mcp-scraper 0.40.2 → 0.40.3

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.
Files changed (62) hide show
  1. package/README.md +1 -1
  2. package/dist/bin/api-server.cjs +1517 -236
  3. package/dist/bin/api-server.cjs.map +1 -1
  4. package/dist/bin/api-server.js +3 -3
  5. package/dist/bin/mcp-scraper-cli.cjs +1 -1
  6. package/dist/bin/mcp-scraper-cli.cjs.map +1 -1
  7. package/dist/bin/mcp-scraper-cli.js +1 -1
  8. package/dist/bin/mcp-scraper-install.cjs +1 -1
  9. package/dist/bin/mcp-scraper-install.cjs.map +1 -1
  10. package/dist/bin/mcp-scraper-install.js +1 -1
  11. package/dist/bin/mcp-stdio-server.cjs +155 -87
  12. package/dist/bin/mcp-stdio-server.cjs.map +1 -1
  13. package/dist/bin/mcp-stdio-server.js +6 -5
  14. package/dist/bin/mcp-stdio-server.js.map +1 -1
  15. package/dist/bin/paa-harvest.cjs +31 -9
  16. package/dist/bin/paa-harvest.cjs.map +1 -1
  17. package/dist/bin/paa-harvest.js +4 -4
  18. package/dist/{chunk-XORPNO3Z.js → chunk-27DUCRAZ.js} +2 -2
  19. package/dist/{chunk-44HZLHDV.js → chunk-5UN33CGU.js} +38 -1
  20. package/dist/chunk-5UN33CGU.js.map +1 -0
  21. package/dist/{chunk-X2LKCX6H.js → chunk-6J57U6HA.js} +3 -3
  22. package/dist/{chunk-LP6E462I.js → chunk-6SZ52BQ5.js} +49 -101
  23. package/dist/chunk-6SZ52BQ5.js.map +1 -0
  24. package/dist/chunk-BBI7RGOT.js +99 -0
  25. package/dist/chunk-BBI7RGOT.js.map +1 -0
  26. package/dist/{chunk-NVXNEOUQ.js → chunk-CS4HE6QY.js} +47 -11
  27. package/dist/chunk-CS4HE6QY.js.map +1 -0
  28. package/dist/{chunk-3PIWJS6Y.js → chunk-EHES33KB.js} +2 -2
  29. package/dist/{chunk-GUVKHCKE.js → chunk-IQTT7CAB.js} +94 -3
  30. package/dist/chunk-IQTT7CAB.js.map +1 -0
  31. package/dist/{chunk-7XBBFBYY.js → chunk-JVIW4GK2.js} +44 -12
  32. package/dist/chunk-JVIW4GK2.js.map +1 -0
  33. package/dist/chunk-XU4ZLIB2.js +7 -0
  34. package/dist/chunk-XU4ZLIB2.js.map +1 -0
  35. package/dist/{db-W3CP562I.js → db-LVVU6NOK.js} +16 -2
  36. package/dist/{extract-bundle-K4PG3RZJ.js → extract-bundle-SKXEDG4Z.js} +4 -4
  37. package/dist/index.cjs +31 -9
  38. package/dist/index.cjs.map +1 -1
  39. package/dist/index.js +4 -4
  40. package/dist/{location-data-repository-RLQX6SNM.js → location-data-repository-NBTCOA7B.js} +3 -3
  41. package/dist/{server-KM3CFWCF.js → server-W76RUYD3.js} +1108 -66
  42. package/dist/server-W76RUYD3.js.map +1 -0
  43. package/dist/{site-extract-repository-2SMMFKKL.js → site-extract-repository-6GX72U4N.js} +4 -4
  44. package/dist/{worker-FXAGFYOE.js → worker-5RWLFSAJ.js} +27 -28
  45. package/dist/worker-5RWLFSAJ.js.map +1 -0
  46. package/package.json +1 -1
  47. package/dist/chunk-44HZLHDV.js.map +0 -1
  48. package/dist/chunk-7XBBFBYY.js.map +0 -1
  49. package/dist/chunk-GUVKHCKE.js.map +0 -1
