@angri450/zvec-grep 0.2.2
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 +201 -0
- package/README.md +319 -0
- package/README_CN.md +297 -0
- package/dist/authorization/index.d.ts +7 -0
- package/dist/authorization/index.js +7 -0
- package/dist/authorization/index.js.map +1 -0
- package/dist/authorization/manager.d.ts +8 -0
- package/dist/authorization/manager.js +19 -0
- package/dist/authorization/manager.js.map +1 -0
- package/dist/authorization/operation.d.ts +7 -0
- package/dist/authorization/operation.js +50 -0
- package/dist/authorization/operation.js.map +1 -0
- package/dist/authorization/planner.d.ts +21 -0
- package/dist/authorization/planner.js +97 -0
- package/dist/authorization/planner.js.map +1 -0
- package/dist/authorization/prompt.d.ts +12 -0
- package/dist/authorization/prompt.js +63 -0
- package/dist/authorization/prompt.js.map +1 -0
- package/dist/authorization/store.d.ts +20 -0
- package/dist/authorization/store.js +251 -0
- package/dist/authorization/store.js.map +1 -0
- package/dist/authorization/target.d.ts +15 -0
- package/dist/authorization/target.js +49 -0
- package/dist/authorization/target.js.map +1 -0
- package/dist/authorization/types.d.ts +61 -0
- package/dist/authorization/types.js +2 -0
- package/dist/authorization/types.js.map +1 -0
- package/dist/cli/args.d.ts +6 -0
- package/dist/cli/args.js +1255 -0
- package/dist/cli/args.js.map +1 -0
- package/dist/cli/auth.d.ts +21 -0
- package/dist/cli/auth.js +209 -0
- package/dist/cli/auth.js.map +1 -0
- package/dist/cli/commands.d.ts +4 -0
- package/dist/cli/commands.js +801 -0
- package/dist/cli/commands.js.map +1 -0
- package/dist/cli/errors.d.ts +9 -0
- package/dist/cli/errors.js +131 -0
- package/dist/cli/errors.js.map +1 -0
- package/dist/cli/format/context.d.ts +9 -0
- package/dist/cli/format/context.js +786 -0
- package/dist/cli/format/context.js.map +1 -0
- package/dist/cli/format/debug.d.ts +4 -0
- package/dist/cli/format/debug.js +60 -0
- package/dist/cli/format/debug.js.map +1 -0
- package/dist/cli/format/highlight.d.ts +3 -0
- package/dist/cli/format/highlight.js +106 -0
- package/dist/cli/format/highlight.js.map +1 -0
- package/dist/cli/format/progress.d.ts +13 -0
- package/dist/cli/format/progress.js +290 -0
- package/dist/cli/format/progress.js.map +1 -0
- package/dist/cli/format/range.d.ts +3 -0
- package/dist/cli/format/range.js +24 -0
- package/dist/cli/format/range.js.map +1 -0
- package/dist/cli/format/status.d.ts +83 -0
- package/dist/cli/format/status.js +530 -0
- package/dist/cli/format/status.js.map +1 -0
- package/dist/cli/help.d.ts +1 -0
- package/dist/cli/help.js +546 -0
- package/dist/cli/help.js.map +1 -0
- package/dist/cli/index.d.ts +2 -0
- package/dist/cli/index.js +40 -0
- package/dist/cli/index.js.map +1 -0
- package/dist/cli/install.d.ts +4 -0
- package/dist/cli/install.js +1735 -0
- package/dist/cli/install.js.map +1 -0
- package/dist/cli/managed-rg.d.ts +10 -0
- package/dist/cli/managed-rg.js +337 -0
- package/dist/cli/managed-rg.js.map +1 -0
- package/dist/cli/server-search.d.ts +3 -0
- package/dist/cli/server-search.js +12 -0
- package/dist/cli/server-search.js.map +1 -0
- package/dist/cli/types.d.ts +87 -0
- package/dist/cli/types.js +10 -0
- package/dist/cli/types.js.map +1 -0
- package/dist/cli/version.d.ts +1 -0
- package/dist/cli/version.js +13 -0
- package/dist/cli/version.js.map +1 -0
- package/dist/client/daemon-client.d.ts +31 -0
- package/dist/client/daemon-client.js +203 -0
- package/dist/client/daemon-client.js.map +1 -0
- package/dist/client/mode-router.d.ts +9 -0
- package/dist/client/mode-router.js +28 -0
- package/dist/client/mode-router.js.map +1 -0
- package/dist/client/search-policy.d.ts +11 -0
- package/dist/client/search-policy.js +15 -0
- package/dist/client/search-policy.js.map +1 -0
- package/dist/daemon/backend.d.ts +78 -0
- package/dist/daemon/backend.js +1118 -0
- package/dist/daemon/backend.js.map +1 -0
- package/dist/daemon/change-set.d.ts +30 -0
- package/dist/daemon/change-set.js +163 -0
- package/dist/daemon/change-set.js.map +1 -0
- package/dist/daemon/config.d.ts +27 -0
- package/dist/daemon/config.js +107 -0
- package/dist/daemon/config.js.map +1 -0
- package/dist/daemon/errors.d.ts +5 -0
- package/dist/daemon/errors.js +11 -0
- package/dist/daemon/errors.js.map +1 -0
- package/dist/daemon/http-server.d.ts +30 -0
- package/dist/daemon/http-server.js +307 -0
- package/dist/daemon/http-server.js.map +1 -0
- package/dist/daemon/index-coordinator.d.ts +22 -0
- package/dist/daemon/index-coordinator.js +55 -0
- package/dist/daemon/index-coordinator.js.map +1 -0
- package/dist/daemon/job-scheduler.d.ts +85 -0
- package/dist/daemon/job-scheduler.js +492 -0
- package/dist/daemon/job-scheduler.js.map +1 -0
- package/dist/daemon/logger.d.ts +10 -0
- package/dist/daemon/logger.js +101 -0
- package/dist/daemon/logger.js.map +1 -0
- package/dist/daemon/model-pool.d.ts +48 -0
- package/dist/daemon/model-pool.js +190 -0
- package/dist/daemon/model-pool.js.map +1 -0
- package/dist/daemon/root-lease.d.ts +14 -0
- package/dist/daemon/root-lease.js +197 -0
- package/dist/daemon/root-lease.js.map +1 -0
- package/dist/daemon/root-runtime.d.ts +80 -0
- package/dist/daemon/root-runtime.js +340 -0
- package/dist/daemon/root-runtime.js.map +1 -0
- package/dist/daemon/runtime-manager.d.ts +47 -0
- package/dist/daemon/runtime-manager.js +252 -0
- package/dist/daemon/runtime-manager.js.map +1 -0
- package/dist/daemon/runtime.d.ts +12 -0
- package/dist/daemon/runtime.js +105 -0
- package/dist/daemon/runtime.js.map +1 -0
- package/dist/daemon/server-controller.d.ts +38 -0
- package/dist/daemon/server-controller.js +327 -0
- package/dist/daemon/server-controller.js.map +1 -0
- package/dist/daemon/watch-manager.d.ts +62 -0
- package/dist/daemon/watch-manager.js +415 -0
- package/dist/daemon/watch-manager.js.map +1 -0
- package/dist/daemon/workspace-read-session-cache.d.ts +31 -0
- package/dist/daemon/workspace-read-session-cache.js +115 -0
- package/dist/daemon/workspace-read-session-cache.js.map +1 -0
- package/dist/engine/code-formats.d.ts +3 -0
- package/dist/engine/code-formats.js +14 -0
- package/dist/engine/code-formats.js.map +1 -0
- package/dist/engine/config.d.ts +57 -0
- package/dist/engine/config.js +351 -0
- package/dist/engine/config.js.map +1 -0
- package/dist/engine/errors.d.ts +20 -0
- package/dist/engine/errors.js +51 -0
- package/dist/engine/errors.js.map +1 -0
- package/dist/engine/extraction/code/adapter.d.ts +19 -0
- package/dist/engine/extraction/code/adapter.js +24 -0
- package/dist/engine/extraction/code/adapter.js.map +1 -0
- package/dist/engine/extraction/code/extractor.d.ts +13 -0
- package/dist/engine/extraction/code/extractor.js +696 -0
- package/dist/engine/extraction/code/extractor.js.map +1 -0
- package/dist/engine/extraction/code/families/c-family.d.ts +2 -0
- package/dist/engine/extraction/code/families/c-family.js +122 -0
- package/dist/engine/extraction/code/families/c-family.js.map +1 -0
- package/dist/engine/extraction/code/families/js-ts.d.ts +12 -0
- package/dist/engine/extraction/code/families/js-ts.js +118 -0
- package/dist/engine/extraction/code/families/js-ts.js.map +1 -0
- package/dist/engine/extraction/code/families/metadata.d.ts +8 -0
- package/dist/engine/extraction/code/families/metadata.js +86 -0
- package/dist/engine/extraction/code/families/metadata.js.map +1 -0
- package/dist/engine/extraction/code/families/name-field.d.ts +3 -0
