opencode-rehydra-core 0.14.0
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 +22 -0
- package/README.en.md +191 -0
- package/README.md +191 -0
- package/dist/browser.d.ts +54 -0
- package/dist/browser.d.ts.map +1 -0
- package/dist/browser.js +74 -0
- package/dist/browser.js.map +1 -0
- package/dist/core/anonymizer.d.ts +309 -0
- package/dist/core/anonymizer.d.ts.map +1 -0
- package/dist/core/anonymizer.js +555 -0
- package/dist/core/anonymizer.js.map +1 -0
- package/dist/core/index.d.ts +6 -0
- package/dist/core/index.d.ts.map +1 -0
- package/dist/core/index.js +6 -0
- package/dist/core/index.js.map +1 -0
- package/dist/crypto/index.d.ts +6 -0
- package/dist/crypto/index.d.ts.map +1 -0
- package/dist/crypto/index.js +6 -0
- package/dist/crypto/index.js.map +1 -0
- package/dist/crypto/pii-map-crypto.d.ts +114 -0
- package/dist/crypto/pii-map-crypto.d.ts.map +1 -0
- package/dist/crypto/pii-map-crypto.js +228 -0
- package/dist/crypto/pii-map-crypto.js.map +1 -0
- package/dist/index.d.ts +53 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +72 -0
- package/dist/index.js.map +1 -0
- package/dist/ner/bio-decoder.d.ts +64 -0
- package/dist/ner/bio-decoder.d.ts.map +1 -0
- package/dist/ner/bio-decoder.js +218 -0
- package/dist/ner/bio-decoder.js.map +1 -0
- package/dist/ner/index.d.ts +12 -0
- package/dist/ner/index.d.ts.map +1 -0
- package/dist/ner/index.js +12 -0
- package/dist/ner/index.js.map +1 -0
- package/dist/ner/inference-server-client.d.ts +81 -0
- package/dist/ner/inference-server-client.d.ts.map +1 -0
- package/dist/ner/inference-server-client.js +120 -0
- package/dist/ner/inference-server-client.js.map +1 -0
- package/dist/ner/inference-server-ner-model.d.ts +43 -0
- package/dist/ner/inference-server-ner-model.d.ts.map +1 -0
- package/dist/ner/inference-server-ner-model.js +119 -0
- package/dist/ner/inference-server-ner-model.js.map +1 -0
- package/dist/ner/model-manager.d.ts +111 -0
- package/dist/ner/model-manager.d.ts.map +1 -0
- package/dist/ner/model-manager.js +334 -0
- package/dist/ner/model-manager.js.map +1 -0
- package/dist/ner/ner-model.d.ts +172 -0
- package/dist/ner/ner-model.d.ts.map +1 -0
- package/dist/ner/ner-model.js +514 -0
- package/dist/ner/ner-model.js.map +1 -0
- package/dist/ner/onnx-runtime.browser.d.ts +46 -0
- package/dist/ner/onnx-runtime.browser.d.ts.map +1 -0
- package/dist/ner/onnx-runtime.browser.js +90 -0
- package/dist/ner/onnx-runtime.browser.js.map +1 -0
- package/dist/ner/onnx-runtime.d.ts +72 -0
- package/dist/ner/onnx-runtime.d.ts.map +1 -0
- package/dist/ner/onnx-runtime.js +132 -0
- package/dist/ner/onnx-runtime.js.map +1 -0
- package/dist/ner/token-window.d.ts +19 -0
- package/dist/ner/token-window.d.ts.map +1 -0
- package/dist/ner/token-window.js +101 -0
- package/dist/ner/token-window.js.map +1 -0
- package/dist/ner/tokenizer.d.ts +142 -0
- package/dist/ner/tokenizer.d.ts.map +1 -0
- package/dist/ner/tokenizer.js +362 -0
- package/dist/ner/tokenizer.js.map +1 -0
- package/dist/opencode-plugin/git-identity.d.ts +3 -0
- package/dist/opencode-plugin/git-identity.d.ts.map +1 -0
- package/dist/opencode-plugin/git-identity.js +37 -0
- package/dist/opencode-plugin/git-identity.js.map +1 -0
- package/dist/opencode-plugin/github-username.d.ts +3 -0
- package/dist/opencode-plugin/github-username.d.ts.map +1 -0
- package/dist/opencode-plugin/github-username.js +41 -0
- package/dist/opencode-plugin/github-username.js.map +1 -0
- package/dist/opencode-plugin/identity-recognizer.d.ts +5 -0
- package/dist/opencode-plugin/identity-recognizer.d.ts.map +1 -0
- package/dist/opencode-plugin/identity-recognizer.js +21 -0
- package/dist/opencode-plugin/identity-recognizer.js.map +1 -0
- package/dist/opencode-plugin/index.d.ts +7 -0
- package/dist/opencode-plugin/index.d.ts.map +1 -0
- package/dist/opencode-plugin/index.js +6 -0
- package/dist/opencode-plugin/index.js.map +1 -0
- package/dist/opencode-plugin/plugin.d.ts +39 -0
- package/dist/opencode-plugin/plugin.d.ts.map +1 -0
- package/dist/opencode-plugin/plugin.js +294 -0
- package/dist/opencode-plugin/plugin.js.map +1 -0
- package/dist/opencode-plugin/types.d.ts +33 -0
- package/dist/opencode-plugin/types.d.ts.map +1 -0
- package/dist/opencode-plugin/types.js +5 -0
- package/dist/opencode-plugin/types.js.map +1 -0
- package/dist/opencode-plugin/v2/anon-state.d.ts +12 -0
- package/dist/opencode-plugin/v2/anon-state.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/anon-state.js +14 -0
- package/dist/opencode-plugin/v2/anon-state.js.map +1 -0
- package/dist/opencode-plugin/v2/deep-rehydrate.d.ts +13 -0
- package/dist/opencode-plugin/v2/deep-rehydrate.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/deep-rehydrate.js +33 -0
- package/dist/opencode-plugin/v2/deep-rehydrate.js.map +1 -0
- package/dist/opencode-plugin/v2/index.d.ts +17 -0
- package/dist/opencode-plugin/v2/index.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/index.js +17 -0
- package/dist/opencode-plugin/v2/index.js.map +1 -0
- package/dist/opencode-plugin/v2/logger.d.ts +24 -0
- package/dist/opencode-plugin/v2/logger.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/logger.js +73 -0
- package/dist/opencode-plugin/v2/logger.js.map +1 -0
- package/dist/opencode-plugin/v2/options.d.ts +40 -0
- package/dist/opencode-plugin/v2/options.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/options.js +132 -0
- package/dist/opencode-plugin/v2/options.js.map +1 -0
- package/dist/opencode-plugin/v2/register-hooks.d.ts +43 -0
- package/dist/opencode-plugin/v2/register-hooks.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/register-hooks.js +160 -0
- package/dist/opencode-plugin/v2/register-hooks.js.map +1 -0
- package/dist/opencode-plugin/v2/rehydrate-tail.d.ts +101 -0
- package/dist/opencode-plugin/v2/rehydrate-tail.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/rehydrate-tail.js +147 -0
- package/dist/opencode-plugin/v2/rehydrate-tail.js.map +1 -0
- package/dist/opencode-plugin/v2/response-rehydrate.d.ts +78 -0
- package/dist/opencode-plugin/v2/response-rehydrate.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/response-rehydrate.js +210 -0
- package/dist/opencode-plugin/v2/response-rehydrate.js.map +1 -0
- package/dist/opencode-plugin/v2/scrub-messages.d.ts +39 -0
- package/dist/opencode-plugin/v2/scrub-messages.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/scrub-messages.js +260 -0
- package/dist/opencode-plugin/v2/scrub-messages.js.map +1 -0
- package/dist/opencode-plugin/v2/sessions.d.ts +17 -0