  50. package/dist/chunk-LP6E462I.js.map +0 -1
  51. package/dist/chunk-NVXNEOUQ.js.map +0 -1
  52. package/dist/chunk-ZUJLSICT.js +0 -7
  53. package/dist/chunk-ZUJLSICT.js.map +0 -1
  54. package/dist/server-KM3CFWCF.js.map +0 -1
  55. package/dist/worker-FXAGFYOE.js.map +0 -1
  56. /package/dist/{chunk-XORPNO3Z.js.map → chunk-27DUCRAZ.js.map} +0 -0
  57. /package/dist/{chunk-X2LKCX6H.js.map → chunk-6J57U6HA.js.map} +0 -0
  58. /package/dist/{chunk-3PIWJS6Y.js.map → chunk-EHES33KB.js.map} +0 -0
  59. /package/dist/{db-W3CP562I.js.map → db-LVVU6NOK.js.map} +0 -0
  60. /package/dist/{extract-bundle-K4PG3RZJ.js.map → extract-bundle-SKXEDG4Z.js.map} +0 -0
  61. /package/dist/{location-data-repository-RLQX6SNM.js.map → location-data-repository-NBTCOA7B.js.map} +0 -0
  62. /package/dist/{site-extract-repository-2SMMFKKL.js.map → site-extract-repository-6GX72U4N.js.map} +0 -0
@@ -1 +0,0 @@
1
- {"version":3,"sources":["../src/api/harvest-problems.ts","../src/api/harvest-attempt-events.ts"],"sourcesContent":["import { CaptchaError, LocationMismatchError, RequestAbortedError } from '../errors.js'\n\nexport type HarvestProblemCode =\n | 'request_aborted'\n | 'captcha_exhausted'\n | 'location_mismatch'\n | 'proxy_tunnel_failed'\n | 'proxy_unavailable'\n | 'harvest_timeout'\n | 'extraction_failed'\n\nexport interface HarvestProblem {\n error_code: HarvestProblemCode\n error_type: string\n message: string\n retryable: boolean\n httpStatus: number\n terminalStatus: 'cancelled' | 'failed'\n}\n\nfunction errorMessage(err: unknown): string {\n return err instanceof Error ? err.message : String(err)\n}\n\nfunction looksLikeTimeout(err: unknown, message: string): boolean {\n if (err instanceof DOMException && (err.name === 'TimeoutError' || err.name === 'AbortError')) return true\n return /timeout|timed out|Timeout \\d+ms exceeded|deadline/i.test(message)\n}\n\nfunction looksLikeCaptcha(message: string): boolean {\n return /captcha|recaptcha|unusual traffic|google\\.com\\/sorry|blocked/i.test(message)\n}\n\nfunction looksLikeProxyTunnelFailure(message: string): boolean {\n return /ERR_TUNNEL_CONNECTION_FAILED|ERR_PROXY_CONNECTION_FAILED|ERR_SOCKS_CONNECTION_FAILED|tunnel connection failed|proxy connection failed|transport error: proxy/i.test(message)\n}\n\nfunction looksLikeProxyUnavailable(message: string): boolean {\n return /proxy unavailable|proxy_unavailable|connection_test_failed|did not return a proxy id|configured fallback/i.test(message)\n}\n\nexport function classifyHarvestProblem(err: unknown): HarvestProblem {\n const message = errorMessage(err)\n\n if (err instanceof RequestAbortedError) {\n return {\n error_code: 'request_aborted',\n error_type: 'request_aborted',\n message,\n retryable: true,\n