- package/dist/engine/extraction/code/families/name-field.js +22 -0
- package/dist/engine/extraction/code/families/name-field.js.map +1 -0
- package/dist/engine/extraction/code/languages/c.d.ts +1 -0
- package/dist/engine/extraction/code/languages/c.js +12 -0
- package/dist/engine/extraction/code/languages/c.js.map +1 -0
- package/dist/engine/extraction/code/languages/cpp.d.ts +1 -0
- package/dist/engine/extraction/code/languages/cpp.js +18 -0
- package/dist/engine/extraction/code/languages/cpp.js.map +1 -0
- package/dist/engine/extraction/code/languages/go.d.ts +2 -0
- package/dist/engine/extraction/code/languages/go.js +70 -0
- package/dist/engine/extraction/code/languages/go.js.map +1 -0
- package/dist/engine/extraction/code/languages/java.d.ts +1 -0
- package/dist/engine/extraction/code/languages/java.js +17 -0
- package/dist/engine/extraction/code/languages/java.js.map +1 -0
- package/dist/engine/extraction/code/languages/javascript.d.ts +2 -0
- package/dist/engine/extraction/code/languages/javascript.js +23 -0
- package/dist/engine/extraction/code/languages/javascript.js.map +1 -0
- package/dist/engine/extraction/code/languages/python.d.ts +2 -0
- package/dist/engine/extraction/code/languages/python.js +52 -0
- package/dist/engine/extraction/code/languages/python.js.map +1 -0
- package/dist/engine/extraction/code/languages/rust.d.ts +2 -0
- package/dist/engine/extraction/code/languages/rust.js +25 -0
- package/dist/engine/extraction/code/languages/rust.js.map +1 -0
- package/dist/engine/extraction/code/languages/typescript.d.ts +2 -0
- package/dist/engine/extraction/code/languages/typescript.js +37 -0
- package/dist/engine/extraction/code/languages/typescript.js.map +1 -0
- package/dist/engine/extraction/code/tree-sitter/grammar.d.ts +4 -0
- package/dist/engine/extraction/code/tree-sitter/grammar.js +49 -0
- package/dist/engine/extraction/code/tree-sitter/grammar.js.map +1 -0
- package/dist/engine/extraction/code/tree-sitter/nodes.d.ts +3 -0
- package/dist/engine/extraction/code/tree-sitter/nodes.js +27 -0
- package/dist/engine/extraction/code/tree-sitter/nodes.js.map +1 -0
- package/dist/engine/extraction/code/tree-sitter/parser.d.ts +2 -0
- package/dist/engine/extraction/code/tree-sitter/parser.js +31 -0
- package/dist/engine/extraction/code/tree-sitter/parser.js.map +1 -0
- package/dist/engine/extraction/ids.d.ts +1 -0
- package/dist/engine/extraction/ids.js +5 -0
- package/dist/engine/extraction/ids.js.map +1 -0
- package/dist/engine/extraction/image/extractor.d.ts +5 -0
- package/dist/engine/extraction/image/extractor.js +35 -0
- package/dist/engine/extraction/image/extractor.js.map +1 -0
- package/dist/engine/extraction/index.d.ts +4 -0
- package/dist/engine/extraction/index.js +3 -0
- package/dist/engine/extraction/index.js.map +1 -0
- package/dist/engine/extraction/markdown/extractor.d.ts +7 -0
- package/dist/engine/extraction/markdown/extractor.js +362 -0
- package/dist/engine/extraction/markdown/extractor.js.map +1 -0
- package/dist/engine/extraction/runtime.d.ts +9 -0
- package/dist/engine/extraction/runtime.js +37 -0
- package/dist/engine/extraction/runtime.js.map +1 -0
- package/dist/engine/extraction/source.d.ts +18 -0
- package/dist/engine/extraction/source.js +22 -0
- package/dist/engine/extraction/source.js.map +1 -0
- package/dist/engine/extraction/text/extractor.d.ts +7 -0
- package/dist/engine/extraction/text/extractor.js +168 -0
- package/dist/engine/extraction/text/extractor.js.map +1 -0
- package/dist/engine/extraction/types.d.ts +4 -0
- package/dist/engine/extraction/types.js +2 -0
- package/dist/engine/extraction/types.js.map +1 -0
- package/dist/engine/extraction/vector-content.d.ts +5 -0
- package/dist/engine/extraction/vector-content.js +83 -0
- package/dist/engine/extraction/vector-content.js.map +1 -0
- package/dist/engine/file-size-policy.d.ts +6 -0
- package/dist/engine/file-size-policy.js +20 -0
- package/dist/engine/file-size-policy.js.map +1 -0
- package/dist/engine/file-type.d.ts +15 -0
- package/dist/engine/file-type.js +152 -0
- package/dist/engine/file-type.js.map +1 -0
- package/dist/engine/index-status.d.ts +9 -0
- package/dist/engine/index-status.js +38 -0
- package/dist/engine/index-status.js.map +1 -0
- package/dist/engine/manifest.d.ts +13 -0
- package/dist/engine/manifest.js +124 -0
- package/dist/engine/manifest.js.map +1 -0
- package/dist/engine/models/artifact-cache-lock.d.ts +16 -0
- package/dist/engine/models/artifact-cache-lock.js +228 -0
- package/dist/engine/models/artifact-cache-lock.js.map +1 -0
- package/dist/engine/models/artifact-downloader.d.ts +77 -0
- package/dist/engine/models/artifact-downloader.js +766 -0
- package/dist/engine/models/artifact-downloader.js.map +1 -0
- package/dist/engine/models/backends/llama-cpp.d.ts +100 -0
- package/dist/engine/models/backends/llama-cpp.js +567 -0
- package/dist/engine/models/backends/llama-cpp.js.map +1 -0
- package/dist/engine/models/backends/model2vec-runtime.d.ts +63 -0
- package/dist/engine/models/backends/model2vec-runtime.js +112 -0
- package/dist/engine/models/backends/model2vec-runtime.js.map +1 -0
- package/dist/engine/models/backends/model2vec-tokenizer.d.ts +2 -0
- package/dist/engine/models/backends/model2vec-tokenizer.js +42 -0
- package/dist/engine/models/backends/model2vec-tokenizer.js.map +1 -0
- package/dist/engine/models/backends/model2vec-worker-pool.d.ts +25 -0
- package/dist/engine/models/backends/model2vec-worker-pool.js +264 -0
- package/dist/engine/models/backends/model2vec-worker-pool.js.map +1 -0
- package/dist/engine/models/backends/model2vec-worker.d.ts +1 -0
- package/dist/engine/models/backends/model2vec-worker.js +41 -0
- package/dist/engine/models/backends/model2vec-worker.js.map +1 -0
- package/dist/engine/models/backends/model2vec.d.ts +38 -0
- package/dist/engine/models/backends/model2vec.js +344 -0
- package/dist/engine/models/backends/model2vec.js.map +1 -0
- package/dist/engine/models/backends/qwen.d.ts +44 -0
- package/dist/engine/models/backends/qwen.js +411 -0
- package/dist/engine/models/backends/qwen.js.map +1 -0
- package/dist/engine/models/backends/transformers-js.d.ts +90 -0
- package/dist/engine/models/backends/transformers-js.js +411 -0
- package/dist/engine/models/backends/transformers-js.js.map +1 -0
- package/dist/engine/models/catalog.d.ts +490 -0
- package/dist/engine/models/catalog.js +522 -0
- package/dist/engine/models/catalog.js.map +1 -0
- package/dist/engine/models/download-progress.d.ts +16 -0
- package/dist/engine/models/download-progress.js +47 -0
- package/dist/engine/models/download-progress.js.map +1 -0
- package/dist/engine/models/embeddings.d.ts +73 -0
- package/dist/engine/models/embeddings.js +125 -0
- package/dist/engine/models/embeddings.js.map +1 -0
- package/dist/engine/models/factory.d.ts +2 -0
- package/dist/engine/models/factory.js +51 -0
- package/dist/engine/models/factory.js.map +1 -0
- package/dist/engine/models/index.d.ts +5 -0
- package/dist/engine/models/index.js +6 -0
- package/dist/engine/models/index.js.map +1 -0
- package/dist/engine/models/local-embedding-parallelism.d.ts +15 -0
- package/dist/engine/models/local-embedding-parallelism.js +63 -0
- package/dist/engine/models/local-embedding-parallelism.js.map +1 -0
- package/dist/engine/models/local-embedding-queue.d.ts +12 -0
- package/dist/engine/models/local-embedding-queue.js +65 -0
- package/dist/engine/models/local-embedding-queue.js.map +1 -0