- package/dist/opencode-plugin/v2/sessions.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/sessions.js +23 -0
- package/dist/opencode-plugin/v2/sessions.js.map +1 -0
- package/dist/opencode-plugin/v2/sse.d.ts +101 -0
- package/dist/opencode-plugin/v2/sse.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/sse.js +235 -0
- package/dist/opencode-plugin/v2/sse.js.map +1 -0
- package/dist/opencode-plugin/v2/types.d.ts +188 -0
- package/dist/opencode-plugin/v2/types.d.ts.map +1 -0
- package/dist/opencode-plugin/v2/types.js +10 -0
- package/dist/opencode-plugin/v2/types.js.map +1 -0
- package/dist/opencode-plugin/vcs-command.d.ts +3 -0
- package/dist/opencode-plugin/vcs-command.d.ts.map +1 -0
- package/dist/opencode-plugin/vcs-command.js +34 -0
- package/dist/opencode-plugin/vcs-command.js.map +1 -0
- package/dist/pipeline/index.d.ts +12 -0
- package/dist/pipeline/index.d.ts.map +1 -0
- package/dist/pipeline/index.js +12 -0
- package/dist/pipeline/index.js.map +1 -0
- package/dist/pipeline/prenormalize.d.ts +48 -0
- package/dist/pipeline/prenormalize.d.ts.map +1 -0
- package/dist/pipeline/prenormalize.js +94 -0
- package/dist/pipeline/prenormalize.js.map +1 -0
- package/dist/pipeline/resolver.d.ts +56 -0
- package/dist/pipeline/resolver.d.ts.map +1 -0
- package/dist/pipeline/resolver.js +239 -0
- package/dist/pipeline/resolver.js.map +1 -0
- package/dist/pipeline/semantic-data-loader.d.ts +171 -0
- package/dist/pipeline/semantic-data-loader.d.ts.map +1 -0
- package/dist/pipeline/semantic-data-loader.js +677 -0
- package/dist/pipeline/semantic-data-loader.js.map +1 -0
- package/dist/pipeline/semantic-enricher.d.ts +112 -0
- package/dist/pipeline/semantic-enricher.d.ts.map +1 -0
- package/dist/pipeline/semantic-enricher.js +322 -0
- package/dist/pipeline/semantic-enricher.js.map +1 -0
- package/dist/pipeline/tagger.d.ts +121 -0
- package/dist/pipeline/tagger.d.ts.map +1 -0
- package/dist/pipeline/tagger.js +545 -0
- package/dist/pipeline/tagger.js.map +1 -0
- package/dist/pipeline/title-extractor.d.ts +79 -0
- package/dist/pipeline/title-extractor.d.ts.map +1 -0
- package/dist/pipeline/title-extractor.js +801 -0
- package/dist/pipeline/title-extractor.js.map +1 -0
- package/dist/pipeline/validator.d.ts +65 -0
- package/dist/pipeline/validator.d.ts.map +1 -0
- package/dist/pipeline/validator.js +292 -0
- package/dist/pipeline/validator.js.map +1 -0
- package/dist/proxy/index.d.ts +13 -0
- package/dist/proxy/index.d.ts.map +1 -0
- package/dist/proxy/index.js +12 -0
- package/dist/proxy/index.js.map +1 -0
- package/dist/proxy/providers/anthropic.d.ts +28 -0
- package/dist/proxy/providers/anthropic.d.ts.map +1 -0
- package/dist/proxy/providers/anthropic.js +265 -0
- package/dist/proxy/providers/anthropic.js.map +1 -0
- package/dist/proxy/providers/index.d.ts +19 -0
- package/dist/proxy/providers/index.d.ts.map +1 -0
- package/dist/proxy/providers/index.js +41 -0
- package/dist/proxy/providers/index.js.map +1 -0
- package/dist/proxy/providers/openai.d.ts +27 -0
- package/dist/proxy/providers/openai.d.ts.map +1 -0
- package/dist/proxy/providers/openai.js +203 -0
- package/dist/proxy/providers/openai.js.map +1 -0
- package/dist/proxy/providers/streaming-tool-loop.d.ts +12 -0
- package/dist/proxy/providers/streaming-tool-loop.d.ts.map +1 -0
- package/dist/proxy/providers/streaming-tool-loop.js +223 -0
- package/dist/proxy/providers/streaming-tool-loop.js.map +1 -0
- package/dist/proxy/providers/types.d.ts +109 -0
- package/dist/proxy/providers/types.d.ts.map +1 -0
- package/dist/proxy/providers/types.js +2 -0
- package/dist/proxy/providers/types.js.map +1 -0
- package/dist/proxy/proxy-server.d.ts +61 -0
- package/dist/proxy/proxy-server.d.ts.map +1 -0
- package/dist/proxy/proxy-server.js +146 -0
- package/dist/proxy/proxy-server.js.map +1 -0
- package/dist/proxy/rehydra-fetch.d.ts +35 -0
- package/dist/proxy/rehydra-fetch.d.ts.map +1 -0
- package/dist/proxy/rehydra-fetch.js +631 -0
- package/dist/proxy/rehydra-fetch.js.map +1 -0
- package/dist/proxy/rehydra-proxy.d.ts +40 -0
- package/dist/proxy/rehydra-proxy.d.ts.map +1 -0
- package/dist/proxy/rehydra-proxy.js +150 -0
- package/dist/proxy/rehydra-proxy.js.map +1 -0
- package/dist/proxy/sse-parser.d.ts +59 -0
- package/dist/proxy/sse-parser.d.ts.map +1 -0
- package/dist/proxy/sse-parser.js +112 -0
- package/dist/proxy/sse-parser.js.map +1 -0
- package/dist/proxy/system-instruction.d.ts +19 -0
- package/dist/proxy/system-instruction.d.ts.map +1 -0
- package/dist/proxy/system-instruction.js +33 -0
- package/dist/proxy/system-instruction.js.map +1 -0
- package/dist/proxy/tool-loop-stream.d.ts +7 -0
- package/dist/proxy/tool-loop-stream.d.ts.map +1 -0
- package/dist/proxy/tool-loop-stream.js +65 -0
- package/dist/proxy/tool-loop-stream.js.map +1 -0
- package/dist/proxy/types.d.ts +138 -0
- package/dist/proxy/types.d.ts.map +1 -0
- package/dist/proxy/types.js +5 -0
- package/dist/proxy/types.js.map +1 -0
- package/dist/proxy/wrap-client.d.ts +47 -0
- package/dist/proxy/wrap-client.d.ts.map +1 -0
- package/dist/proxy/wrap-client.js +70 -0
- package/dist/proxy/wrap-client.js.map +1 -0
- package/dist/recognizers/address.d.ts +2 -0
- package/dist/recognizers/address.d.ts.map +1 -0
- package/dist/recognizers/address.js +23 -0
- package/dist/recognizers/address.js.map +1 -0
- package/dist/recognizers/base.d.ts +78 -0
- package/dist/recognizers/base.d.ts.map +1 -0
- package/dist/recognizers/base.js +100 -0
- package/dist/recognizers/base.js.map +1 -0
- package/dist/recognizers/bic-swift.d.ts +10 -0
- package/dist/recognizers/bic-swift.d.ts.map +1 -0
- package/dist/recognizers/bic-swift.js +107 -0
- package/dist/recognizers/bic-swift.js.map +1 -0
- package/dist/recognizers/credit-card.d.ts +32 -0
- package/dist/recognizers/credit-card.d.ts.map +1 -0
- package/dist/recognizers/credit-card.js +160 -0
- package/dist/recognizers/credit-card.js.map +1 -0
- package/dist/recognizers/custom-id.d.ts +28 -0
- package/dist/recognizers/custom-id.d.ts.map +1 -0
- package/dist/recognizers/custom-id.js +116 -0
- package/dist/recognizers/custom-id.js.map +1 -0
- package/dist/recognizers/date.d.ts +10 -0
- package/dist/recognizers/date.d.ts.map +1 -0
- package/dist/recognizers/date.js +228 -0
- package/dist/recognizers/date.js.map +1 -0
- package/dist/recognizers/email.d.ts +10 -0
- package/dist/recognizers/email.d.ts.map +1 -0
- package/dist/recognizers/email.js +75 -0
- package/dist/recognizers/email.js.map +1 -0
- package/dist/recognizers/iban.d.ts +14 -0
- package/dist/recognizers/iban.d.ts.map +1 -0
- package/dist/recognizers/iban.js +67 -0
- package/dist/recognizers/iban.js.map +1 -0
