@cosmocoder/mcp-web-docs 2.0.21 → 2.0.23

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 (89) hide show
  1. package/build/__mocks__/embeddings.d.ts +0 -4
  2. package/build/__mocks__/embeddings.js +0 -9
  3. package/build/__mocks__/embeddings.js.map +1 -1
  4. package/build/config.d.ts +0 -2
  5. package/build/config.js +0 -40
  6. package/build/config.js.map +1 -1
  7. package/build/config.test.js +1 -36
  8. package/build/config.test.js.map +1 -1
  9. package/build/crawler/auth.d.ts +0 -17
  10. package/build/crawler/auth.js +1 -40
  11. package/build/crawler/auth.js.map +1 -1
  12. package/build/crawler/auth.test.js +4 -22
  13. package/build/crawler/auth.test.js.map +1 -1
  14. package/build/crawler/base.d.ts +0 -13
  15. package/build/crawler/base.js +4 -116
  16. package/build/crawler/base.js.map +1 -1
  17. package/build/crawler/base.test.js +2 -172
  18. package/build/crawler/base.test.js.map +1 -1
  19. package/build/crawler/content-utils.d.ts +0 -1
  20. package/build/crawler/content-utils.js +0 -8
  21. package/build/crawler/content-utils.js.map +1 -1
  22. package/build/crawler/content-utils.test.js +1 -41
  23. package/build/crawler/content-utils.test.js.map +1 -1
  24. package/build/crawler/crawlee-crawler.d.ts +2 -21
  25. package/build/crawler/crawlee-crawler.js +0 -1
  26. package/build/crawler/crawlee-crawler.js.map +1 -1
  27. package/build/crawler/crawlee-crawler.test.js +13 -0
  28. package/build/crawler/crawlee-crawler.test.js.map +1 -1
  29. package/build/crawler/docs-crawler.d.ts +3 -4
  30. package/build/crawler/docs-crawler.js +3 -5
  31. package/build/crawler/docs-crawler.js.map +1 -1
  32. package/build/crawler/docs-crawler.test.js +3 -9
  33. package/build/crawler/docs-crawler.test.js.map +1 -1
  34. package/build/crawler/github.d.ts +1 -1
  35. package/build/crawler/github.js +7 -7
  36. package/build/crawler/github.js.map +1 -1
  37. package/build/crawler/github.test.js +8 -19
  38. package/build/crawler/github.test.js.map +1 -1
  39. package/build/crawler/llms-txt.js +5 -3
  40. package/build/crawler/llms-txt.js.map +1 -1
  41. package/build/crawler/llms-txt.test.js +2 -0
  42. package/build/crawler/llms-txt.test.js.map +1 -1
  43. package/build/crawler/site-rules.js +6 -4
  44. package/build/crawler/site-rules.js.map +1 -1
  45. package/build/index.js +7 -1687
  46. package/build/index.js.map +1 -1
  47. package/build/index.test.js +124 -422
  48. package/build/index.test.js.map +1 -1
  49. package/build/indexing/queue-manager.js +3 -2
  50. package/build/indexing/queue-manager.js.map +1 -1
  51. package/build/indexing/status.d.ts +0 -5
  52. package/build/indexing/status.js +1 -26
  53. package/build/indexing/status.js.map +1 -1
  54. package/build/indexing/status.test.js +0 -24
  55. package/build/indexing/status.test.js.map +1 -1
  56. package/build/indexing/workflow.d.ts +38 -0
  57. package/build/indexing/workflow.js +223 -0
  58. package/build/indexing/workflow.js.map +1 -0
  59. package/build/indexing/workflow.test.d.ts +1 -0
  60. package/build/indexing/workflow.test.js +218 -0
  61. package/build/indexing/workflow.test.js.map +1 -0
  62. package/build/processor/processor.d.ts +2 -2
  63. package/build/processor/processor.test.js +3 -3
  64. package/build/processor/processor.test.js.map +1 -1
  65. package/build/server.d.ts +88 -0
  66. package/build/server.js +1460 -0
  67. package/build/server.js.map +1 -0