- package/dist/engine/models/ranking.d.ts +19 -0
- package/dist/engine/models/ranking.js +2 -0
- package/dist/engine/models/ranking.js.map +1 -0
- package/dist/engine/models/resolution.d.ts +8 -0
- package/dist/engine/models/resolution.js +22 -0
- package/dist/engine/models/resolution.js.map +1 -0
- package/dist/engine/pipeline/indexing/index.d.ts +14 -0
- package/dist/engine/pipeline/indexing/index.js +1219 -0
- package/dist/engine/pipeline/indexing/index.js.map +1 -0
- package/dist/engine/pipeline/indexing/input-budget.d.ts +2 -0
- package/dist/engine/pipeline/indexing/input-budget.js +51 -0
- package/dist/engine/pipeline/indexing/input-budget.js.map +1 -0
- package/dist/engine/pipeline/indexing/root-paths.d.ts +7 -0
- package/dist/engine/pipeline/indexing/root-paths.js +148 -0
- package/dist/engine/pipeline/indexing/root-paths.js.map +1 -0
- package/dist/engine/pipeline/indexing/scanner/index.d.ts +14 -0
- package/dist/engine/pipeline/indexing/scanner/index.js +776 -0
- package/dist/engine/pipeline/indexing/scanner/index.js.map +1 -0
- package/dist/engine/pipeline/search/index.d.ts +12 -0
- package/dist/engine/pipeline/search/index.js +811 -0
- package/dist/engine/pipeline/search/index.js.map +1 -0
- package/dist/engine/service/index.d.ts +3 -0
- package/dist/engine/service/index.js +2 -0
- package/dist/engine/service/index.js.map +1 -0
- package/dist/engine/service/lexical.d.ts +28 -0
- package/dist/engine/service/lexical.js +446 -0
- package/dist/engine/service/lexical.js.map +1 -0
- package/dist/engine/service/root.d.ts +15 -0
- package/dist/engine/service/root.js +57 -0
- package/dist/engine/service/root.js.map +1 -0
- package/dist/engine/service/structure-enrichment.d.ts +8 -0
- package/dist/engine/service/structure-enrichment.js +178 -0
- package/dist/engine/service/structure-enrichment.js.map +1 -0
- package/dist/engine/service/types.d.ts +212 -0
- package/dist/engine/service/types.js +2 -0
- package/dist/engine/service/types.js.map +1 -0
- package/dist/engine/service/workspace-index.d.ts +26 -0
- package/dist/engine/service/workspace-index.js +157 -0
- package/dist/engine/service/workspace-index.js.map +1 -0
- package/dist/engine/service/zvec-grep.d.ts +15 -0
- package/dist/engine/service/zvec-grep.js +1520 -0
- package/dist/engine/service/zvec-grep.js.map +1 -0
- package/dist/engine/storage/index.d.ts +53 -0
- package/dist/engine/storage/index.js +3 -0
- package/dist/engine/storage/index.js.map +1 -0
- package/dist/engine/storage/layout.d.ts +10 -0
- package/dist/engine/storage/layout.js +29 -0
- package/dist/engine/storage/layout.js.map +1 -0
- package/dist/engine/storage/zvec.d.ts +4 -0
- package/dist/engine/storage/zvec.js +976 -0
- package/dist/engine/storage/zvec.js.map +1 -0
- package/dist/engine/types.d.ts +303 -0
- package/dist/engine/types.js +14 -0
- package/dist/engine/types.js.map +1 -0
- package/dist/engine/utils/atomic-file.d.ts +3 -0
- package/dist/engine/utils/atomic-file.js +39 -0
- package/dist/engine/utils/atomic-file.js.map +1 -0
- package/dist/engine/utils/daemon-lease.d.ts +17 -0
- package/dist/engine/utils/daemon-lease.js +175 -0
- package/dist/engine/utils/daemon-lease.js.map +1 -0
- package/dist/engine/utils/file-selection.d.ts +10 -0
- package/dist/engine/utils/file-selection.js +157 -0
- package/dist/engine/utils/file-selection.js.map +1 -0
- package/dist/engine/utils/glob.d.ts +9 -0
- package/dist/engine/utils/glob.js +208 -0
- package/dist/engine/utils/glob.js.map +1 -0
- package/dist/engine/utils/hash.d.ts +2 -0
- package/dist/engine/utils/hash.js +8 -0
- package/dist/engine/utils/hash.js.map +1 -0
- package/dist/engine/utils/json.d.ts +8 -0
- package/dist/engine/utils/json.js +83 -0
- package/dist/engine/utils/json.js.map +1 -0
- package/dist/engine/utils/lock.d.ts +19 -0
- package/dist/engine/utils/lock.js +254 -0
- package/dist/engine/utils/lock.js.map +1 -0
- package/dist/engine/utils/path.d.ts +4 -0
- package/dist/engine/utils/path.js +18 -0
- package/dist/engine/utils/path.js.map +1 -0
- package/dist/engine/utils/timing.d.ts +19 -0
- package/dist/engine/utils/timing.js +87 -0
- package/dist/engine/utils/timing.js.map +1 -0
- package/dist/index-progress.d.ts +10 -0
- package/dist/index-progress.js +159 -0
- package/dist/index-progress.js.map +1 -0
- package/dist/index.d.ts +6 -0
- package/dist/index.js +4 -0
- package/dist/index.js.map +1 -0
- package/dist/mcp/http-transport.d.ts +27 -0
- package/dist/mcp/http-transport.js +182 -0
- package/dist/mcp/http-transport.js.map +1 -0
- package/dist/mcp/input-normalization.d.ts +38 -0
- package/dist/mcp/input-normalization.js +104 -0
- package/dist/mcp/input-normalization.js.map +1 -0
- package/dist/mcp/progress-heartbeat.d.ts +17 -0
- package/dist/mcp/progress-heartbeat.js +46 -0
- package/dist/mcp/progress-heartbeat.js.map +1 -0
- package/dist/mcp/request-metadata.d.ts +2 -0
- package/dist/mcp/request-metadata.js +11 -0
- package/dist/mcp/request-metadata.js.map +1 -0
- package/dist/mcp/request-state.d.ts +40 -0
- package/dist/mcp/request-state.js +210 -0
- package/dist/mcp/request-state.js.map +1 -0
- package/dist/mcp/result-format.d.ts +13 -0
- package/dist/mcp/result-format.js +75 -0
- package/dist/mcp/result-format.js.map +1 -0
- package/dist/mcp/schemas.d.ts +668 -0
- package/dist/mcp/schemas.js +512 -0
- package/dist/mcp/schemas.js.map +1 -0
- package/dist/mcp/stdio-bridge.d.ts +15 -0
- package/dist/mcp/stdio-bridge.js +190 -0
- package/dist/mcp/stdio-bridge.js.map +1 -0
- package/dist/mcp/tools.d.ts +153 -0
- package/dist/mcp/tools.js +488 -0
- package/dist/mcp/tools.js.map +1 -0
- package/dist/mcp/toolset.d.ts +5 -0
- package/dist/mcp/toolset.js +12 -0
- package/dist/mcp/toolset.js.map +1 -0
- package/dist/observability/trace-context.d.ts +11 -0
- package/dist/observability/trace-context.js +138 -0
- package/dist/observability/trace-context.js.map +1 -0
- package/dist/prompts/zvec-grep-guidance.d.ts +2 -0
- package/dist/prompts/zvec-grep-guidance.js +9 -0
- package/dist/prompts/zvec-grep-guidance.js.map +1 -0
- package/docs/01-agents.md +217 -0
- package/docs/02-cli.md +259 -0
- package/docs/03-mcp.md +185 -0
- package/docs/04-pipeline.md +234 -0
- package/docs/05-architecture.md +86 -0
- package/docs/06-server.md +279 -0
- package/docs/07-embedding.md +227 -0
- package/docs/08-roadmap.md +85 -0
- package/docs/README.md +53 -0
- package/package.json +87 -0
|
@@ -0,0 +1,112 @@
|
|
|
1
|
+
export async function embedModel2VecTexts(texts, tokenizer, table, maxInputTokens, normalize) {
|
|
2
|
+
const tokenized = await tokenizeModel2VecTexts(texts, tokenizer, maxInputTokens);
|
|
3
|
+
return {
|
|
4
|
+
vectors: embedModel2VecTokenLists(tokenized.tokenLists, table, normalize),
|
|
5
|
+
truncated: tokenized.truncated,
|
|
6
|
+
};
|
|
7
|
+
}
|
|
8
|
+
export async function tokenizeModel2VecTexts(texts, tokenizer, maxInputTokens) {
|
|
9
|
+
const truncatedInputIndexes = [];
|
|
10
|
+
const tokenLists = await Promise.all(texts.map(async (text, index) => {
|
|
11
|
+
const encoded = await tokenizer(text, {
|
|
12
|
+
add_special_tokens: false,
|
|
13
|
+
truncation: true,
|
|
14
|
+
max_length: maxInputTokens + 1,
|
|
15
|
+
});
|
|
16
|
+
const encodedTokenIds = Array.from(encoded.input_ids.data, Number);
|
|
17
|
+
if (encodedTokenIds.length > maxInputTokens) {
|
|
18
|
+
truncatedInputIndexes.push(index);
|
|
19
|
+
}
|
|
20
|
+
const unknownTokenId = tokenizer.unk_token_id;
|
|
21