- package/dist/recognizers/index.d.ts +25 -0
- package/dist/recognizers/index.d.ts.map +1 -0
- package/dist/recognizers/index.js +53 -0
- package/dist/recognizers/index.js.map +1 -0
- package/dist/recognizers/ip-address.d.ts +14 -0
- package/dist/recognizers/ip-address.d.ts.map +1 -0
- package/dist/recognizers/ip-address.js +183 -0
- package/dist/recognizers/ip-address.js.map +1 -0
- package/dist/recognizers/phone.d.ts +10 -0
- package/dist/recognizers/phone.d.ts.map +1 -0
- package/dist/recognizers/phone.js +145 -0
- package/dist/recognizers/phone.js.map +1 -0
- package/dist/recognizers/postal-code.d.ts +16 -0
- package/dist/recognizers/postal-code.d.ts.map +1 -0
- package/dist/recognizers/postal-code.js +42 -0
- package/dist/recognizers/postal-code.js.map +1 -0
- package/dist/recognizers/registry.d.ts +59 -0
- package/dist/recognizers/registry.d.ts.map +1 -0
- package/dist/recognizers/registry.js +113 -0
- package/dist/recognizers/registry.js.map +1 -0
- package/dist/recognizers/secrets/api-key.d.ts +6 -0
- package/dist/recognizers/secrets/api-key.d.ts.map +1 -0
- package/dist/recognizers/secrets/api-key.js +40 -0
- package/dist/recognizers/secrets/api-key.js.map +1 -0
- package/dist/recognizers/secrets/aws-credentials.d.ts +7 -0
- package/dist/recognizers/secrets/aws-credentials.d.ts.map +1 -0
- package/dist/recognizers/secrets/aws-credentials.js +68 -0
- package/dist/recognizers/secrets/aws-credentials.js.map +1 -0
- package/dist/recognizers/secrets/config-secret.d.ts +8 -0
- package/dist/recognizers/secrets/config-secret.d.ts.map +1 -0
- package/dist/recognizers/secrets/config-secret.js +59 -0
- package/dist/recognizers/secrets/config-secret.js.map +1 -0
- package/dist/recognizers/secrets/connection-string.d.ts +6 -0
- package/dist/recognizers/secrets/connection-string.d.ts.map +1 -0
- package/dist/recognizers/secrets/connection-string.js +40 -0
- package/dist/recognizers/secrets/connection-string.js.map +1 -0
- package/dist/recognizers/secrets/env-var.d.ts +8 -0
- package/dist/recognizers/secrets/env-var.d.ts.map +1 -0
- package/dist/recognizers/secrets/env-var.js +54 -0
- package/dist/recognizers/secrets/env-var.js.map +1 -0
- package/dist/recognizers/secrets/index.d.ts +22 -0
- package/dist/recognizers/secrets/index.d.ts.map +1 -0
- package/dist/recognizers/secrets/index.js +41 -0
- package/dist/recognizers/secrets/index.js.map +1 -0
- package/dist/recognizers/secrets/jwt.d.ts +6 -0
- package/dist/recognizers/secrets/jwt.d.ts.map +1 -0
- package/dist/recognizers/secrets/jwt.js +47 -0
- package/dist/recognizers/secrets/jwt.js.map +1 -0
- package/dist/recognizers/secrets/key-patterns.d.ts +10 -0
- package/dist/recognizers/secrets/key-patterns.d.ts.map +1 -0
- package/dist/recognizers/secrets/key-patterns.js +15 -0
- package/dist/recognizers/secrets/key-patterns.js.map +1 -0
- package/dist/recognizers/secrets/literal-value.d.ts +10 -0
- package/dist/recognizers/secrets/literal-value.d.ts.map +1 -0
- package/dist/recognizers/secrets/literal-value.js +48 -0
- package/dist/recognizers/secrets/literal-value.js.map +1 -0
- package/dist/recognizers/secrets/private-key.d.ts +6 -0
- package/dist/recognizers/secrets/private-key.d.ts.map +1 -0
- package/dist/recognizers/secrets/private-key.js +16 -0
- package/dist/recognizers/secrets/private-key.js.map +1 -0
- package/dist/recognizers/url.d.ts +14 -0
- package/dist/recognizers/url.d.ts.map +1 -0
- package/dist/recognizers/url.js +121 -0
- package/dist/recognizers/url.js.map +1 -0
- package/dist/storage/browser.d.ts +13 -0
- package/dist/storage/browser.d.ts.map +1 -0
- package/dist/storage/browser.js +13 -0
- package/dist/storage/browser.js.map +1 -0
- package/dist/storage/in-memory.d.ts +60 -0
- package/dist/storage/in-memory.d.ts.map +1 -0
- package/dist/storage/in-memory.js +108 -0
- package/dist/storage/in-memory.js.map +1 -0
- package/dist/storage/index.d.ts +11 -0
- package/dist/storage/index.d.ts.map +1 -0
- package/dist/storage/index.js +11 -0
- package/dist/storage/index.js.map +1 -0
- package/dist/storage/indexeddb.d.ts +69 -0
- package/dist/storage/indexeddb.d.ts.map +1 -0
- package/dist/storage/indexeddb.js +274 -0
- package/dist/storage/indexeddb.js.map +1 -0
- package/dist/storage/session-base.d.ts +34 -0
- package/dist/storage/session-base.d.ts.map +1 -0
- package/dist/storage/session-base.js +122 -0
- package/dist/storage/session-base.js.map +1 -0
- package/dist/storage/session.d.ts +19 -0
- package/dist/storage/session.d.ts.map +1 -0
- package/dist/storage/session.js +31 -0
- package/dist/storage/session.js.map +1 -0
- package/dist/storage/sqlite.d.ts +80 -0
- package/dist/storage/sqlite.d.ts.map +1 -0
- package/dist/storage/sqlite.js +255 -0
- package/dist/storage/sqlite.js.map +1 -0
- package/dist/storage/types.d.ts +208 -0
- package/dist/storage/types.d.ts.map +1 -0
- package/dist/storage/types.js +6 -0
- package/dist/storage/types.js.map +1 -0
- package/dist/streaming/anonymizer-stream.d.ts +63 -0
- package/dist/streaming/anonymizer-stream.d.ts.map +1 -0
- package/dist/streaming/anonymizer-stream.js +184 -0
- package/dist/streaming/anonymizer-stream.js.map +1 -0
- package/dist/streaming/index.d.ts +9 -0
- package/dist/streaming/index.d.ts.map +1 -0
- package/dist/streaming/index.js +8 -0
- package/dist/streaming/index.js.map +1 -0
- package/dist/streaming/sentence-buffer.d.ts +78 -0
- package/dist/streaming/sentence-buffer.d.ts.map +1 -0
- package/dist/streaming/sentence-buffer.js +239 -0
- package/dist/streaming/sentence-buffer.js.map +1 -0
- package/dist/streaming/stream-factory.d.ts +38 -0
- package/dist/streaming/stream-factory.d.ts.map +1 -0
- package/dist/streaming/stream-factory.js +69 -0
- package/dist/streaming/stream-factory.js.map +1 -0
- package/dist/streaming/types.d.ts +121 -0
- package/dist/streaming/types.d.ts.map +1 -0
- package/dist/streaming/types.js +5 -0
- package/dist/streaming/types.js.map +1 -0
- package/dist/types/index.d.ts +269 -0
- package/dist/types/index.d.ts.map +1 -0
- package/dist/types/index.js +95 -0
- package/dist/types/index.js.map +1 -0
- package/dist/types/pii-types.d.ts +64 -0
- package/dist/types/pii-types.d.ts.map +1 -0
- package/dist/types/pii-types.js +153 -0
- package/dist/types/pii-types.js.map +1 -0
- package/dist/utils/iban-checksum.d.ts +23 -0
- package/dist/utils/iban-checksum.d.ts.map +1 -0
- package/dist/utils/iban-checksum.js +106 -0
- package/dist/utils/iban-checksum.js.map +1 -0
- package/dist/utils/index.d.ts +11 -0
- package/dist/utils/index.d.ts.map +1 -0
- package/dist/utils/index.js +11 -0