  68. package/build/server.test.d.ts +1 -0
  69. package/build/server.test.js +27 -0
  70. package/build/server.test.js.map +1 -0
  71. package/build/storage/storage.d.ts +2 -7
  72. package/build/storage/storage.js +1 -55
  73. package/build/storage/storage.js.map +1 -1
  74. package/build/storage/storage.test.js +25 -19
  75. package/build/storage/storage.test.js.map +1 -1
  76. package/build/types.d.ts +1 -29
  77. package/build/util/docs.js +1 -2
  78. package/build/util/docs.js.map +1 -1
  79. package/build/util/docs.test.js +8 -2
  80. package/build/util/docs.test.js.map +1 -1
  81. package/build/util/security.d.ts +1 -6
  82. package/build/util/security.js +7 -14
  83. package/build/util/security.js.map +1 -1
  84. package/build/util/security.test.js +7 -14
  85. package/build/util/security.test.js.map +1 -1
  86. package/package.json +2 -6
  87. package/build/crawler/content-extractors.d.ts +0 -9
  88. package/build/crawler/content-extractors.js +0 -9
  89. package/build/crawler/content-extractors.js.map +0 -1
@@ -1 +1 @@
1
- {"version":3,"file":"llms-txt.test.js","sourceRoot":"","sources":["../../src/crawler/llms-txt.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,eAAe,CAAC;AAEtE,QAAQ,CAAC,UAAU,EAAE,GAAG,EAAE;IACxB,QAAQ,CAAC,cAAc,EAAE,GAAG,EAAE;QAC5B,MAAM,OAAO,GAAG,+BAA+B,CAAC;QAEhD,EAAE,CAAC,iDAAiD,EAAE,GAAG,EAAE;YACzD,MAAM,OAAO,GAAG;;;;;;CAMrB,CAAC;YACI,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC;gBACnB,+CAA+C;gBAC/C,mCAAmC;gBACnC,sCAAsC;aACvC,CAAC,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,iDAAiD,EAAE,GAAG,EAAE;YACzD,MAAM,OAAO,GAAG;gDAC0B,CAAC;YAE3C,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,qCAAqC,CAAC,CAAC,CAAC;QACvG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yCAAyC,EAAE,GAAG,EAAE;YACjD,MAAM,OAAO,GAAG;wBACE,CAAC;YAEnB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,CAAC,CAAC,CAAC;QAChE,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yBAAyB,EAAE,GAAG,EAAE;YACjC,MAAM,OAAO,GAAG;;kDAE4B,CAAC;YAE7C,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,oCAAoC,CAAC,CAAC,CAAC;QAC/D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,6BAA6B,EAAE,GAAG,EAAE;YACrC,MAAM,OAAO,GAAG;sBACA,CAAC;YAEjB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,oCAAoC,CAAC,CAAC,CAAC;QAC/D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,gCAAgC,EAAE,GAAG,EAAE;YACxC,MAAM,OAAO,GAAG;;uBAEC,CAAC;YAElB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,CAAC,CAAC,CAAC;QAChE,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qDAAqD,EAAE,GAAG,EAAE;YAC7D,MAAM,CAAC,YAAY,CAAC,EAAE,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAC9C,MAAM,CAAC,YAAY,CAAC,kCAAkC,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAChF,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;YAC7C,MAAM,OAAO,GAAG,4EAA4E,CAAC;YAE7F,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,mCAAmC,CAAC,CAAC,CAAC;QACrG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,uDAAuD,EAAE,GAAG,EAAE;YAC/D,MAAM,OAAO,GAAG;8BACQ,CAAC;YAEzB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,wCAAwC,CAAC,CAAC,CAAC;QACnE,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,yBAAyB,EAAE,GAAG,EAAE;QACvC,UAAU,CAAC,GAAG,EAAE;YACd,SAAS,CAAC,UAAU,EAAE,CAAC;QACzB,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,iCAAiC,EAAE,KAAK,IAAI,EAAE;YAC/C,MAAM,WAAW,GAAG;;yBAED,CAAC;YAEpB,SAAS,CAAC,gBAAgB,CAAC,WAAW,EAAE;gBACtC,OAAO,EAAE,EAAE,cAAc,EAAE,YAAY,EAAE;aAC1C,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