httpStatus: 408,\n terminalStatus: 'cancelled',\n }\n }\n\n if (err instanceof CaptchaError || looksLikeCaptcha(message)) {\n return {\n error_code: 'captcha_exhausted',\n error_type: 'captcha',\n message,\n retryable: true,\n httpStatus: 503,\n terminalStatus: 'failed',\n }\n }\n\n if (err instanceof LocationMismatchError) {\n return {\n error_code: 'location_mismatch',\n error_type: 'location_mismatch',\n message,\n retryable: true,\n httpStatus: 503,\n terminalStatus: 'failed',\n }\n }\n\n if (looksLikeProxyTunnelFailure(message)) {\n return {\n error_code: 'proxy_tunnel_failed',\n error_type: 'proxy_tunnel_failed',\n message,\n retryable: true,\n httpStatus: 503,\n terminalStatus: 'failed',\n }\n }\n\n if (looksLikeProxyUnavailable(message)) {\n return {\n error_code: 'proxy_unavailable',\n error_type: 'proxy_unavailable',\n message,\n retryable: true,\n httpStatus: 503,\n terminalStatus: 'failed',\n }\n }\n\n if (looksLikeTimeout(err, message)) {\n return {\n error_code: 'harvest_timeout',\n error_type: 'timeout',\n message,\n retryable: true,\n httpStatus: 504,\n terminalStatus: 'failed',\n }\n }\n\n return {\n error_code: 'extraction_failed',\n error_type: 'extraction',\n message,\n retryable: false,\n httpStatus: 500,\n terminalStatus: 'failed',\n }\n}\n\nexport function serializeHarvestProblem(problem: HarvestProblem): string {\n return JSON.stringify({\n error_code: problem.error_code,\n error_type: problem.error_type,\n message: problem.message,\n retryable: problem.retryable,\n })\n}\n\nexport function harvestProblemResponse(problem: HarvestProblem): {\n error: string\n error_code: HarvestProblemCode\n error_type: string\n retryable: boolean\n} {\n return {\n error: problem.message,\n error_code: problem.error_code,\n error_type: problem.error_type,\n retryable: problem.retryable,\n }\n}\n","import type { HarvestAttemptLogEvent } from '../harvest.js'\nimport { finishHarvestAttempt, startHarvestAttempt } from './db.js'\n\nexport function createHarvestAttemptRecorder(jobId: string, userId: number | bigint) {\n return async (event: HarvestAttemptLogEvent): Promise<void> => {\n if (event.type === 'started') {\n await startHarvestAttempt({\n jobId,\n userId,\n attemptNumber: event.attemptNumber,\n maxAttempts: event.maxAttempts,\n query: event.query,\n location: event.location,\n maxQuestions: event.maxQuestions,\n startedAt: event.startedAt,\n })\n return\n }\n\n await finishHarvestAttempt({\n jobId,\n attemptNumber: event.attemptNumber,\n outcome: event.outcome,\n kernelSessionId: event.kernelSessionId,\n questionCount: event.questionCount,\n durationMs: event.durationMs,\n error: event.error,\n willRetry: event.willRetry,\n kernelDeleteStarted: event.cleanup.kernelDeleteStarted,\n