- package/dist/utils/index.js.map +1 -0
- package/dist/utils/json-walk.d.ts +21 -0
- package/dist/utils/json-walk.d.ts.map +1 -0
- package/dist/utils/json-walk.js +44 -0
- package/dist/utils/json-walk.js.map +1 -0
- package/dist/utils/luhn.d.ts +17 -0
- package/dist/utils/luhn.d.ts.map +1 -0
- package/dist/utils/luhn.js +55 -0
- package/dist/utils/luhn.js.map +1 -0
- package/dist/utils/offsets.d.ts +86 -0
- package/dist/utils/offsets.d.ts.map +1 -0
- package/dist/utils/offsets.js +124 -0
- package/dist/utils/offsets.js.map +1 -0
- package/dist/utils/path.d.ts +34 -0
- package/dist/utils/path.d.ts.map +1 -0
- package/dist/utils/path.js +96 -0
- package/dist/utils/path.js.map +1 -0
- package/dist/utils/regex.d.ts +11 -0
- package/dist/utils/regex.d.ts.map +1 -0
- package/dist/utils/regex.js +14 -0
- package/dist/utils/regex.js.map +1 -0
- package/dist/utils/storage-browser.d.ts +51 -0
- package/dist/utils/storage-browser.d.ts.map +1 -0
- package/dist/utils/storage-browser.js +381 -0
- package/dist/utils/storage-browser.js.map +1 -0
- package/dist/utils/storage-node.d.ts +51 -0
- package/dist/utils/storage-node.d.ts.map +1 -0
- package/dist/utils/storage-node.js +153 -0
- package/dist/utils/storage-node.js.map +1 -0
- package/dist/utils/storage.browser.d.ts +75 -0
- package/dist/utils/storage.browser.d.ts.map +1 -0
- package/dist/utils/storage.browser.js +69 -0
- package/dist/utils/storage.browser.js.map +1 -0
- package/dist/utils/storage.d.ts +73 -0
- package/dist/utils/storage.d.ts.map +1 -0
- package/dist/utils/storage.js +69 -0
- package/dist/utils/storage.js.map +1 -0
- package/package.json +148 -0
|
@@ -0,0 +1,362 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* HuggingFace Tokenizer
|
|
3
|
+
* Loads and uses tokenizers from HuggingFace's tokenizer.json format
|
|
4
|
+
* Supports Unigram (SentencePiece) and BPE tokenizers
|
|
5
|
+
*/
|
|
6
|
+
/**
|
|
7
|
+
* Default tokenizer configuration
|
|
8
|
+
*/
|
|
9
|
+
export const DEFAULT_TOKENIZER_CONFIG = {
|
|
10
|
+
maxLength: 512,
|
|
11
|
+
doLowerCase: false, // XLM-RoBERTa doesn't lowercase
|
|
12
|
+
};
|
|
13
|
+
/**
|
|
14
|
+
* Creates a new trie node
|
|
15
|
+
*/
|
|
16
|
+
function createTrieNode() {
|
|
17
|
+
return { children: new Map(), tokenId: null, token: null };
|
|
18
|
+
}
|
|
19
|
+
/**
|
|
20
|
+
* WordPiece Tokenizer - supports both HuggingFace JSON and vocab.txt formats
|
|
21
|
+
* Uses a Trie for O(token_length) lookups instead of O(vocab_size)
|
|
22
|
+
*/
|
|
23
|
+
export class WordPieceTokenizer {
|
|
24
|
+
vocab;
|
|
25
|
+
inverseVocab;
|
|
26
|
+
config;
|
|
27
|
+
trie;
|
|
28
|
+
// Special token IDs (XLM-RoBERTa style)
|
|
29
|
+
clsId = 0; // <s>
|
|
30
|
+
sepId = 2; // </s>
|
|
31
|
+
padId = 1; // <pad>
|
|
32
|
+
unkId = 3; // <unk>
|
|
33
|
+
// Special token strings
|
|
34
|
+
clsToken = '<s>';
|
|
35
|
+
sepToken = '</s>';
|
|
36
|
+
padToken = '<pad>';
|
|
37
|
+
unkToken = '<unk>';
|
|
38
|
+
constructor(vocab, config = {}) {
|
|
39
|
+
this.vocab = vocab;
|
|
40
|
+
this.config = { ...DEFAULT_TOKENIZER_CONFIG, ...config };
|
|
41
|
+
// Build inverse vocab
|
|
42
|
+
this.inverseVocab = new Map();
|
|
43
|
+
for (const [token, id] of vocab) {
|
|
44
|
+
this.inverseVocab.set(id, token);
|
|
45
|
+
}
|
|
46
|
+
// Build trie for fast prefix matching
|
|
47
|
+
this.trie = createTrieNode();
|
|
48
|
+
for (const [token, id] of vocab) {
|
|
49
|
+
// Skip special tokens in trie (we handle them explicitly)
|
|
50
|
+
if (token.startsWith('<') || token.startsWith('['))
|
|
51
|
+
continue;
|
|
52
|
+
this.insertIntoTrie(token, id);
|
|
53
|
+
}
|
|
54
|
+
// Try to detect special tokens from vocab
|
|
55
|
+
this.detectSpecialTokens();
|
|
56
|
+
}
|
|
57
|
+
/**
|
|
58
|
+
* Insert a token into the trie
|
|
59
|
+
*/
|
|
60
|
+
insertIntoTrie(token, id) {
|
|
61
|
+
let node = this.trie;
|
|
62
|
+
for (const char of token) {
|
|
63
|
+
if (!node.children.has(char)) {
|
|
64
|
+
node.children.set(char, createTrieNode());
|
|
65
|
+
}
|
|
66
|
+
node = node.children.get(char);
|
|
67
|
+
}
|
|
68
|
+
node.tokenId = id;
|
|
69
|
+
node.token = token;
|
|
70
|
+
}
|
|
71
|
+
/**
|
|
72
|
+
* Detect special tokens from vocabulary
|
|
73
|
+
*/
|
|
74
|
+
detectSpecialTokens() {
|
|
75
|
+
// XLM-RoBERTa style
|
|
76
|
+
if (this.vocab.has('<s>')) {
|
|
77
|
+
this.clsToken = '<s>';
|
|
78
|
+
this.clsId = this.vocab.get('<s>') ?? 0;
|
|
79
|
+
this.sepToken = '</s>';
|
|
80
|
+
this.sepId = this.vocab.get('</s>') ?? 2;
|
|
81
|
+
this.padToken = '<pad>';
|
|
82
|
+
this.padId = this.vocab.get('<pad>') ?? 1;
|
|
83
|
+
this.unkToken = '<unk>';
|
|
84
|
+
this.unkId = this.vocab.get('<unk>') ?? 3;
|
|
85
|
+
}
|
|
86
|
+
// BERT style
|
|
87
|
+
else if (this.vocab.has('[CLS]')) {
|
|
88
|
+
this.clsToken = '[CLS]';
|
|
89
|
+
this.clsId = this.vocab.get('[CLS]') ?? 101;
|
|
90
|
+
this.sepToken = '[SEP]';
|
|
91
|
+
this.sepId = this.vocab.get('[SEP]') ?? 102;
|
|
92
|
+
this.padToken = '[PAD]';
|
|
93
|
+
this.padId = this.vocab.get('[PAD]') ?? 0;
|
|
94
|
+
this.unkToken = '[UNK]';
|
|
95
|
+
this.unkId = this.vocab.get('[UNK]') ?? 100;
|
|
96
|
+
}
|
|
97
|
+
}
|
|
98
|
+
/**
|
|
99
|
+
* Tokenizes text into tokens with offset tracking
|
|
100
|
+
*/
|
|
101
|
+
tokenize(text, options = {}) {
|
|
102
|
+
const tokens = [];
|
|
103
|
+
const tokenToCharSpan = [];
|
|
104
|
+
// Add CLS token
|
|
105
|
+
tokens.push({
|
|
106
|
+
id: this.clsId,
|
|
107
|
+
token: this.clsToken,
|
|
108
|
+
start: 0,
|
|
109
|
+
end: 0,
|
|
110
|
+
isContinuation: false,
|
|
111
|
+
isSpecial: true,
|
|
112
|
+
});
|
|
113
|
+
tokenToCharSpan.push(null);
|
|
114
|
+
// Preprocess text
|
|
115
|
+
const processedText = this.config.doLowerCase ? text.toLowerCase() : text;
|
|
116
|
+
// Token budget, counting the CLS already pushed and the SEP still to come.
|
|
117
|
+
// maxTokens wins over truncate; otherwise truncate caps at maxLength.
|
|
118
|
+
const maxTokens = options.maxTokens ?? ((options.truncate ?? true) ? this.config.maxLength : undefined);
|
|
119
|
+
let truncated = false;
|
|
120
|
+
// Tokenize using greedy longest-match
|
|
121
|
+
let pos = 0;
|
|
122
|
+
while (pos < processedText.length) {
|
|
123
|
+
// Skip whitespace
|
|
124
|
+
if (/\s/.test(processedText[pos])) {
|
|
125
|
+
pos++;
|
|
126
|
+
continue;
|
|
127
|
+
}
|
|
128
|
+
// Stop before exceeding the budget (reserve one slot for the SEP token).
|
|
129
|
+
// Breaking here — rather than tokenizing everything and slicing — keeps
|
|
130
|
+
// memory bounded on huge inputs.