,SAAS,CAAC,CAAC,oBAAoB,CACpC,kCAAkC,EAClC,MAAM,CAAC,gBAAgB,CAAC;gBACtB,OAAO,EAAE,MAAM,CAAC,gBAAgB,CAAC;oBAC/B,MAAM,EAAE,gCAAgC;iBACzC,CAAC;aACH,CAAC,CACH,CAAC;YACF,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,qCAAqC,CAAC,CAAC,CAAC;QACvG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,kCAAkC,EAAE,KAAK,IAAI,EAAE;YAChD,SAAS,CAAC,gBAAgB,CAAC,WAAW,EAAE,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;YAEzD,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,4CAA4C,EAAE,KAAK,IAAI,EAAE;YAC1D,SAAS,CAAC,cAAc,CAAC,IAAI,KAAK,CAAC,eAAe,CAAC,CAAC,CAAC;YAErD,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,4DAA4D,EAAE,KAAK,IAAI,EAAE;YAC1E,SAAS,CAAC,gBAAgB,CAAC,qCAAqC,EAAE;gBAChE,OAAO,EAAE,EAAE,cAAc,EAAE,0BAA0B,EAAE;aACxD,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,8CAA8C,EAAE,KAAK,IAAI,EAAE;YAC5D,SAAS,CAAC,gBAAgB,CAAC,EAAE,EAAE;gBAC7B,OAAO,EAAE,EAAE,cAAc,EAAE,YAAY,EAAE;aAC1C,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
1
+ {"version":3,"file":"llms-txt.test.js","sourceRoot":"","sources":["../../src/crawler/llms-txt.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,eAAe,CAAC;AAEtE,QAAQ,CAAC,UAAU,EAAE,GAAG,EAAE;IACxB,QAAQ,CAAC,cAAc,EAAE,GAAG,EAAE;QAC5B,MAAM,OAAO,GAAG,+BAA+B,CAAC;QAEhD,EAAE,CAAC,iDAAiD,EAAE,GAAG,EAAE;YACzD,MAAM,OAAO,GAAG;;;;;;CAMrB,CAAC;YACI,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC;gBACnB,+CAA+C;gBAC/C,mCAAmC;gBACnC,sCAAsC;aACvC,CAAC,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,iDAAiD,EAAE,GAAG,EAAE;YACzD,MAAM,OAAO,GAAG;gDAC0B,CAAC;YAE3C,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,qCAAqC,CAAC,CAAC,CAAC;QACvG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yCAAyC,EAAE,GAAG,EAAE;YACjD,MAAM,OAAO,GAAG;wBACE,CAAC;YAEnB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,CAAC,CAAC,CAAC;QAChE,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yBAAyB,EAAE,GAAG,EAAE;YACjC,MAAM,OAAO,GAAG;;kDAE4B,CAAC;YAE7C,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,oCAAoC,CAAC,CAAC,CAAC;QAC/D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,6BAA6B,EAAE,GAAG,EAAE;YACrC,MAAM,OAAO,GAAG;sBACA,CAAC;YAEjB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,oCAAoC,CAAC,CAAC,CAAC;QAC/D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,gCAAgC,EAAE,GAAG,EAAE;YACxC,MAAM,OAAO,GAAG;;uBAEC,CAAC;YAElB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,CAAC,CAAC,CAAC;QAChE,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qDAAqD,EAAE,GAAG,EAAE;YAC7D,MAAM,CAAC,YAAY,CAAC,EAAE,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAC9C,MAAM,CAAC,YAAY,CAAC,kCAAkC,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAChF,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;YAC7C,MAAM,OAAO,GAAG,4EAA4E,CAAC;YAE7F,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,mCAAmC,CAAC,CAAC,CAAC;QACrG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,uDAAuD,EAAE,GAAG,EAAE;YAC/D,MAAM,OAAO,GAAG;8BACQ,CAAC;YAEzB,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE5C,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,wCAAwC,CAAC,CAAC,CAAC;QACnE,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,yBAAyB,EAAE,GAAG,EAAE;QACvC,UAAU,CAAC,GAAG,EAAE;YACd,SAAS,CAAC,UAAU,EAAE,CAAC;QACzB,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,iCAAiC,EAAE,KAAK,IAAI,EAAE;YAC/C,MAAM,WAAW,GAAG;;yBAED,CAAC;YAEpB,SAAS,CAAC,gBAAgB,CAAC,WAAW,EAAE;gBACtC,OAAO,EAAE,EAAE,cAAc,EAAE,YAAY,EAAE;aAC1C,