|
+
return encodedTokenIds
|
|
22
|
+
.slice(0, maxInputTokens)
|
|
23
|
+
.filter((id) => unknownTokenId === undefined ||
|
|
24
|
+
unknownTokenId === null ||
|
|
25
|
+
id !== unknownTokenId);
|
|
26
|
+
}));
|
|
27
|
+
return {
|
|
28
|
+
tokenLists,
|
|
29
|
+
truncated: truncatedInputIndexes.sort((left, right) => left - right),
|
|
30
|
+
};
|
|
31
|
+
}
|
|
32
|
+
export function embedModel2VecTokenLists(tokenLists, table, normalize) {
|
|
33
|
+
return tokenLists.map((tokenIds) => embedStaticTokenList(tokenIds, table, normalize));
|
|
34
|
+
}
|
|
35
|
+
export function sharedStaticEmbeddingTable(table) {
|
|
36
|
+
if (table.data.buffer instanceof SharedArrayBuffer) {
|
|
37
|
+
return table;
|
|
38
|
+
}
|
|
39
|
+
const sharedBuffer = new SharedArrayBuffer(table.data.byteLength);
|
|
40
|
+
new Uint8Array(sharedBuffer).set(new Uint8Array(table.data.buffer, table.data.byteOffset, table.data.byteLength));
|
|
41
|
+
return {
|
|
42
|
+
...table,
|
|
43
|
+
data: table.dtype === "F16"
|
|
44
|
+
? new Uint16Array(sharedBuffer)
|
|
45
|
+
: new Float32Array(sharedBuffer),
|
|
46
|
+
};
|
|
47
|
+
}
|
|
48
|
+
export function staticEmbeddingTableFromWorkerData(data) {
|
|
49
|
+
return {
|
|
50
|
+
data: data.dtype === "F16"
|
|
51
|
+
? new Uint16Array(data.tableBuffer)
|
|
52
|
+
: new Float32Array(data.tableBuffer),
|
|
53
|
+
dimension: data.dimension,
|
|
54
|
+
dtype: data.dtype,
|
|
55
|
+
rows: data.rows,
|
|
56
|
+
};
|
|
57
|
+
}
|
|
58
|
+
function embedStaticTokenList(tokenIds, table, normalize) {
|
|
59
|
+
const dimension = table.dimension;
|
|
60
|
+
const halfValues = table.dtype === "F16" ? halfFloatValues() : null;
|
|
61
|
+
const vector = Array.from({ length: dimension }, () => 0);
|
|
62
|
+
if (tokenIds.length === 0) {
|
|
63
|
+
return vector;
|
|
64
|
+
}
|
|
65
|
+
for (let index = 0; index < tokenIds.length; index++) {
|
|
66
|
+
const tokenId = tokenIds[index];
|
|
67
|
+
if (!Number.isInteger(tokenId) || tokenId < 0 || tokenId >= table.rows) {
|
|
68
|
+
throw new Error(`Tokenizer returned out-of-range token id: id=${tokenId} rows=${table.rows}`);
|
|
69
|
+
}
|
|
70
|
+
const start = tokenId * dimension;
|
|
71
|
+
for (let column = 0; column < dimension; column++) {
|
|
72
|
+
const value = table.data[start + column];
|
|
73
|
+
vector[column] += halfValues ? halfValues[value] : value;
|
|
74
|
+
}
|
|
75
|
+
}
|
|
76
|
+
let squaredNorm = 0;
|
|
77
|
+
for (let column = 0; column < dimension; column++) {
|
|
78
|
+
vector[column] /= tokenIds.length;
|
|
79
|
+
squaredNorm += vector[column] * vector[column];
|
|
80
|
+
}
|
|
81
|
+
if (normalize && squaredNorm > 0) {
|
|
82
|
+
const inverseNorm = 1 / Math.sqrt(squaredNorm);
|
|
83
|
+
for (let column = 0; column < dimension; column++) {
|
|
84
|
+
vector[column] *= inverseNorm;
|
|
85
|
+
}
|
|
86
|
+
}
|
|
87
|
+
return vector;
|
|
88
|
+
}
|
|
89
|
+
let cachedHalfFloatValues = null;
|
|
90
|
+
function halfFloatValues() {
|
|
91
|
+
if (cachedHalfFloatValues) {
|
|
92
|
+
return cachedHalfFloatValues;
|
|
93
|
+
}
|
|
94
|
+
const values = new Float32Array(65_536);
|
|
95
|
+
for (let bits = 0; bits < values.length; bits++) {
|
|
96
|
+
const sign = bits & 0x8000 ? -1 : 1;
|
|
97
|
+
const exponent = (bits >> 10) & 0x1f;
|
|
98
|
+
const fraction = bits & 0x03ff;
|
|
99
|
+
if (exponent === 0) {
|
|
100
|
+
values[bits] = sign * 2 ** -14 * (fraction / 1024);
|
|
101
|
+
}
|
|
102
|
+
else if (exponent === 0x1f) {
|
|
103
|
+
values[bits] = fraction === 0 ? sign * Infinity : Number.NaN;
|
|
104
|
+
}
|
|
105
|
+
else {
|
|
106
|
+
values[bits] = sign * 2 ** (exponent - 15) * (1 + fraction / 1024);
|
|
107
|
+
}
|
|
108
|
+
}
|
|
109
|
+
cachedHalfFloatValues = values;
|
|
110
|
+
return values;
|
|
111
|
+
}
|
|
112
|
+
//# sourceMappingURL=model2vec-runtime.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"model2vec-runtime.js","sourceRoot":"","sources":["../../../../src/engine/models/backends/model2vec-runtime.ts"],"names":[],"mappings":"AAsEA,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,KAAwB,EACxB,SAAwB,EACxB,KAA2B,EAC3B,cAAsB,EACtB,SAAkB;IAElB,MAAM,SAAS,GAAG,MAAM,sBAAsB,CAC5C,KAAK,EACL,SAAS,EACT,cAAc,CACf,CAAC;IAEF,OAAO;QACL,OAAO,EAAE,wBAAwB,CAAC,SAAS,CAAC,UAAU,EAAE,KAAK,EAAE,SAAS,CAAC;QACzE,SAAS,EAAE,SAAS,CAAC,SAAS;KAC/B,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,KAAwB,EACxB,SAAwB,EACxB,cAAsB;IAEtB,MAAM,qBAAqB,GAAa,EAAE,CAAC;IAC3C,MAAM,UAAU,GAAG,MAAM,OAAO,CAAC,GAAG,CAClC,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE;QAC9B,MAAM,OAAO,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE;YACpC,kBAAkB,EAAE,KAAK;YACzB,UAAU,EAAE,IAAI;YAChB,UAAU,EAAE,cAAc,GAAG,CAAC;SAC/B,CAAC,CAAC;QACH,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;QACnE,IAAI,eAAe,CAAC,MAAM,GAAG,cAAc,EAAE,CAAC;YAC5C,qBAAqB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;QACD,MAAM,cAAc,GAAG,SAAS,CAAC,YAAY,CAAC;QAC9C,OAAO,eAAe;aACnB,KAAK,CAAC,CAAC,EAAE,cAAc,CAAC;aACxB,MAAM,CACL,CAAC,EAAE,EAAE,EAAE,CACL,cAAc,KAAK,SAAS;YAC5B,cAAc,KAAK,IAAI;YACvB,EAAE,KAAK,cAAc,CACxB,CAAC;IACN,CAAC,CAAC,CACH,CAAC;IAEF,OAAO;QACL,UAAU;QACV,SAAS,EAAE,qBAAqB,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,GAAG,KAAK,CAAC;KACrE,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,wBAAwB,CACtC,UAA0C,EAC1C,KAA2B,EAC3B,SAAkB;IAElB,OAAO,UAAU,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CACjC,oBAAoB,CAAC,QAAQ,EAAE,KAAK,EAAE,SAAS,CAAC,CACjD,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,KAA2B;IAE3B,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,YAAY,iBAAiB,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACf,CAAC;IAED,MAAM,YAAY,GAAG,IAAI,iBAAiB,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAClE,IAAI,UAAU,CAAC,YAAY,CAAC,CAAC,GAAG,CAC9B,IAAI,UAAU,CACZ,KAAK,CAAC,IAAI,CAAC,MAAM,EACjB,KAAK,CAAC,IAAI,CAAC,UAAU,EACrB,KAAK,CAAC,IAAI,CAAC,UAAU,CACtB,CACF,CAAC;IACF,OAAO;QACL,GAAG,KAAK;QACR,IAAI,EACF,KAAK,CAAC,KAAK,KAAK,KAAK;YACnB,CAAC,CAAC,IAAI,WAAW,CAAC,YAAY,CAAC;YAC/B,CAAC,CAAC,IAAI,YAAY,CAAC,YAAY,CAAC;KACrC,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,kCAAkC,CAChD,IAAyB;IAEzB,OAAO;QACL,IAAI,EACF,IAAI,CAAC,KAAK,KAAK,KAAK;YAClB,CAAC,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,WAAW,CAAC;YACnC,CAAC,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC;QACxC,SAAS,EAAE,IAAI,CAAC,SAAS;QACzB,KAAK,EAAE,IAAI,CAAC,KAAK;QACjB,IAAI,EAAE,IAAI,CAAC,IAAI;KAChB,CAAC;AACJ,CAAC;AAED,SAAS,oBAAoB,CAC3B,QAA2B,EAC3B,KAA2B,EAC3B,SAAkB;IAElB,MAAM,SAAS,GAAG,KAAK,CAAC,SAAS,CAAC;IAClC,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC,CAAC,eAAe,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;IACpE,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;IAC1D,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC1B,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QACrD,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAChC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC,IAAI,OAAO,GAAG,CAAC,IAAI,OAAO,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;YACvE,MAAM,IAAI,KAAK,CACb,gDAAgD,OAAO,SAAS,KAAK,CAAC,IAAI,EAAE,CAC7E,CAAC;QACJ,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,GAAG,SAAS,CAAC;QAClC,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,SAAS,EAAE,MAAM,EAAE,EAAE,CAAC;YAClD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,MAAM,CAAC,CAAC;YACzC,MAAM,CAAC,MAAM,CAAC,IAAI,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;QAC3D,CAAC;IACH,CAAC;IAED,IAAI,WAAW,GAAG,CAAC,CAAC;IACpB,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,SAAS,EAAE,MAAM,EAAE,EAAE,CAAC;QAClD,MAAM,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC;QAClC,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;IACjD,CAAC;IACD,IAAI,SAAS,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;QACjC,MAAM,WAAW,GAAG,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;QAC/C,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,SAAS,EAAE,MAAM,EAAE,EAAE,CAAC;YAClD,MAAM,CAAC,MAAM,CAAC,IAAI,WAAW,CAAC;QAChC,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,IAAI,qBAAqB,GAAwB,IAAI,CAAC;AAEtD,SAAS,eAAe;IACtB,IAAI,qBAAqB,EAAE,CAAC;QAC1B,OAAO,qBAAqB,CAAC;IAC/B,CAAC;IACD,MAAM,MAAM,GAAG,IAAI,YAAY,CAAC,MAAM,CAAC,CAAC;IACxC,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACpC,MAAM,QAAQ,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC;QACrC,MAAM,QAAQ,GAAG,IAAI,GAAG,MAAM,CAAC;QAC/B,IAAI,QAAQ,KAAK,CAAC,EAAE,CAAC;YACnB,MAAM,CAAC,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,QAAQ,GAAG,IAAI,CAAC,CAAC;QACrD,CAAC;aAAM,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;YAC7B,MAAM,CAAC,IAAI,CAAC,GAAG,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC;QAC/D,CAAC;aAAM,CAAC;YACN,MAAM,CAAC,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,GAAG,QAAQ,GAAG,IAAI,CAAC,CAAC;QACrE,CAAC;IACH,CAAC;IACD,qBAAqB,GAAG,MAAM,CAAC;IAC/B,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -0,0 +1,42 @@
|
|
|
1
|
+
import { Tokenizer } from "@huggingface/tokenizers";
|
|
2
|
+
import { readFile } from "node:fs/promises";
|
|
3
|
+
import { join } from "node:path";
|
|
4
|
+
export async function loadModel2VecTokenizer(source) {
|
|
5
|
+
const [tokenizerJson, tokenizerConfig] = await Promise.all([
|
|
6
|
+
readJson(join(source, "tokenizer.json")),
|
|
7
|
+
readJson(join(source, "tokenizer_config.json")),
|
|
8
|
+
]);
|
|
9
|
+
const tokenizer = new Tokenizer(tokenizerJson, tokenizerConfig);
|
|
10
|
+
const unknownTokenId = resolveUnknownTokenId(tokenizer, tokenizerJson);
|
|
11
|
+
const tokenize = (text, options) => {
|
|
12
|
+
const ids = tokenizer.encode(text, {
|
|
13
|
+
add_special_tokens: options.add_special_tokens,
|
|
14
|
+
}).ids;
|
|
15
|
+
return {
|
|
16
|
+
input_ids: {
|
|
17
|
+
data: ids.length > options.max_length
|
|
18
|
+
? ids.slice(0, options.max_length)
|
|
19
|
+
: ids,
|
|
20
|
+
},
|
|
21
|
+
};
|
|
22
|
+
};
|
|
23
|
+
return Object.assign(tokenize, { unk_token_id: unknownTokenId });
|
|
24
|
+
}
|
|
25
|
+
async function readJson(path) {
|
|
26
|
+
return JSON.parse(await readFile(path, "utf8"));
|
|
27
|
+
}
|
|
28
|
+
function resolveUnknownTokenId(tokenizer, tokenizerJson) {
|
|
29
|
+
const configuredId = tokenizerJson.model?.unk_id;
|
|
30
|
+
if (typeof configuredId === "number" && Number.isInteger(configuredId)) {
|
|
31
|
+
return configuredId;
|
|
32
|
+
}
|
|
33
|
+
const configuredToken = tokenizerJson.model?.unk_token;
|
|
34
|
+
if (typeof configuredToken !== "string") {
|
|
35
|
+
return null;
|
|
36
|
+
}
|
|
37
|
+
const tokenId = tokenizer.token_to_id(configuredToken);
|
|
38
|
+
return typeof tokenId === "number" && Number.isInteger(tokenId)
|
|
39
|
+
? tokenId
|
|
40
|
+
: null;
|
|
41
|
+
}
|
|
42
|
+
//# sourceMappingURL=model2vec-tokenizer.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"model2vec-tokenizer.js","sourceRoot":"","sources":["../../../../src/engine/models/backends/model2vec-tokenizer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAC;AACpD,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAUjC,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,MAAc;IAEd,MAAM,CAAC,aAAa,EAAE,eAAe,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACzD,QAAQ,CAAC,IAAI,CAAC,MAAM,EAAE,gBAAgB,CAAC,CAAC;QACxC,QAAQ,CAAC,IAAI,CAAC,MAAM,EAAE,uBAAuB,CAAC,CAAC;KAChD,CAAC,CAAC;IACH,MAAM,SAAS,GAAG,IAAI,SAAS,CAAC,aAAa,EAAE,eAAe,CAAC,CAAC;IAChE,MAAM,cAAc,GAAG,qBAAqB,CAAC,SAAS,EAAE,aAAa,CAAC,CAAC;IACvE,MAAM,QAAQ,GAAG,CACf,IAAY,EACZ,OAIC,EACD,EAAE;QACF,MAAM,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC,IAAI,EAAE;YACjC,kBAAkB,EAAE,OAAO,CAAC,kBAAkB;SAC/C,CAAC,CAAC,GAAG,CAAC;QACP,OAAO;YACL,SAAS,EAAE;gBACT,IAAI,EACF,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,UAAU;oBAC7B,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC;oBAClC,CAAC,CAAC,GAAG;aACV;SACF,CAAC;IACJ,CAAC,CAAC;IAEF,OAAO,MAAM,CAAC,MAAM,CAAC,QAAQ,EAAE,EAAE,YAAY,EAAE,cAAc,EAAE,CAAC,CAAC;AACnE,CAAC;AAED,KAAK,UAAU,QAAQ,CAAC,IAAY;IAClC,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAA4B,CAAC;AAC7E,CAAC;AAED,SAAS,qBAAqB,CAC5B,SAAoB,EACpB,aAA4B;IAE5B,MAAM,YAAY,GAAG,aAAa,CAAC,KAAK,EAAE,MAAM,CAAC;IACjD,IAAI,OAAO,YAAY,KAAK,QAAQ,IAAI,MAAM,CAAC,SAAS,CAAC,YAAY,CAAC,EAAE,CAAC;QACvE,OAAO,YAAY,CAAC;IACtB,CAAC;IAED,MAAM,eAAe,GAAG,aAAa,CAAC,KAAK,EAAE,SAAS,CAAC;IACvD,IAAI,OAAO,eAAe,KAAK,QAAQ,EAAE,CAAC;QACxC,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,OAAO,GAAY,SAAS,CAAC,WAAW,CAAC,eAAe,CAAC,CAAC;IAChE,OAAO,OAAO,OAAO,KAAK,QAAQ,IAAI,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC;QAC7D,CAAC,CAAC,OAAO;QACT,CAAC,CAAC,IAAI,CAAC;AACX,CAAC"}
|
|
@@ -0,0 +1,25 @@
|
|
|
1
|
+
import type { EmbeddingResult } from "../embeddings.js";
|
|
2
|
+
import type { Model2VecWorkerData } from "./model2vec-runtime.js";
|
|
3
|
+
export declare class Model2VecWorkerPool {
|
|
4
|
+
private readonly data;
|
|
5
|
+
private readonly workerUrl;
|
|
6
|
+
private disposed;
|
|
7
|
+
private nextJobId;
|
|
8
|
+
private readonly maxWorkers;
|
|
9
|
+
private readonly queue;
|
|
10
|
+
private readonly slots;
|
|
11
|
+
constructor(data: Model2VecWorkerData, maxWorkers?: number, workerUrl?: URL);
|
|
12
|
+
start(): Promise<void>;
|
|
13
|
+
run(texts: readonly string[], signal?: AbortSignal): Promise<EmbeddingResult>;
|
|
14
|
+
dispose(): Promise<void>;
|
|
15
|
+
private spawnWorker;
|
|
16
|
+
private handleMessage;
|
|
17
|
+
private dispatch;
|
|
18
|
+
private postJob;
|
|
19
|
+
private drain;
|
|
20
|
+
private nextQueuedJob;
|
|
21
|
+
private failSlot;
|
|
22
|
+
private resolveJob;
|
|
23
|
+
private rejectJob;
|
|
24
|
+
private ensureNotDisposed;
|
|
25
|
+
}
|
|
@@ -0,0 +1,264 @@
|
|
|
1
|
+
import { availableParallelism } from "node:os";
|
|
2
|
+
import { Worker } from "node:worker_threads";
|
|
3
|
+
export class Model2VecWorkerPool {
|
|
4
|
+
data;
|
|
5
|
+
workerUrl;
|
|
6
|
+
disposed = false;
|
|
7
|
+
nextJobId = 1;
|
|
8
|
+
maxWorkers;
|
|
9
|
+
queue = [];
|
|
10
|
+
slots = [];
|
|
11
|
+
constructor(data, maxWorkers = availableParallelism(), workerUrl = new URL("./model2vec-worker.js", import.meta.url)) {
|
|
12
|
+
this.data = data;
|
|
13
|
+
this.workerUrl = workerUrl;
|
|
14
|
+
this.maxWorkers = Math.max(1, Math.floor(maxWorkers));
|
|
15
|
+
}
|
|
16
|
+
async start() {
|
|
17
|
+
this.ensureNotDisposed();
|
|
18
|
+
if (this.slots.length === 0) {
|
|
19
|
+
await this.spawnWorker().readyPromise;
|
|
20
|
+
}
|
|
21
|
+
}
|
|
22
|
+
async run(texts, signal) {
|
|
23
|
+
this.ensureNotDisposed();
|
|
24
|
+
throwIfAborted(signal);
|
|
25
|
+
return await new Promise((resolve, reject) => {
|
|
26
|
+
const job = {
|
|
27
|
+
id: this.nextJobId++,
|
|
28
|
+
texts: [...texts],
|
|
29
|
+
signal,
|
|
30
|
+
resolve,
|
|
31
|
+
reject,
|
|
32
|
+
settled: false,
|
|
33
|
+
};
|
|
34
|
+
job.abort = () => {
|
|
35
|
+
if (job.settled) {
|
|
36
|
+
return;
|
|
37
|
+
}
|
|
38
|
+
job.settled = true;
|
|
39
|
+
const queuedIndex = this.queue.indexOf(job);
|
|