|
|
131
|
+
if (maxTokens !== undefined && tokens.length >= maxTokens - 1) {
|
|
132
|
+
truncated = true;
|
|
133
|
+
break;
|
|
134
|
+
}
|
|
135
|
+
// Find the longest matching token starting at this position
|
|
136
|
+
const { token, id, length } = this.findBestToken(processedText, pos);
|
|
137
|
+
const isFirstOfWord = pos === 0 || /\s/.test(processedText[pos - 1]);
|
|
138
|
+
tokens.push({
|
|
139
|
+
id,
|
|
140
|
+
token,
|
|
141
|
+
start: pos,
|
|
142
|
+
end: pos + length,
|
|
143
|
+
isContinuation: !isFirstOfWord && !token.startsWith('▁'),
|
|
144
|
+
isSpecial: false,
|
|
145
|
+
});
|
|
146
|
+
tokenToCharSpan.push([pos, pos + length]);
|
|
147
|
+
pos += length;
|
|
148
|
+
}
|
|
149
|
+
// Add SEP token
|
|
150
|
+
tokens.push({
|
|
151
|
+
id: this.sepId,
|
|
152
|
+
token: this.sepToken,
|
|
153
|
+
start: text.length,
|
|
154
|
+
end: text.length,
|
|
155
|
+
isContinuation: false,
|
|
156
|
+
isSpecial: true,
|
|
157
|
+
});
|
|
158
|
+
tokenToCharSpan.push(null);
|
|
159
|
+
// Build arrays
|
|
160
|
+
const inputIds = tokens.map((t) => t.id);
|
|
161
|
+
const attentionMask = tokens.map(() => 1);
|
|
162
|
+
const tokenTypeIds = tokens.map(() => 0);
|
|
163
|
+
return {
|
|
164
|
+
tokens,
|
|
165
|
+
inputIds,
|
|
166
|
+
attentionMask,
|
|
167
|
+
tokenTypeIds,
|
|
168
|
+
tokenToCharSpan,
|
|
169
|
+
truncated,
|
|
170
|
+
};
|
|
171
|
+
}
|
|
172
|
+
/**
|
|
173
|
+
* Find the best matching token using trie-based greedy longest-match
|
|
174
|
+
* O(max_token_length) instead of O(vocab_size)
|
|
175
|
+
*/
|
|
176
|
+
findBestToken(text, startPos) {
|
|
177
|
+
// Check if this starts a new word (preceded by space or start)
|
|
178
|
+
const isWordStart = startPos === 0 || /\s/.test(text[startPos - 1]);
|
|
179
|
+
// For SentencePiece models, word-initial tokens start with ▁
|
|
180
|
+
// Try matching with ▁ prefix first for word starts
|
|
181
|
+
if (isWordStart) {
|
|
182
|
+
const result = this.findLongestTrieMatch('▁', text, startPos);
|
|
183
|
+
if (result) {
|
|
184
|
+
return result;
|
|
185
|
+
}
|
|
186
|
+
}
|
|
187
|
+
// Try matching without prefix
|
|
188
|
+
const result = this.findLongestTrieMatch('', text, startPos);
|
|
189
|
+
if (result) {
|
|
190
|
+
return result;
|
|
191
|
+
}
|
|
192
|
+
// Single character fallback
|
|
193
|
+
const char = text[startPos];
|
|
194
|
+
const charId = this.vocab.get(char) ?? this.vocab.get('▁' + char) ?? this.unkId;
|
|
195
|
+
return { token: char, id: charId, length: 1 };
|
|
196
|
+
}
|
|
197
|
+
/**
|
|
198
|
+
* Find the longest matching token in the trie
|
|
199
|
+
*/
|
|
200
|
+
findLongestTrieMatch(prefix, text, startPos) {
|
|
201
|
+
let node = this.trie;
|
|
202
|
+
let bestMatch = null;
|
|
203
|
+
// First, traverse the prefix (e.g., '▁')
|
|
204
|
+
for (const char of prefix) {
|
|
205
|
+
const child = node.children.get(char);
|
|
206
|
+
if (!child)
|
|
207
|
+
return null;
|
|
208
|
+
node = child;
|
|
209
|
+
}
|
|
210
|
+
// Now traverse the actual text
|
|
211
|
+
let matchLength = 0;
|
|
212
|
+
for (let i = startPos; i < text.length; i++) {
|
|
213
|
+
const char = text[i];
|
|
214
|
+
const child = node.children.get(char);
|
|
215
|
+
if (!child)
|
|
216
|
+
break;
|
|
217
|
+
node = child;
|
|
218
|
+
matchLength++;
|
|
219
|
+
// If this node represents a complete token, record it as best match so far
|
|
220
|
+
if (node.tokenId !== null && node.token !== null) {
|
|
221
|
+
bestMatch = {
|
|
222
|
+
token: node.token,
|
|
223
|
+
id: node.tokenId,
|
|
224
|
+
length: matchLength, // Length in original text (without prefix)
|
|
225
|
+
};
|
|
226
|
+
}
|
|
227
|
+
}
|
|
228
|
+
return bestMatch;
|
|
229
|
+
}
|
|
230
|
+
/**
|
|
231
|
+
* Decodes token IDs back to text
|
|
232
|
+
*/
|
|
233
|
+
decode(tokenIds) {
|
|
234
|
+
const parts = [];
|
|
235
|
+
for (const id of tokenIds) {
|
|
236
|
+
const token = this.inverseVocab.get(id);
|
|
237
|
+
if (token === undefined)
|
|
238
|
+
continue;
|
|
239
|
+
if (token === this.clsToken || token === this.sepToken || token === this.padToken)
|
|
240
|
+
continue;
|
|
241
|
+
// SentencePiece uses ▁ to mark word boundaries
|
|
242
|
+
if (token.startsWith('▁')) {
|
|
243
|
+
parts.push(' ' + token.slice(1));
|
|
244
|
+
}
|
|
245
|
+
else {
|
|
246
|
+
parts.push(token);
|
|
247
|
+
}
|
|
248
|
+
}
|
|
249
|
+
return parts.join('').trim();
|
|
250
|
+
}
|
|
251
|
+
/**
|
|
252
|
+
* Gets vocabulary size
|
|
253
|
+
*/
|
|
254
|
+
get vocabSize() {
|
|
255
|
+
return this.vocab.size;
|
|
256
|
+
}
|
|
257
|
+
/**
|
|
258
|
+
* Gets a token ID by string
|
|
259
|
+
*/
|
|
260
|
+
getTokenId(token) {
|
|
261
|
+
return this.vocab.get(token);
|
|
262
|
+
}
|
|
263
|
+
/**
|
|
264
|
+
* Gets a token string by ID
|
|
265
|
+
*/
|
|
266
|
+
getToken(id) {
|
|
267
|
+
return this.inverseVocab.get(id);
|
|
268
|
+
}
|
|
269
|
+
}
|
|
270
|
+
/**
|
|
271
|
+
* Loads vocabulary from a file (supports tokenizer.json and vocab.txt)
|
|
272
|
+
* Uses storage abstraction for browser compatibility
|
|
273
|
+
*/
|
|
274
|
+
export async function loadVocabFromFile(filePath) {
|
|
275
|
+
const { getStorageProvider } = await import('#storage');
|
|
276
|
+
const storage = await getStorageProvider();
|
|
277
|
+
const content = await storage.readTextFile(filePath);
|
|
278
|
+
// Detect format
|
|
279
|
+
if (filePath.endsWith('.json') || content.trim().startsWith('{')) {
|
|
280
|
+
return parseHFTokenizerJson(content);
|
|
281
|
+
}
|
|
282
|
+
else {
|
|
283
|
+
return parseVocab(content);
|
|
284
|
+
}
|
|
285
|
+
}
|
|
286
|
+
/**
|
|
287
|
+
* Parses HuggingFace tokenizer.json format
|
|
288
|
+
*/
|
|
289
|
+
export function parseHFTokenizerJson(content) {
|
|
290
|
+
const vocab = new Map();
|
|
291
|
+
try {
|
|
292
|
+
const config = JSON.parse(content);
|
|
293
|
+
// Add special tokens first
|
|
294
|
+
if (Array.isArray(config.added_tokens)) {
|
|
295
|
+
for (const token of config.added_tokens) {
|
|
296
|
+
vocab.set(token.content, token.id);
|
|
297
|
+
}
|
|
298
|
+
}
|
|
299
|
+
// Add model vocabulary
|
|
300
|
+
if (config.model !== undefined && config.model.vocab !== undefined) {
|
|
301
|
+
if (Array.isArray(config.model.vocab)) {
|
|
302
|
+
// Unigram format: array of [token, score] pairs
|
|
303
|
+
for (let i = 0; i < config.model.vocab.length; i++) {
|
|
304
|
+
const entry = config.model.vocab[i];
|
|
305
|
+
if (entry && typeof entry[0] === 'string') {
|
|
306
|
+
vocab.set(entry[0], i);
|
|
307
|
+
}
|
|
308
|
+
}
|
|
309
|
+
}
|
|
310
|
+
else {
|
|
311
|
+
// BPE/WordPiece format: object mapping token -> id
|
|
312
|
+
for (const [token, id] of Object.entries(config.model.vocab)) {
|
|
313
|
+
vocab.set(token, id);
|
|
314
|
+
}
|
|
315
|
+
}
|
|
316
|
+
}
|
|
317
|
+
}
|
|
318
|
+
catch (e) {
|
|
319
|
+
throw new Error(`Failed to parse tokenizer.json: ${String(e)}`);
|
|
320
|
+
}
|
|
321
|
+
return vocab;
|
|
322
|
+
}
|
|
323
|
+
/**
|
|
324
|
+
* Parses vocabulary from string content (vocab.txt format)
|
|
325
|
+
*/
|
|
326
|
+
export function parseVocab(content) {
|
|
327
|
+
const vocab = new Map();
|
|
328
|
+
const lines = content.split('\n');
|
|
329
|
+
for (let i = 0; i < lines.length; i++) {
|
|
330
|
+
const token = lines[i]?.trim();
|
|
331
|
+
if (token !== undefined && token.length > 0) {
|
|
332
|
+
