
60
|
+
//# sourceMappingURL=file-storage.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"file-storage.d.ts","sourceRoot":"","sources":["../../../src/indexer/wal/file-storage.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAIH,OAAO,KAAK,EACV,UAAU,EACV,QAAQ,EACR,QAAQ,EACR,cAAc,EACd,YAAY,EACb,MAAM,YAAY,CAAC;AAEpB,qBAAa,cAAe,YAAW,UAAU;IAC/C,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,OAAO,CAAS;IACxB,OAAO,CAAC,cAAc,CAAS;IAC/B,OAAO,CAAC,YAAY,CAAS;gBAEjB,GAAG,EAAE,MAAM;IAWjB,WAAW,CAAC,IAAI,EAAE,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC;IAyB7C,WAAW,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAa1C,OAAO,IAAI,OAAO,CAAC,QAAQ,GAAG,IAAI,CAAC;IAYnC,eAAe,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAQpD;;;OAGG;IACG,cAAc,IAAI,OAAO,CAAC,IAAI,CAAC;IAUrC;;;OAGG;IACG,mBAAmB,CAAC,YAAY,EAAE,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC;IAgC7D,WAAW,CAAC,KAAK,EAAE,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC;IAI3C,iBAAiB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,QAAQ,EAAE,CAAC;IA2BnD,aAAa,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC;IAIpC,iBAAiB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAS7C,QAAQ,IAAI,OAAO,CAAC,IAAI,CAAC;IAczB,eAAe,CAAC,IAAI,EAAE,cAAc,GAAG,OAAO,CAAC,IAAI,CAAC;IAIpD,aAAa,IAAI,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC;IAI/C,eAAe,IAAI,OAAO,CAAC,IAAI,CAAC;IAchC,aAAa,CAAC,IAAI,EAAE,YAAY,GAAG,OAAO,CAAC,IAAI,CAAC;IAIhD,WAAW,IAAI,OAAO,CAAC,YAAY,GAAG,IAAI,CAAC;IAI3C,aAAa,IAAI,OAAO,CAAC,IAAI,CAAC;IAc9B,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;YAad,WAAW;YAMX,YAAY;IAkB1B;;OAEG;IACH,MAAM,CAAC,YAAY,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO;IAUzC;;OAEG;IACH,QAAQ,IAAI;QACV,IAAI,EAAE,MAAM,CAAC;QACb,GAAG,EAAE,MAAM,CAAC;QACZ,UAAU,EAAE,MAAM,CAAC;QACnB,QAAQ,EAAE,MAAM,CAAC;KAClB;CAQF"}
|