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,SAAS,CAAC,CAAC,oBAAoB,CACpC,kCAAkC,EAClC,MAAM,CAAC,gBAAgB,CAAC;gBACtB,OAAO,EAAE,MAAM,CAAC,gBAAgB,CAAC;oBAC/B,MAAM,EAAE,gCAAgC;iBACzC,CAAC;aACH,CAAC,CACH,CAAC;YACF,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,qCAAqC,EAAE,qCAAqC,CAAC,CAAC,CAAC;QACvG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,kCAAkC,EAAE,KAAK,IAAI,EAAE;YAChD,SAAS,CAAC,gBAAgB,CAAC,WAAW,EAAE,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;YAEzD,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,4CAA4C,EAAE,KAAK,IAAI,EAAE;YAC1D,MAAM,eAAe,GAAG,EAAE,CAAC,KAAK,CAAC,MAAM,EAAE,cAAc,CAAC,CAAC;YACzD,SAAS,CAAC,cAAc,CAAC,IAAI,KAAK,CAAC,eAAe,CAAC,CAAC,CAAC;YAErD,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YACzB,MAAM,CAAC,eAAe,CAAC,CAAC,oBAAoB,EAAE,CAAC;QACjD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,4DAA4D,EAAE,KAAK,IAAI,EAAE;YAC1E,SAAS,CAAC,gBAAgB,CAAC,qCAAqC,EAAE;gBAChE,OAAO,EAAE,EAAE,cAAc,EAAE,0BAA0B,EAAE;aACxD,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,8CAA8C,EAAE,KAAK,IAAI,EAAE;YAC5D,SAAS,CAAC,gBAAgB,CAAC,EAAE,EAAE;gBAC7B,OAAO,EAAE,EAAE,cAAc,EAAE,YAAY,EAAE;aAC1C,CAAC,CAAC;YAEH,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,+BAA+B,CAAC,CAAC;YAE5E,MAAM,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAC3B,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
@@ -1,8 +1,10 @@
1
- import { contentExtractors } from './content-extractors.js';
1
+ import { StorybookExtractor } from './storybook-extractor.js';
2
+ import { GitHubPagesExtractor } from './github-pages-extractor.js';
3
+ import { DefaultExtractor } from './default-extractor.js';
2
4
  export const siteRules = [
3
5
  {
4
6
  type: 'storybook',
5
- extractor: contentExtractors.storybook,
7
+ extractor: new StorybookExtractor(),
6
8
  detect: async (page) => {
7
9
  return page.evaluate(() => {
8
10
  return !!(document.querySelector('#storybook-root, .sbdocs, [data-nodetype="root"]') ||
@@ -88,7 +90,7 @@ export const siteRules = [
88
90
  },
89
91
  {
90
92
  type: 'github',
91
- extractor: contentExtractors.github,
93
+ extractor: new GitHubPagesExtractor(),
92
94
  detect: async (page) => {
93
95
  return page.evaluate(() => {
94
96
  return ((window.location.hostname === 'github.io' || window.location.hostname.endsWith('.github.io')) &&
@@ -98,7 +100,7 @@ export const siteRules = [
98
100
  },
99
101
  {
100
102
  type: 'default',
101
- extractor: contentExtractors.default,
103
+ extractor: new DefaultExtractor(),
102
104
  detect: async () => true,
103
105
  },
104
106
  ];
@@ -1 +1 @@
1
- {"version":3,"file":"site-rules.js","sourceRoot":"","sources":["../../src/crawler/site-rules.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAU5D,MAAM,CAAC,MAAM,SAAS,GAAwB;IAC5C;QACE,IAAI,EAAE,WAAW;QACjB,SAAS,EAAE,iBAAiB,CAAC,SAAS;QACtC,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE;YACrB,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACxB,OAAO,CAAC,CAAC,CACP,QAAQ,CAAC,aAAa,CAAC,kDAAkD,CAAC;oBAC1E,QAAQ,CAAC,aAAa,CAAC,gCAAgC,CAAC;oBACxD,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,aAAa,CAAC;oBACzC,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,cAAc,CAAC;oBACzC,MAA4D,CAAC,wBAAwB,CACvF,CAAC;YACJ,CAAC,CAAC,CAAC;QACL,CAAC;QACD,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE;YAC3B,MAAM,OAAO,CAAC,GAAG,CAAC;gBAChB,IAAI,CAAC,gBAAgB,CAAC,aAAa,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,0CAA0C,CAAC,CAAC;gBAC1H,IAAI;qBACD,eAAe