vocab.set(token, i);
|
|
333
|
+
}
|
|
334
|
+
}
|
|
335
|
+
return vocab;
|
|
336
|
+
}
|
|
337
|
+
/**
|
|
338
|
+
* Creates a minimal vocabulary for testing
|
|
339
|
+
*/
|
|
340
|
+
export function createTestVocab() {
|
|
341
|
+
const tokens = [
|
|
342
|
+
'<s>',
|
|
343
|
+
'<pad>',
|
|
344
|
+
'</s>',
|
|
345
|
+
'<unk>',
|
|
346
|
+
'▁Hello',
|
|
347
|
+
'▁John',
|
|
348
|
+
'▁Smith',
|
|
349
|
+
'▁from',
|
|
350
|
+
'▁Acme',
|
|
351
|
+
'▁Corp',
|
|
352
|
+
'▁in',
|
|
353
|
+
'▁Berlin',
|
|
354
|
+
'!',
|
|
355
|
+
];
|
|
356
|
+
const vocab = new Map();
|
|
357
|
+
tokens.forEach((token, index) => {
|
|
358
|
+
vocab.set(token, index);
|
|
359
|
+
});
|
|
360
|
+
return vocab;
|
|
361
|
+
}
|
|
362
|
+
//# sourceMappingURL=tokenizer.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"tokenizer.js","sourceRoot":"","sources":["../../src/ner/tokenizer.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAqFH;;GAEG;AACH,MAAM,CAAC,MAAM,wBAAwB,GAAoB;IACvD,SAAS,EAAE,GAAG;IACd,WAAW,EAAE,KAAK,EAAE,gCAAgC;CACrD,CAAC;AAWF;;GAEG;AACH,SAAS,cAAc;IACrB,OAAO,EAAE,QAAQ,EAAE,IAAI,GAAG,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;AAC7D,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,kBAAkB;IACrB,KAAK,CAAsB;IAC3B,YAAY,CAAsB;IAClC,MAAM,CAAkB;IACxB,IAAI,CAAW;IAEvB,wCAAwC;IAChC,KAAK,GAAW,CAAC,CAAC,CAAE,MAAM;IAC1B,KAAK,GAAW,CAAC,CAAC,CAAE,OAAO;IAC3B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAC5B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAEpC,wBAAwB;IAChB,QAAQ,GAAW,KAAK,CAAC;IACzB,QAAQ,GAAW,MAAM,CAAC;IAC1B,QAAQ,GAAW,OAAO,CAAC;IAC3B,QAAQ,GAAW,OAAO,CAAC;IAEnC,YAAY,KAA0B,EAAE,SAAmC,EAAE;QAC3E,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,MAAM,GAAG,EAAE,GAAG,wBAAwB,EAAE,GAAG,MAAM,EAAE,CAAC;QAEzD,sBAAsB;QACtB,IAAI,CAAC,YAAY,GAAG,IAAI,GAAG,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC;YAChC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;QACnC,CAAC;QAED,sCAAsC;QACtC,IAAI,CAAC,IAAI,GAAG,cAAc,EAAE,CAAC;QAC7B,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC;YAChC,0DAA0D;YAC1D,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC7D,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACjC,CAAC;QAED,0CAA0C;QAC1C,IAAI,CAAC,mBAAmB,EAAE,CAAC;IAC7B,CAAC;IAED;;OAEG;IACK,cAAc,CAAC,KAAa,EAAE,EAAU;QAC9C,IAAI,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QACrB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC7B,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,EAAE,cAAc,EAAE,CAAC,CAAC;YAC5C,CAAC;YACD,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAE,CAAC;QAClC,CAAC;QACD,IAAI,CAAC,OAAO,GAAG,EAAE,CAAC;QAClB,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;IACrB,CAAC;IAED;;OAEG;IACK,mBAAmB;QACzB,oBAAoB;QACpB,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAC1B,IAAI,CAAC,QAAQ,GAAG,KAAK,CAAC;YACtB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,CAAC,QAAQ,GAAG,MAAM,CAAC;YACvB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YACzC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QAC5C,CAAC;QACD,aAAa;aACR,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YACjC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;QAC9C,CAAC;IACH,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,IAAY,EAAE,UAA2B,EAAE;QAClD,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,eAAe,GAAmC,EAAE,CAAC;QAE3D,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,CAAC;YACR,GAAG,EAAE,CAAC;YACN,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,kBAAkB;QAClB,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QAE1E,2EAA2E;QAC3E,sEAAsE;QACtE,MAAM,SAAS,GACb,OAAO,CAAC,SAAS,IAAI,CAAC,CAAC,OAAO,CAAC,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;QACxF,IAAI,SAAS,GAAG,KAAK,CAAC;QAEtB,sCAAsC;QACtC,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,OAAO,GAAG,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC;YAClC,kBAAkB;YAClB,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAE,CAAC,EAAE,CAAC;gBACnC,GAAG,EAAE,CAAC;gBACN,SAAS;YACX,CAAC;YAED,yEAAyE;YACzE,wEAAwE;YACxE,iCAAiC;YACjC,IAAI,SAAS,KAAK,SAAS,IAAI,MAAM,CAAC,MAAM,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;gBAC9D,SAAS,GAAG,IAAI,CAAC;gBACjB,MAAM;YACR,CAAC;YAED,4DAA4D;YAC5D,MAAM,EAAE,KAAK,EAAE,EAAE,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC,aAAa,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC;YAErE,MAAM,aAAa,GAAG,GAAG,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,GAAG,CAAC,CAAE,CAAC,CAAC;YAEtE,MAAM,CAAC,IAAI,CAAC;gBACV,EAAE;gBACF,KAAK;gBACL,KAAK,EAAE,GAAG;gBACV,GAAG,EAAE,GAAG,GAAG,MAAM;gBACjB,cAAc,EAAE,CAAC,aAAa,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;gBACxD,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,eAAe,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC;YAE1C,GAAG,IAAI,MAAM,CAAC;QAChB,CAAC;QAED,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,IAAI,CAAC,MAAM;YAClB,GAAG,EAAE,IAAI,CAAC,MAAM;YAChB,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,eAAe;QACf,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACzC,MAAM,aAAa,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAEzC,OAAO;YACL,MAAM;YACN,QAAQ;YACR,aAAa;YACb,YAAY;YACZ,eAAe;YACf,SAAS;SACV,CAAC;IACJ,CAAC;IAED;;;OAGG;IACK,aAAa,CAAC,IAAY,EAAE,QAAgB;QAClD,+DAA+D;QAC/D,MAAM,WAAW,GAAG,QAAQ,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAE,CAAC,CAAC;QAErE,6DAA6D;QAC7D,mDAAmD;QACnD,IAAI,WAAW,EAAE,CAAC;YAChB,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,CAAC,GAAG,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;YAC9D,IAAI,MAAM,EAAE,CAAC;gBACX,OAAO,MAAM,CAAC;YAChB,CAAC;QACH,CAAC;QAED,8BAA8B;QAC9B,MAAM,MAAM,GAAG,IAAI,CAAC,oBAAoB,CAAC,EAAE,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;QAC7D,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,4BAA4B;QAC5B,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAE,CAAC;QAC7B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,GAAG,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC;QAChF,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAChD,CAAC;IAED;;OAEG;IACK,oBAAoB,CAC1B,MAAc,EACd,IAAY,EACZ,QAAgB;QAEhB,IAAI,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;QACrB,IAAI,SAAS,GAAyD,IAAI,CAAC;QAE3E,yCAAyC;QACzC,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtC,IAAI,CAAC,KAAK;gBAAE,OAAO,IAAI,CAAC;YACxB,IAAI,GAAG,KAAK,CAAC;QACf,CAAC;QAED,+BAA+B;QAC/B,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,KAAK,IAAI,CAAC,GAAG,QAAQ,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC5C,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAE,CAAC;YACtB,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtC,IAAI,CAAC,KAAK;gBAAE,MAAM;YAElB,IAAI,GAAG,KAAK,CAAC;YACb,WAAW,EAAE,CAAC;YAEd,2EAA2E;YAC3E,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,KAAK,KAAK,IAAI,EAAE,CAAC;gBACjD,SAAS,GAAG;oBACV,KAAK,EAAE,IAAI,CAAC,KAAK;oBACjB,EAAE,EAAE,IAAI,CAAC,OAAO;oBAChB,MAAM,EAAE,WAAW,EAAE,2CAA2C;iBACjE,CAAC;YACJ,CAAC;QACH,CAAC;QAED,OAAO,SAAS,CAAC;IACnB,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,QAAkB;QACvB,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACxC,IAAI,KAAK,KAAK,SAAS;gBAAE,SAAS;YAClC,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ;gBAAE,SAAS;YAE5F,+CAA+C;YAC/C,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC1B,KAAK,CAAC,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACnC,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACpB,CAAC;QACH,CAAC;