40
|
+
if (queuedIndex >= 0) {
|
|
41
|
+
this.queue.splice(queuedIndex, 1);
|
|
42
|
+
}
|
|
43
|
+
reject(abortError(signal));
|
|
44
|
+
};
|
|
45
|
+
signal?.addEventListener("abort", job.abort, { once: true });
|
|
46
|
+
const idle = this.slots.find((slot) => slot.ready && !slot.closed && !slot.job);
|
|
47
|
+
if (idle) {
|
|
48
|
+
this.dispatch(idle, job);
|
|
49
|
+
}
|
|
50
|
+
else if (this.slots.length < this.maxWorkers) {
|
|
51
|
+
const slot = this.spawnWorker(job);
|
|
52
|
+
void slot.readyPromise.catch(() => undefined);
|
|
53
|
+
}
|
|
54
|
+
else {
|
|
55
|
+
this.queue.push(job);
|
|
56
|
+
}
|
|
57
|
+
});
|
|
58
|
+
}
|
|
59
|
+
async dispose() {
|
|
60
|
+
if (this.disposed) {
|
|
61
|
+
return;
|
|
62
|
+
}
|
|
63
|
+
this.disposed = true;
|
|
64
|
+
const error = new Error("Model2Vec worker pool is disposed");
|
|
65
|
+
for (const job of this.queue.splice(0)) {
|
|
66
|
+
this.rejectJob(job, error);
|
|
67
|
+
}
|
|
68
|
+
for (const slot of this.slots) {
|
|
69
|
+
if (slot.job) {
|
|
70
|
+
this.rejectJob(slot.job, error);
|
|
71
|
+
slot.job = undefined;
|
|
72
|
+
}
|
|
73
|
+
slot.closed = true;
|
|
74
|
+
}
|
|
75
|
+
await Promise.allSettled(this.slots.map((slot) => slot.worker.terminate()));
|
|
76
|
+
this.slots.length = 0;
|
|
77
|
+
}
|
|
78
|
+
spawnWorker(job) {
|
|
79
|
+
const worker = new Worker(this.workerUrl, { workerData: this.data });
|
|
80
|
+
let resolveReady;
|
|
81
|
+
let rejectReady;
|
|
82
|
+
const readyPromise = new Promise((resolve, reject) => {
|
|
83
|
+
resolveReady = resolve;
|
|
84
|
+
rejectReady = reject;
|
|
85
|
+
});
|
|
86
|
+
const slot = {
|
|
87
|
+
worker,
|
|
88
|
+
ready: false,
|
|
89
|
+
closed: false,
|
|
90
|
+
job,
|
|
91
|
+
resolveReady,
|
|
92
|
+
rejectReady,
|
|
93
|
+
readyPromise,
|
|
94
|
+
};
|
|
95
|
+
this.slots.push(slot);
|
|
96
|
+
worker.on("message", (response) => {
|
|
97
|
+
this.handleMessage(slot, response);
|
|
98
|
+
});
|
|
99
|
+
worker.on("error", (error) => {
|
|
100
|
+
this.failSlot(slot, error);
|
|
101
|
+
});
|
|
102
|
+
worker.on("exit", (code) => {
|
|
103
|
+
if (!slot.closed) {
|
|
104
|
+
this.failSlot(slot, new Error(`Model2Vec worker exited unexpectedly with code ${code}`));
|
|
105
|
+
}
|
|
106
|
+
});
|
|
107
|
+
return slot;
|
|
108
|
+
}
|
|
109
|
+
handleMessage(slot, response) {
|
|
110
|
+
if (slot.closed) {
|
|
111
|
+
return;
|
|
112
|
+
}
|
|
113
|
+
if (response.type === "ready") {
|
|
114
|
+
slot.ready = true;
|
|
115
|
+
slot.resolveReady();
|
|
116
|
+
if (slot.job) {
|
|
117
|
+
this.postJob(slot, slot.job);
|
|
118
|
+
}
|
|
119
|
+
else {
|
|
120
|
+
slot.worker.unref();
|
|
121
|
+
this.drain();
|
|
122
|
+
}
|
|
123
|
+
return;
|
|
124
|
+
}
|
|
125
|
+
const job = slot.job;
|
|
126
|
+
if (!job || response.id !== job.id) {
|
|
127
|
+
this.failSlot(slot, new Error("Model2Vec worker response is out of order"));
|
|
128
|
+
return;
|
|
129
|
+
}
|
|
130
|
+
slot.job = undefined;
|
|
131
|
+
slot.worker.unref();
|
|
132
|
+
if (response.type === "error") {
|
|
133
|
+
this.rejectJob(job, workerError(response.error));
|
|
134
|
+
}
|
|
135
|
+
else {
|
|
136
|
+
try {
|
|
137
|
+
this.resolveJob(job, workerEmbeddingResult(response, this.data.dimension));
|
|
138
|
+
}
|
|
139
|
+
catch (error) {
|
|
140
|
+
this.rejectJob(job, error);
|
|
141
|
+
this.failSlot(slot, error);
|
|
142
|
+
return;
|
|
143
|
+
}
|
|
144
|
+
}
|
|
145
|
+
this.drain();
|
|
146
|
+
}
|
|
147
|
+
dispatch(slot, job) {
|
|
148
|
+
slot.worker.ref();
|
|
149
|
+
slot.job = job;
|
|
150
|
+
this.postJob(slot, job);
|
|
151
|
+
}
|
|
152
|
+
postJob(slot, job) {
|
|
153
|
+
if (job.settled) {
|
|
154
|
+
slot.job = undefined;
|
|
155
|
+
slot.worker.unref();
|
|
156
|
+
this.drain();
|
|
157
|
+
return;
|
|
158
|
+
}
|
|
159
|
+
slot.worker.postMessage({
|
|
160
|
+
id: job.id,
|
|
161
|
+
texts: job.texts,
|
|
162
|
+
});
|
|
163
|
+
}
|
|
164
|
+
drain() {
|
|
165
|
+
if (this.disposed) {
|
|
166
|
+
return;
|
|
167
|
+
}
|
|
168
|
+
for (const slot of this.slots) {
|
|
169
|
+
if (!slot.ready || slot.closed || slot.job) {
|
|
170
|
+
continue;
|
|
171
|
+
}
|
|
172
|
+
const job = this.nextQueuedJob();
|
|
173
|
+
if (!job) {
|
|
174
|
+
return;
|
|
175
|
+
}
|
|
176
|
+
this.dispatch(slot, job);
|
|
177
|
+
}
|
|
178
|
+
while (this.queue.length > 0 && this.slots.length < this.maxWorkers) {
|
|
179
|
+
const job = this.nextQueuedJob();
|
|
180
|
+
if (!job) {
|
|
181
|
+
return;
|
|
182
|
+
}
|
|
183
|
+
const slot = this.spawnWorker(job);
|
|
184
|
+
void slot.readyPromise.catch(() => undefined);
|
|
185
|
+
}
|
|
186
|
+
}
|
|
187
|
+
nextQueuedJob() {
|
|
188
|
+
while (this.queue.length > 0) {
|
|
189
|
+
const job = this.queue.shift();
|
|
190
|
+
if (job && !job.settled) {
|
|
191
|
+
return job;
|
|
192
|
+
}
|
|
193
|
+
}
|
|
194
|
+
return undefined;
|
|
195
|
+
}
|
|
196
|
+
failSlot(slot, error) {
|
|
197
|
+
if (slot.closed) {
|
|
198
|
+
return;
|
|
199
|
+
}
|
|
200
|
+
slot.closed = true;
|
|
201
|
+
slot.rejectReady(error);
|
|
202
|
+
if (slot.job) {
|
|
203
|
+
this.rejectJob(slot.job, error);
|
|
204
|
+
slot.job = undefined;
|
|
205
|
+
}
|
|
206
|
+
const index = this.slots.indexOf(slot);
|
|
207
|
+
if (index >= 0) {
|
|
208
|
+
this.slots.splice(index, 1);
|
|
209
|
+
}
|
|
210
|
+
void slot.worker.terminate().catch(() => undefined);
|
|
211
|
+
this.drain();
|
|
212
|
+
}
|
|
213
|
+
resolveJob(job, result) {
|
|
214
|
+
if (job.settled) {
|
|
215
|
+
return;
|
|
216
|
+
}
|
|
217
|
+
job.settled = true;
|
|
218
|
+
job.signal?.removeEventListener("abort", job.abort);
|
|
219
|
+
job.resolve(result);
|
|
220
|
+
}
|
|
221
|
+
rejectJob(job, error) {
|
|
222
|
+
if (job.settled) {
|
|
223
|
+
return;
|
|
224
|
+
}
|
|
225
|
+
job.settled = true;
|
|
226
|
+
job.signal?.removeEventListener("abort", job.abort);
|
|
227
|
+
job.reject(error);
|
|
228
|
+
}
|
|
229
|
+
ensureNotDisposed() {
|
|
230
|
+
if (this.disposed) {
|
|
231
|
+
throw new Error("Model2Vec worker pool is disposed");
|
|
232
|
+
}
|
|
233
|
+
}
|
|
234
|
+
}
|
|
235
|
+
function workerEmbeddingResult(response, dimension) {
|
|
236
|
+
const flat = new Float32Array(response.vectors);
|
|
237
|
+
if (flat.length !== response.vectorCount * dimension) {
|
|
238
|
+
throw new Error("Model2Vec worker returned an invalid vector buffer");
|
|
239
|
+
}
|
|
240
|
+
const vectors = [];
|
|
241
|
+
for (let index = 0; index < response.vectorCount; index++) {
|
|
242
|
+
vectors.push(Array.from(flat.subarray(index * dimension, (index + 1) * dimension)));
|
|
243
|
+
}
|
|
244
|
+
return { vectors, truncated: response.truncated };
|
|
245
|
+
}
|
|
246
|
+
function workerError(serialized) {
|
|
247
|
+
const error = new Error(serialized.message);
|
|
248
|
+
error.name = serialized.name ?? "Error";
|
|
249
|
+
if (serialized.stack) {
|
|
250
|
+
error.stack = serialized.stack;
|
|
251
|
+
}
|
|
252
|
+
return error;
|
|
253
|
+
}
|
|
254
|
+
function throwIfAborted(signal) {
|
|
255
|
+
if (signal?.aborted) {
|
|
256
|
+
throw abortError(signal);
|
|
257
|
+
}
|
|
258
|
+
}
|
|
259
|
+
function abortError(signal) {
|
|
260
|
+
return signal?.reason instanceof Error
|
|
261
|
+
? signal.reason
|
|
262
|
+
: new Error("Model2Vec embedding was cancelled");
|
|
263
|
+
}
|
|
264
|
+