kernelDeleteSucceeded: event.cleanup.kernelDeleteSucceeded,\n kernelDeleteError: event.cleanup.kernelDeleteError,\n browserCloseSucceeded: event.cleanup.browserCloseSucceeded,\n browserCloseError: event.cleanup.browserCloseError,\n debug: event.debug,\n completedAt: event.completedAt,\n })\n }\n}\n"],"mappings":";;;;;;;;;;;AAoBA,SAAS,aAAa,KAAsB;AAC1C,SAAO,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AACxD;AAEA,SAAS,iBAAiB,KAAc,SAA0B;AAChE,MAAI,eAAe,iBAAiB,IAAI,SAAS,kBAAkB,IAAI,SAAS,cAAe,QAAO;AACtG,SAAO,qDAAqD,KAAK,OAAO;AAC1E;AAEA,SAAS,iBAAiB,SAA0B;AAClD,SAAO,gEAAgE,KAAK,OAAO;AACrF;AAEA,SAAS,4BAA4B,SAA0B;AAC7D,SAAO,gKAAgK,KAAK,OAAO;AACrL;AAEA,SAAS,0BAA0B,SAA0B;AAC3D,SAAO,4GAA4G,KAAK,OAAO;AACjI;AAEO,SAAS,uBAAuB,KAA8B;AACnE,QAAM,UAAU,aAAa,GAAG;AAEhC,MAAI,eAAe,qBAAqB;AACtC,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,MAAI,eAAe,gBAAgB,iBAAiB,OAAO,GAAG;AAC5D,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,MAAI,eAAe,uBAAuB;AACxC,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,MAAI,4BAA4B,OAAO,GAAG;AACxC,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,MAAI,0BAA0B,OAAO,GAAG;AACtC,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,MAAI,iBAAiB,KAAK,OAAO,GAAG;AAClC,WAAO;AAAA,MACL,YAAY;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,WAAW;AAAA,MACX,YAAY;AAAA,MACZ,gBAAgB;AAAA,IAClB;AAAA,EACF;AAEA,SAAO;AAAA,IACL,YAAY;AAAA,IACZ,YAAY;AAAA,IACZ;AAAA,IACA,WAAW;AAAA,IACX,YAAY;AAAA,IACZ,gBAAgB;AAAA,EAClB;AACF;AAEO,SAAS,wBAAwB,SAAiC;AACvE,SAAO,KAAK,UAAU;AAAA,IACpB,YAAY,QAAQ;AAAA,IACpB,YAAY,QAAQ;AAAA,IACpB,SAAS,QAAQ;AAAA,IACjB,WAAW,QAAQ;AAAA,EACrB,CAAC;AACH;AAEO,SAAS,uBAAuB,SAKrC;AACA,SAAO;AAAA,IACL,OAAO,QAAQ;AAAA,IACf,YAAY,QAAQ;AAAA,IACpB,YAAY,QAAQ;AAAA,IACpB,WAAW,QAAQ;AAAA,EACrB;AACF;;;AC1IO,SAAS,6BAA6B,OAAe,QAAyB;AACnF,SAAO,OAAO,UAAiD;AAC7D,QAAI,MAAM,SAAS,WAAW;AAC5B,YAAM,oBAAoB;AAAA,QACxB;AAAA,QACA;AAAA,QACA,eAAe,MAAM;AAAA,QACrB,aAAa,MAAM;AAAA,QACnB,OAAO,MAAM;AAAA,QACb,UAAU,MAAM;AAAA,QAChB,cAAc,MAAM;AAAA,QACpB,WAAW,MAAM;AAAA,MACnB,CAAC;AACD;AAAA,IACF;AAEA,UAAM,qBAAqB;AAAA,MACzB;AAAA,MACA,eAAe,MAAM;AAAA,MACrB,SAAS,MAAM;AAAA,MACf,iBAAiB,MAAM;AAAA,MACvB,eAAe,MAAM;AAAA,MACrB,YAAY,MAAM;AAAA,MAClB,OAAO,MAAM;AAAA,MACb,WAAW,MAAM;AAAA,MACjB,qBAAqB,MAAM,QAAQ;AAAA,MACnC,uBAAuB,MAAM,QAAQ;AAAA,MACrC,mBAAmB,MAAM,QAAQ;AAAA,MACjC,uBAAuB,MAAM,QAAQ;AAAA,MACrC,mBAAmB,MAAM,QAAQ;AAAA,MACjC,OAAO,MAAM;AAAA,MACb,aAAa,MAAM;AAAA,IACrB,CAAC;AAAA,EACH;AACF;","names":[]}
@@ -1,7 +0,0 @@
1
- // src/version.ts
2
- var PACKAGE_VERSION = "0.40.2";
3
-
4
- export {
5
- PACKAGE_VERSION
6
- };
7
- //# sourceMappingURL=chunk-ZUJLSICT.js.map
@@ -1 +0,0 @@
1
- {"version":3,"sources":["../src/version.ts"],"sourcesContent":["export const PACKAGE_VERSION = '0.40.2'\n"],"mappings":";AAAO,IAAM,kBAAkB;","names":[]}