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,IAAI,SAAS;QACX,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,UAAU,CAAC,KAAa;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,EAAU;QACjB,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IACnC,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,QAAgB;IACtD,MAAM,EAAE,kBAAkB,EAAE,GAAG,MAAM,MAAM,CAAC,UAAU,CAAC,CAAC;IACxD,MAAM,OAAO,GAAG,MAAM,kBAAkB,EAAE,CAAC;IAC3C,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IAErD,gBAAgB;IAChB,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QACjE,OAAO,oBAAoB,CAAC,OAAO,CAAC,CAAC;IACvC,CAAC;SAAM,CAAC;QACN,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC;IAC7B,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,oBAAoB,CAAC,OAAe;IAClD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IAExC,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAsB,CAAC;QAExD,2BAA2B;QAC3B,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC;YACvC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAED,uBAAuB;QACvB,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;YACnE,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtC,gDAAgD;gBAChD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACnD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;oBACpC,IAAI,KAAK,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;wBAC1C,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;oBACzB,CAAC;gBACH,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,mDAAmD;gBACnD,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;oBAC7D,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;gBACvB,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,mCAAmC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAClE,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,OAAe;IACxC,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC/B,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5C,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,eAAe;IAC7B,MAAM,MAAM,GAAG;QACb,KAAK;QACL,OAAO;QACP,MAAM;QACN,OAAO;QACP,QAAQ;QACR,OAAO;QACP,QAAQ;QACR,OAAO;QACP,OAAO;QACP,OAAO;QACP,KAAK;QACL,SAAS;QACT,GAAG;KACJ,CAAC;IAEF,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC9B,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;IAC1B,CAAC,CAAC,CAAC;IAEH,OAAO,KAAK,CAAC;AACf,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"git-identity.d.ts","sourceRoot":"","sources":["../../src/opencode-plugin/git-identity.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AAE1D,eAAO,MAAM,qBAAqB,EAAE,UAwCnC,CAAC"}
|
|
@@ -0,0 +1,37 @@
|
|
|
1
|
+
import { DetectionSource, PIIType } from "../types/index.js";
|
|
2
|
+
export const gitIdentityRecognizer = {
|
|
3
|
+
type: PIIType.PERSON,
|
|
4
|
+
name: "GitIdentityRecognizer",
|
|
5
|
+
defaultConfidence: 1,
|
|
6
|
+
find(text) {
|
|
7
|
+
const names = new Set();
|
|
8
|
+
for (const match of text.matchAll(/^(?:Author|Commit|Committer):[\t ]*(.+)$/gm)) {
|
|
9
|
+
names.add(match[1].replace(/[\t ]+<[^>]*>.*$/, "").trim());
|
|
10
|
+
}
|
|
11
|
+
for (const match of text.matchAll(/^(?:author|committer) (.+)$/gm)) {
|
|
12
|
+
names.add(match[1].replace(/[\t ]+<[^>]*>.*$/, '').trim());
|
|
13
|
+
}
|
|
14
|
+
for (const match of text.matchAll(/^.*\((.+?)\s+\d{4}-\d{2}-\d{2}\s[^)\n]*\)/gm)) {
|
|
15
|
+
names.add(match[1].trim());
|
|
16
|
+
}
|
|
17
|
+
const matches = [];
|
|
18
|
+
for (const name of names) {
|
|
19
|
+
if (name === "")
|
|
20
|
+
continue;
|
|
21
|
+
const escaped = name.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
|
|
22
|
+
const pattern = new RegExp(`(?<![\\p{L}\\p{N}])${escaped}(?![\\p{L}\\p{N}])`, "giu");
|
|
23
|
+
for (const match of text.matchAll(pattern)) {
|
|
24
|
+
matches.push({
|
|
25
|
+
type: PIIType.PERSON,
|
|
26
|
+
start: match.index,
|
|
27
|
+
end: match.index + match[0].length,
|
|
28
|
+
confidence: 1,
|
|
29
|
+
source: DetectionSource.REGEX,
|
|
30
|
+
text: match[0],
|
|
31
|
+
});
|
|
32
|
+
}
|
|
33
|
+
}
|
|
34
|
+
return matches;
|
|
35
|
+
},
|
|
36
|
+
};
|
|
37
|
+
//# sourceMappingURL=git-identity.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"git-identity.js","sourceRoot":"","sources":["../../src/opencode-plugin/git-identity.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,OAAO,EAAkB,MAAM,mBAAmB,CAAC;AAG7E,MAAM,CAAC,MAAM,qBAAqB,GAAe;IAC/C,IAAI,EAAE,OAAO,CAAC,MAAM;IACpB,IAAI,EAAE,uBAAuB;IAC7B,iBAAiB,EAAE,CAAC;IAEpB,IAAI,CAAC,IAAY;QACf,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;QAEhC,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,4CAA4C,CAAC,EAAE,CAAC;YAChF,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC9D,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,+BAA+B,CAAC,EAAE,CAAC;YACnE,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC9D,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,6CAA6C,CAAC,EAAE,CAAC;YACjF,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC,CAAC;QAC9B,CAAC;QAED,MAAM,OAAO,GAAgB,EAAE,CAAC;QAChC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,IAAI,KAAK,EAAE;gBAAE,SAAS;YAC1B,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;YAC5D,MAAM,OAAO,GAAG,IAAI,MAAM,CACxB,sBAAsB,OAAO,oBAAoB,EACjD,KAAK,CACN,CAAC;YACF,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC3C,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,OAAO,CAAC,MAAM;oBACpB,KAAK,EAAE,KAAK,CAAC,KAAK;oBAClB,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM;oBAClC,UAAU,EAAE,CAAC;oBACb,MAAM,EAAE,eAAe,CAAC,KAAK;oBAC7B,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;iBACf,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;CACF,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"github-username.d.ts","sourceRoot":"","sources":["../../src/opencode-plugin/github-username.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AAI1D,eAAO,MAAM,wBAAwB,EAAE,UAsCtC,CAAC"}
|
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
import { DetectionSource, PIIType } from "../types/index.js";
|
|
2
|
+
const LOGIN = String.raw `[A-Za-z0-9](?:[A-Za-z0-9-]{0,37}[A-Za-z0-9])?(?:\[bot\])?`;
|
|
3
|
+
export const githubUsernameRecognizer = {
|
|
4
|
+
type: PIIType.GITHUB_USERNAME,
|
|
5
|
+
name: "GitHubUsernameRecognizer",
|
|
6
|
+
defaultConfidence: 1,
|
|
7
|
+
find(text) {
|
|
8
|
+
const logins = new Set();
|
|
9
|
+
const matches = new Map();
|
|
10
|
+
function add(start, login) {
|
|
11
|
+
logins.add(login);
|
|
12
|
+
matches.set(start, {
|
|
13
|
+
type: PIIType.GITHUB_USERNAME, start, end: start + login.length,
|
|
14
|
+
confidence: 1, source: DetectionSource.REGEX, text: login,
|
|
15
|
+
});
|
|
16
|
+
}
|
|
17
|
+
// Restrict JSON matches to values, so a login like "id" cannot mask JSON keys.
|
|
18
|
+
for (const match of text.matchAll(new RegExp(String.raw `"login"\s*:\s*"(${LOGIN})"`, 'g'))) {
|
|
19
|
+
add(match.index + match[0].lastIndexOf(match[1]), match[1]);
|
|
20
|
+
}
|
|
21
|
+
for (const match of text.matchAll(/^(?:author|assignees|reviewers):[\t ]*(.+)$/gim)) {
|
|
22
|
+
const list = match[1];
|
|
23
|
+
const offset = match.index + match[0].indexOf(list);
|
|
24
|
+
const entry = new RegExp(String.raw `(?:^|,)[\t ]*(${LOGIN})(?=[\t ,(]|$)`, 'g');
|
|
25
|
+
for (const item of list.matchAll(entry)) {
|
|
26
|
+
add(offset + item.index + item[0].lastIndexOf(item[1]), item[1]);
|
|
27
|
+
}
|
|
28
|
+
}
|
|
29
|
+
// Cover mentions of discovered participants, but preserve npm scopes.