//# sourceMappingURL=model2vec-worker-pool.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"model2vec-worker-pool.js","sourceRoot":"","sources":["../../../../src/engine/models/backends/model2vec-worker-pool.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,SAAS,CAAC;AAC/C,OAAO,EAAE,MAAM,EAAE,MAAM,qBAAqB,CAAC;AA6B7C,MAAM,OAAO,mBAAmB;IAQX;IAEA;IATX,QAAQ,GAAG,KAAK,CAAC;IACjB,SAAS,GAAG,CAAC,CAAC;IACL,UAAU,CAAS;IACnB,KAAK,GAAgB,EAAE,CAAC;IACxB,KAAK,GAAiB,EAAE,CAAC;IAE1C,YACmB,IAAyB,EAC1C,UAAU,GAAG,oBAAoB,EAAE,EAClB,YAAY,IAAI,GAAG,CAClC,uBAAuB,EACvB,MAAM,CAAC,IAAI,CAAC,GAAG,CAChB;QALgB,SAAI,GAAJ,IAAI,CAAqB;QAEzB,cAAS,GAAT,SAAS,CAGzB;QAED,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,KAAK,CAAC,KAAK;QACT,IAAI,CAAC,iBAAiB,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC5B,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC,YAAY,CAAC;QACxC,CAAC;IACH,CAAC;IAED,KAAK,CAAC,GAAG,CACP,KAAwB,EACxB,MAAoB;QAEpB,IAAI,CAAC,iBAAiB,EAAE,CAAC;QACzB,cAAc,CAAC,MAAM,CAAC,CAAC;QAEvB,OAAO,MAAM,IAAI,OAAO,CAAkB,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YAC5D,MAAM,GAAG,GAAc;gBACrB,EAAE,EAAE,IAAI,CAAC,SAAS,EAAE;gBACpB,KAAK,EAAE,CAAC,GAAG,KAAK,CAAC;gBACjB,MAAM;gBACN,OAAO;gBACP,MAAM;gBACN,OAAO,EAAE,KAAK;aACf,CAAC;YACF,GAAG,CAAC,KAAK,GAAG,GAAG,EAAE;gBACf,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;oBAChB,OAAO;gBACT,CAAC;gBACD,GAAG,CAAC,OAAO,GAAG,IAAI,CAAC;gBACnB,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;gBAC5C,IAAI,WAAW,IAAI,CAAC,EAAE,CAAC;oBACrB,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;gBACpC,CAAC;gBACD,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC;YAC7B,CAAC,CAAC;YACF,MAAM,EAAE,gBAAgB,CAAC,OAAO,EAAE,GAAG,CAAC,KAAK,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC;YAE7D,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAC1B,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,CAClD,CAAC;YACF,IAAI,IAAI,EAAE,CAAC;gBACT,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;YAC3B,CAAC;iBAAM,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;gBAC/C,MAAM,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;gBACnC,KAAK,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;YAChD,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACvB,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,KAAK,CAAC,OAAO;QACX,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QAC7D,KAAK,MAAM,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;YACvC,IAAI,CAAC,SAAS,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC7B,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;gBAChC,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC;YACvB,CAAC;YACD,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;QACrB,CAAC;QACD,MAAM,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC;QAC5E,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;IACxB,CAAC;IAEO,WAAW,CAAC,GAAe;QACjC,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,SAAS,EAAE,EAAE,UAAU,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;QACrE,IAAI,YAAyB,CAAC;QAC9B,IAAI,WAAsC,CAAC;QAC3C,MAAM,YAAY,GAAG,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACzD,YAAY,GAAG,OAAO,CAAC;YACvB,WAAW,GAAG,MAAM,CAAC;QACvB,CAAC,CAAC,CAAC;QACH,MAAM,IAAI,GAAe;YACvB,MAAM;YACN,KAAK,EAAE,KAAK;YACZ,MAAM,EAAE,KAAK;YACb,GAAG;YACH,YAAY;YACZ,WAAW;YACX,YAAY;SACb,CAAC;QACF,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtB,MAAM,CAAC,EAAE,CAAC,SAAS,EAAE,CAAC,QAAiC,EAAE,EAAE;YACzD,IAAI,CAAC,aAAa,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;QACrC,CAAC,CAAC,CAAC;QACH,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,KAAK,EAAE,EAAE;YAC3B,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC7B,CAAC,CAAC,CAAC;QACH,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,EAAE;YACzB,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;gBACjB,IAAI,CAAC,QAAQ,CACX,IAAI,EACJ,IAAI,KAAK,CAAC,kDAAkD,IAAI,EAAE,CAAC,CACpE,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,CAAC;QACH,OAAO,IAAI,CAAC;IACd,CAAC;IAEO,aAAa,CACnB,IAAgB,EAChB,QAAiC;QAEjC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,OAAO;QACT,CAAC;QACD,IAAI,QAAQ,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;YAClB,IAAI,CAAC,YAAY,EAAE,CAAC;YACpB,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC;YAC/B,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;gBACpB,IAAI,CAAC,KAAK,EAAE,CAAC;YACf,CAAC;YACD,OAAO;QACT,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;QACrB,IAAI,CAAC,GAAG,IAAI,QAAQ,CAAC,EAAE,KAAK,GAAG,CAAC,EAAE,EAAE,CAAC;YACnC,IAAI,CAAC,QAAQ,CACX,IAAI,EACJ,IAAI,KAAK,CAAC,2CAA2C,CAAC,CACvD,CAAC;YACF,OAAO;QACT,CAAC;QACD,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC;QACrB,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QACpB,IAAI,QAAQ,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,IAAI,CAAC,SAAS,CAAC,GAAG,EAAE,WAAW,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;QACnD,CAAC;aAAM,CAAC;YACN,IAAI,CAAC;gBACH,IAAI,CAAC,UAAU,CACb,GAAG,EACH,qBAAqB,CAAC,QAAQ,EAAE,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CACrD,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,CAAC,SAAS,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;gBAC3B,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;gBAC3B,OAAO;YACT,CAAC;QACH,CAAC;QACD,IAAI,CAAC,KAAK,EAAE,CAAC;IACf,CAAC;IAEO,QAAQ,CAAC,IAAgB,EAAE,GAAc;QAC/C,IAAI,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC;QAClB,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;QACf,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAC1B,CAAC;IAEO,OAAO,CAAC,IAAgB,EAAE,GAAc;QAC9C,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;YAChB,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC;YACrB,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;YACpB,IAAI,CAAC,KAAK,EAAE,CAAC;YACb,OAAO;QACT,CAAC;QACD,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC;YACtB,EAAE,EAAE,GAAG,CAAC,EAAE;YACV,KAAK,EAAE,GAAG,CAAC,KAAK;SACgB,CAAC,CAAC;IACtC,CAAC;IAEO,KAAK;QACX,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO;QACT,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;gBAC3C,SAAS;YACX,CAAC;YACD,MAAM,GAAG,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC;YACjC,IAAI,CAAC,GAAG,EAAE,CAAC;gBACT,OAAO;YACT,CAAC;YACD,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;QAC3B,CAAC;QACD,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;YACpE,MAAM,GAAG,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC;YACjC,IAAI,CAAC,GAAG,EAAE,CAAC;gBACT,OAAO;YACT,CAAC;YACD,MAAM,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACnC,KAAK,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;IAEO,aAAa;QACnB,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC7B,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;YAC/B,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC;gBACxB,OAAO,GAAG,CAAC;YACb,CAAC;QACH,CAAC;QACD,OAAO,SAAS,CAAC;IACnB,CAAC;IAEO,QAAQ,CAAC,IAAgB,EAAE,KAAc;QAC/C,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;QACnB,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC;QACxB,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;YACb,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YAChC,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC;QACvB,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACvC,IAAI,KAAK,IAAI,CAAC,EAAE,CAAC;YACf,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QAC9B,CAAC;QACD,KAAK,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;QACpD,IAAI,CAAC,KAAK,EAAE,CAAC;IACf,CAAC;IAEO,UAAU,CAAC,GAAc,EAAE,MAAuB;QACxD,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;YAChB,OAAO;QACT,CAAC;QACD,GAAG,CAAC,OAAO,GAAG,IAAI,CAAC;QACnB,GAAG,CAAC,MAAM,EAAE,mBAAmB,CAAC,OAAO,EAAE,GAAG,CAAC,KAAM,CAAC,CAAC;QACrD,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IACtB,CAAC;IAEO,SAAS,CAAC,GAAc,EAAE,KAAc;QAC9C,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;YAChB,OAAO;QACT,CAAC;QACD,GAAG,CAAC,OAAO,GAAG,IAAI,CAAC;QACnB,GAAG,CAAC,MAAM,EAAE,mBAAmB,CAAC,OAAO,EAAE,GAAG,CAAC,KAAM,CAAC,CAAC;QACrD,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACpB,CAAC;IAEO,iBAAiB;QACvB,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;QACvD,CAAC;IACH,CAAC;CACF;AAED,SAAS,qBAAqB,CAC5B,QAA8D,EAC9D,SAAiB;IAEjB,MAAM,IAAI,GAAG,IAAI,YAAY,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;IAChD,IAAI,IAAI,CAAC,MAAM,KAAK,QAAQ,CAAC,WAAW,GAAG,SAAS,EAAE,CAAC;QACrD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAC;IACxE,CAAC;IACD,MAAM,OAAO,GAAe,EAAE,CAAC;IAC/B,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,KAAK,EAAE,EAAE,CAAC;QAC1D,OAAO,CAAC,IAAI,CACV,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,GAAG,SAAS,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,CACtE,CAAC;IACJ,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,EAAE,CAAC;AACpD,CAAC;AAED,SAAS,WAAW,CAAC,UAAiC;IACpD,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;IAC5C,KAAK,CAAC,IAAI,GAAG,UAAU,CAAC,IAAI,IAAI,OAAO,CAAC;IACxC,IAAI,UAAU,CAAC,KAAK,EAAE,CAAC;QACrB,KAAK,CAAC,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC;IACjC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,cAAc,CAAC,MAAoB;IAC1C,IAAI,MAAM,EAAE,OAAO,EAAE,CAAC;QACpB,MAAM,UAAU,CAAC,MAAM,CAAC,CAAC;IAC3B,CAAC;AACH,CAAC;AAED,SAAS,UAAU,CAAC,MAAoB;IACtC,OAAO,MAAM,EAAE,MAAM,YAAY,KAAK;QACpC,CAAC,CAAC,MAAM,CAAC,MAAM;QACf,CAAC,CAAC,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;AACrD,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
export {};
|
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
import { parentPort, workerData } from "node:worker_threads";
|
|
2
|
+
import { loadModel2VecTokenizer } from "./model2vec-tokenizer.js";
|
|
3
|
+
import { embedModel2VecTexts, staticEmbeddingTableFromWorkerData, } from "./model2vec-runtime.js";
|
|
4
|
+
const port = parentPort;
|
|
5
|
+
if (!port) {
|
|
6
|
+
throw new Error("Model2Vec worker requires a parent port");
|
|
7
|
+
}
|
|
8
|
+
const data = workerData;
|
|
9
|
+
const tokenizer = await loadModel2VecTokenizer(data.tokenizerSource);
|
|
10
|
+
const table = staticEmbeddingTableFromWorkerData(data);
|
|
11
|
+
port.postMessage({ type: "ready" });
|
|
12
|
+
port.on("message", async (request) => {
|
|
13
|
+
try {
|
|
14
|
+
const result = await embedModel2VecTexts(request.texts, tokenizer, table, data.maxInputTokens, data.normalize);
|
|
15
|
+
const flatVectors = new Float32Array(result.vectors.length * data.dimension);
|
|
16
|
+
for (const [index, vector] of result.vectors.entries()) {
|
|
17
|
+
flatVectors.set(vector, index * data.dimension);
|
|
18
|
+
}
|
|
19
|
+
const response = {
|
|
20
|
+
type: "result",
|
|
21
|
+
id: request.id,
|
|
22
|
+
vectors: flatVectors.buffer,
|
|
23
|
+
vectorCount: result.vectors.length,
|
|
24
|
+
truncated: result.truncated,
|
|
25
|
+
};
|
|
26
|
+
port.postMessage(response, [flatVectors.buffer]);
|
|
27
|
+
}
|
|
28
|
+
catch (error) {
|
|
29
|
+
const cause = error instanceof Error ? error : new Error(String(error));
|
|
30
|
+
port.postMessage({
|
|
31
|
+
type: "error",
|
|
32
|
+
id: request.id,
|
|
33
|
+
error: {
|
|
34
|
+
name: cause.name,
|
|
35
|
+
message: cause.message,
|
|
36
|
+
stack: cause.stack,
|
|
37
|
+
},
|
|
38
|
+
});
|
|
39
|
+
}
|
|
40
|
+
});
|
|
41
|
+
//# sourceMappingURL=model2vec-worker.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"model2vec-worker.js","sourceRoot":"","sources":["../../../../src/engine/models/backends/model2vec-worker.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAC7D,OAAO,EAAE,sBAAsB,EAAE,MAAM,0BAA0B,CAAC;AAClE,OAAO,EACL,mBAAmB,EACnB,kCAAkC,GAInC,MAAM,wBAAwB,CAAC;AAEhC,MAAM,IAAI,GAAG,UAAU,CAAC;AACxB,IAAI,CAAC,IAAI,EAAE,CAAC;IACV,MAAM,IAAI,KAAK,CAAC,yCAAyC,CAAC,CAAC;AAC7D,CAAC;AAED,MAAM,IAAI,GAAG,UAAiC,CAAC;AAC/C,MAAM,SAAS,GAAG,MAAM,sBAAsB,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;AACrE,MAAM,KAAK,GAAG,kCAAkC,CAAC,IAAI,CAAC,CAAC;AAEvD,IAAI,CAAC,WAAW,CAAC,EAAE,IAAI,EAAE,OAAO,EAAoC,CAAC,CAAC;AACtE,IAAI,CAAC,EAAE,CAAC,SAAS,EAAE,KAAK,EAAE,OAA+B,EAAE,EAAE;IAC3D,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,mBAAmB,CACtC,OAAO,CAAC,KAAK,EACb,SAAS,EACT,KAAK,EACL,IAAI,CAAC,cAAc,EACnB,IAAI,CAAC,SAAS,CACf,CAAC;QACF,MAAM,WAAW,GAAG,IAAI,YAAY,CAClC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,SAAS,CACvC,CAAC;QACF,KAAK,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC;YACvD,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC;QAClD,CAAC;QACD,MAAM,QAAQ,GAA4B;YACxC,IAAI,EAAE,QAAQ;YACd,EAAE,EAAE,OAAO,CAAC,EAAE;YACd,OAAO,EAAE,WAAW,CAAC,MAAM;YAC3B,WAAW,EAAE,MAAM,CAAC,OAAO,CAAC,MAAM;YAClC,SAAS,EAAE,MAAM,CAAC,SAAS;SAC5B,CAAC;QACF,IAAI,CAAC,WAAW,CAAC,QAAQ,EAAE,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC;IACnD,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,KAAK,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;QACxE,IAAI,CAAC,WAAW,CAAC;YACf,IAAI,EAAE,OAAO;YACb,EAAE,EAAE,OAAO,CAAC,EAAE;YACd,KAAK,EAAE;gBACL,IAAI,EAAE,KAAK,CAAC,IAAI;gBAChB,OAAO,EAAE,KAAK,CAAC,OAAO;gBACtB,KAAK,EAAE,KAAK,CAAC,KAAK;aACnB;SACgC,CAAC,CAAC;IACvC,CAAC;AACH,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,38 @@
|
|
|
1
|
+
import type { Content } from "../../types.js";
|
|
2
|
+
import { BaseEmbeddingModel, type CreateEmbeddingModelOptions, type EmbeddingModelInfo, type EmbeddingOptions, type EmbeddingResult, type NormalizedEmbeddingOptions } from "../embeddings.js";
|
|
3
|
+
import type { Model2VecEmbeddingCatalogEntry } from "../catalog.js";
|
|
4
|
+
import { resolveModelArtifacts } from "../artifact-downloader.js";
|
|
5
|
+
import { type StaticEmbeddingTable, type TokenizerLike } from "./model2vec-runtime.js";
|
|
6
|
+
type Model2VecDependencies = {
|
|
7
|
+
loadTokenizer(repo: string, options: {
|
|
8
|
+
cache_dir: string;
|
|
9
|
+
revision: string;
|
|
10
|
+
local_files_only?: boolean;
|
|
11
|
+
}): Promise<TokenizerLike>;
|
|
12
|
+
loadSafetensors(path: string, tensorName: string, dimension: number): Promise<StaticEmbeddingTable>;
|
|
13
|
+
resolveArtifacts: typeof resolveModelArtifacts;
|
|
14
|
+
};
|
|
15
|
+
export declare class Model2VecEmbeddingModel extends BaseEmbeddingModel {
|
|
16
|
+
private readonly entry;
|
|
17
|
+
readonly info: EmbeddingModelInfo;
|
|
18
|
+
private readonly modelCacheDir;
|
|
19
|
+
private readonly dependencies;
|
|
20
|
+
private tokenizer;
|
|
21
|
+
private staticTable;
|
|
22
|
+
private workerPool;
|
|
23
|
+
private loadPromise;
|
|
24
|
+
private readonly useWorkerPool;
|
|
25
|
+
private sourceFallbackWarningReported;
|
|
26
|
+
private disposed;
|
|
27
|
+
constructor(entry: Model2VecEmbeddingCatalogEntry, options: CreateEmbeddingModelOptions, dependencies?: Partial<Model2VecDependencies>);
|
|
28
|
+
protected doEmbed(contents: readonly Content[], options: NormalizedEmbeddingOptions): Promise<EmbeddingResult>;
|
|
29
|
+
prepare(options?: Pick<EmbeddingOptions, "signal" | "onProgress">): Promise<void>;
|
|
30
|
+
private embedBatch;
|
|
31
|
+
dispose(): Promise<void>;
|
|
32
|
+
private ensureLoaded;
|
|
33
|
+
private loadModel;
|
|
34
|
+
private createArtifactSources;
|
|
35
|
+
private embedTexts;
|
|
36
|
+
private ensureNotDisposed;
|
|
37
|
+
}
|
|
38
|
+
export {};
|