|
|
30
|
+
for (const login of logins) {
|
|
31
|
+
const escaped = login.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
|
|
32
|
+
for (const match of text.matchAll(new RegExp(`(?<![A-Za-z0-9_/@])@(${escaped})(?![A-Za-z0-9-/])`, 'gi'))) {
|
|
33
|
+
const value = match[1];
|
|
34
|
+
matches.set(match.index + 1, { type: PIIType.GITHUB_USERNAME, start: match.index + 1,
|
|
35
|
+
end: match.index + 1 + value.length, confidence: 1, source: DetectionSource.REGEX, text: value });
|
|
36
|
+
}
|
|
37
|
+
}
|
|
38
|
+
return [...matches.values()];
|
|
39
|
+
},
|
|
40
|
+
};
|
|
41
|
+
//# sourceMappingURL=github-username.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"github-username.js","sourceRoot":"","sources":["../../src/opencode-plugin/github-username.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,OAAO,EAAkB,MAAM,mBAAmB,CAAC;AAG7E,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAA,2DAA2D,CAAC;AAEpF,MAAM,CAAC,MAAM,wBAAwB,GAAe;IAClD,IAAI,EAAE,OAAO,CAAC,eAAe;IAC7B,IAAI,EAAE,0BAA0B;IAChC,iBAAiB,EAAE,CAAC;IAEpB,IAAI,CAAC,IAAY;QACf,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;QACjC,MAAM,OAAO,GAAG,IAAI,GAAG,EAAqB,CAAC;QAC7C,SAAS,GAAG,CAAC,KAAa,EAAE,KAAa;YACvC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YAClB,OAAO,CAAC,GAAG,CAAC,KAAK,EAAE;gBACjB,IAAI,EAAE,OAAO,CAAC,eAAe,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM;gBAC/D,UAAU,EAAE,CAAC,EAAE,MAAM,EAAE,eAAe,CAAC,KAAK,EAAE,IAAI,EAAE,KAAK;aAC1D,CAAC,CAAC;QACL,CAAC;QACD,+EAA+E;QAC/E,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,mBAAmB,KAAK,IAAI,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;YAC3F,GAAG,CAAC,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,EAAE,KAAK,CAAC,CAAC,CAAE,CAAC,CAAC;QAChE,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,gDAAgD,CAAC,EAAE,CAAC;YACpF,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;YACvB,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YACpD,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,iBAAiB,KAAK,gBAAgB,EAAE,GAAG,CAAC,CAAC;YAChF,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;gBACxC,GAAG,CAAC,MAAM,GAAG,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,EAAE,IAAI,CAAC,CAAC,CAAE,CAAC,CAAC;YACrE,CAAC;QACH,CAAC;QACD,sEAAsE;QACtE,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;YAC7D,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,wBAAwB,OAAO,oBAAoB,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC;gBACzG,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,eAAe,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,GAAG,CAAC;oBAClF,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,UAAU,EAAE,CAAC,EAAE,MAAM,EAAE,eAAe,CAAC,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;YACtG,CAAC;QACH,CAAC;QACD,OAAO,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAC/B,CAAC;CACF,CAAC"}
|
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
import type { Recognizer } from '../recognizers/base.js';
|
|
2
|
+
import type { TagFormat } from '../types/index.js';
|
|
3
|
+
/** Keep repeated history transforms from discovering an existing tag as a name. */
|
|
4
|
+
export declare function protectIdentityTags(recognizer: Recognizer, format: TagFormat): Recognizer;
|
|
5
|
+
//# sourceMappingURL=identity-recognizer.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"identity-recognizer.d.ts","sourceRoot":"","sources":["../../src/opencode-plugin/identity-recognizer.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEzD,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAEnD,mFAAmF;AACnF,wBAAgB,mBAAmB,CAAC,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,GAAG,UAAU,CAkBzF"}
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
import { extractTagsStrict } from '../pipeline/tagger.js';
|
|
2
|
+
/** Keep repeated history transforms from discovering an existing tag as a name. */
|
|
3
|
+
export function protectIdentityTags(recognizer, format) {
|
|
4
|
+
return {
|
|
5
|
+
...recognizer,
|
|
6
|
+
find(text) {
|
|
7
|
+
const tags = extractTagsStrict(text, format);
|
|
8
|
+
if (tags.length === 0)
|
|
9
|
+
return recognizer.find(text);
|
|
10
|
+
const pieces = [];
|
|
11
|
+
let offset = 0;
|
|
12
|
+
for (const tag of tags) {
|
|
13
|
+
pieces.push(text.slice(offset, tag.position), ' '.repeat(tag.matchedText.length));
|
|
14
|
+
offset = tag.position + tag.matchedText.length;
|
|
15
|
+
}
|
|
16
|
+
pieces.push(text.slice(offset));
|
|
17
|
+
return recognizer.find(pieces.join('')).filter(match => !tags.some(tag => match.start < tag.position + tag.matchedText.length && match.end > tag.position));
|
|
18
|
+
},
|
|
19
|
+
};
|
|
20
|
+
}
|
|
21
|
+
//# sourceMappingURL=identity-recognizer.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"identity-recognizer.js","sourceRoot":"","sources":["../../src/opencode-plugin/identity-recognizer.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,iBAAiB,EAAE,MAAM,uBAAuB,CAAC;AAG1D,mFAAmF;AACnF,MAAM,UAAU,mBAAmB,CAAC,UAAsB,EAAE,MAAiB;IAC3E,OAAO;QACL,GAAG,UAAU;QACb,IAAI,CAAC,IAAI;YACP,MAAM,IAAI,GAAG,iBAAiB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;YAC7C,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACpD,MAAM,MAAM,GAAa,EAAE,CAAC;YAC5B,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;gBACvB,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,EAAE,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC;gBAClF,MAAM,GAAG,GAAG,CAAC,QAAQ,GAAG,GAAG,CAAC,WAAW,CAAC,MAAM,CAAC;YACjD,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;YAChC,OAAO,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CACrD,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,KAAK,CAAC,KAAK,GAAG,GAAG,CAAC,QAAQ,GAAG,GAAG,CAAC,WAAW,CAAC,MAAM,IAAI,KAAK,CAAC,GAAG,GAAG,GAAG,CAAC,QAAQ,CAAC,CACnG,CAAC;QACJ,CAAC;KACF,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,7 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Rehydra OpenCode Plugin (V2)
|
|
3
|
+
* Anonymizes secrets in LLM context via OpenCode's Plugin V2 hooks.
|
|
4
|
+
*/
|
|
5
|
+
export { default, plugin, createRehydraPlugin } from "./plugin.js";
|
|
6
|
+
export type { RehydraPluginOptions } from "./types.js";
|
|
7
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/opencode-plugin/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAC;AACnE,YAAY,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/opencode-plugin/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAC"}
|
|
@@ -0,0 +1,39 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Rehydra OpenCode Plugin — Plugin V2.
|
|
3
|
+
*
|
|
4
|
+
* Uses OpenCode Plugin V2 hooks to anonymize secrets before they reach the LLM
|
|
5
|
+
* and rehydrate PII tags before local tool execution:
|
|
6
|
+
* - session.hook("context" | "compaction" | "generate") — anonymize all request
|
|
7
|
+
* messages in place and inject the rehydra instruction when needed.
|
|
8
|
+
* - session.hook("title") — anonymize the title request messages only.
|
|
9
|
+
* - tool.hook("execute.before") — rehydrate PII tags in tool arguments.
|
|
10
|
+
* - tool.hook("execute.after") — rehydrate PII tags in completed tool results.
|
|
11
|
+
* - session.hook("http.response") — rehydrate PII tags in the primary answer
|
|
12
|
+
* body (JSON and SSE) before OpenCode renders it.
|
|
13
|
+
* - session.hook("experimental.ws.receive") — rehydrate PII tags in WebSocket
|
|
14
|
+
* frames carrying model output.
|
|
15
|
+
*/
|
|
16
|
+
import type { RehydraPluginOptions } from "./types.js";
|
|
17
|
+
import type { Plugin, RehydraPluginContext } from "./v2/types.js";
|
|
18
|
+
/**
|
|
19
|
+
* Resolves the project directory from the V2 context, falling back through the
|
|
20
|
+
* legacy V1 fields and finally the process working directory.
|
|
21
|
+
*/
|
|
22
|
+
export declare function directoryFromContext(ctx: RehydraPluginContext): string;
|
|
23
|
+
/**
|
|
24
|
+
* Creates a Rehydra plugin for OpenCode.
|
|
25
|
+
*
|
|
26
|
+
* @example
|
|
27
|
+
* ```typescript
|
|
28
|
+
* // .opencode/plugins/rehydra.ts
|
|
29
|
+
* import { createRehydraPlugin } from "rehydra/opencode-plugin";
|
|
30
|
+
* export default createRehydraPlugin({
|
|
31
|
+
* envFiles: [".env", ".env.local"],
|
|
32
|
+
* });
|
|
33
|
+
* ```
|
|
34
|
+
*/
|
|
35
|
+
export declare function createRehydraPlugin(options?: RehydraPluginOptions): Plugin;
|
|
36
|
+
/** Default plugin instance used by `opencode.json`. */
|
|
37
|
+
export declare const plugin: Plugin;
|
|
38
|
+
export default plugin;
|
|
39
|
+
//# sourceMappingURL=plugin.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"plugin.d.ts","sourceRoot":"","sources":["../../src/opencode-plugin/plugin.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAcH,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAC;AA2BvD,OAAO,KAAK,EAGV,MAAM,EACN,oBAAoB,EAGrB,MAAM,eAAe,CAAC;AAEvB;;;GAGG;AACH,wBAAgB,oBAAoB,CAAC,GAAG,EAAE,oBAAoB,GAAG,MAAM,CAUtE;AAkFD;;;;;;;;;;;GAWG;AACH,wBAAgB,mBAAmB,CAAC,OAAO,CAAC,EAAE,oBAAoB,GAAG,MAAM,CAiN1E;AAED,uDAAuD;AACvD,eAAO,MAAM,MAAM,EAAE,MAA8B,CAAC;AAEpD,eAAe,MAAM,CAAC"}
|