,CAAC,6DAA6D,EAAE;oBAC9E,OAAO,EAAE,IAAI;iBACd,CAAC;qBACD,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,yCAAyC,CAAC,CAAC;aACrE,CAAC,CAAC;YAEH,+BAA+B;YAC/B,MAAM,IAAI,CAAC,eAAe,CAAC,oBAAoB,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC,CAAC;YAE/G,mCAAmC;YACnC,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,kCAAkC,CAAC,CAAC,CAAC;gBAC1F,OAAO,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,CAAE,MAA4B,CAAC,KAAK,EAAE,CAAC,CAAC;YACrE,CAAC,CAAC,CAAC;YAEH,qCAAqC;YACrC,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,+CAA+C;YAC/C,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,yBAAyB,CAAC,CAAC,CAAC;gBACvF,aAAa,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,CAAE,MAA4B,CAAC,KAAK,EAAE,CAAC,CAAC;YAC3E,CAAC,CAAC,CAAC;YAEH,0DAA0D;YAC1D,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;YAEhC,8DAA8D;YAC9D,MAAM,IAAI,CAAC,QAAQ,CAAC,KAAK,IAAI,EAAE;gBAC7B,0BAA0B;gBAC1B,MAAM,MAAM,GAAG,QAAQ,CAAC,aAAa,CAAC,QAAQ,CAAsB,CAAC;gBACrE,MAAM,UAAU,GAAG,MAAM,EAAE,eAAe,IAAI,QAAQ,CAAC;gBAEvD,uCAAuC;gBACvC,MAAM,eAAe,GAAG,UAAU,CAAC,aAAa,CAAC,mCAAmC,CAAC,CAAC;gBACtF,IAAI,eAAe,EAAE,CAAC;oBACpB,eAAe,CAAC,QAAQ,CAAC,CAAC,EAAE,eAAe,CAAC,YAAY,CAAC,CAAC;oBAC1D,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC;oBACzD,eAAe,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;gBACjC,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,uDAAuD;YACvD,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CACrF,GAAG,CAAC,WAAW,EAAE,WAAW,EAAE,CAAC,QAAQ,CAAC,WAAW,CAAC,CACrD,CAAC;gBACF,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,mBAAmB;YAChF,CAAC,CAAC,CAAC;YAEH,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,qDAAqD;YACrD,MAAM,IAAI,CAAC,QAAQ,CAAC,KAAK,IAAI,EAAE;gBAC7B,MAAM,MAAM,GAAG,QAAQ,CAAC,aAAa,CAAC,QAAQ,CAAsB,CAAC;gBACrE,MAAM,UAAU,GAAG,MAAM,EAAE,eAAe,IAAI,QAAQ,CAAC;gBAEvD,qCAAqC;gBACrC,MAAM,UAAU,GAAG,UAAU,CAAC,gBAAgB,CAC5C,sDAAsD,GAAG,mDAAmD,CAC7G,CAAC;gBACF,UAAU,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAE,GAAyB,CAAC,KAAK,EAAE,CAAC,CAAC;gBAEhE,+DAA+D;gBAC/D,MAAM,YAAY,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAC3E,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,WAAW,EAAE,QAAQ,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,WAAW,EAAE,QAAQ,CAAC,MAAM,CAAC,CAChF,CAAC;gBACF,KAAK,MAAM,GAAG,IAAI,YAAY,EAAE,CAAC;oBAC/B,GAAG,CAAC,KAAK,EAAE,CAAC;oBACZ,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC;gBAC3D,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,8CAA8C;YAC9C,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACzC,OAAO,QAAQ,CAAC,gBAAgB,CAAC,0EAA0E,CAAC,CAAC,MAAM,CAAC;YACtH,CAAC,CAAC,CAAC;YACH,GAAG,CAAC,KAAK,CAAC,SAAS,SAAS,gCAAgC,CAAC,CAAC;QAChE,CAAC;QACD,aAAa,EAAE;YACb,iBAAiB;YACjB,0BAA0B;YAC1B,2BAA2B;YAC3B,8BAA8B;YAC9B,2BAA2B;YAC3B,kBAAkB;SACnB;KACF;IACD;QACE,IAAI,EAAE,QAAQ;QACd,SAAS,EAAE,iBAAiB,CAAC,MAAM;QACnC,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE;YACrB,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACxB,OAAO,CACL,CAAC,MAAM,CAAC,QAAQ,CAAC,QAAQ,KAAK,WAAW,IAAI,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC;oBAC7F,QAAQ,CAAC,aAAa,CAAC,4CAA4C,CAAC,KAAK,IAAI,CAC9E,CAAC;YACJ,CAAC,CAAC,CAAC;QACL,CAAC;KACF;IACD;QACE,IAAI,EAAE,SAAS;QACf,SAAS,EAAE,iBAAiB,CAAC,OAAO;QACpC,MAAM,EAAE,KAAK,IAAI,EAAE,CAAC,IAAI;KACzB;CACF,CAAC"}
1
+ {"version":3,"file":"site-rules.js","sourceRoot":"","sources":["../../src/crawler/site-rules.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAC9D,OAAO,EAAE,oBAAoB,EAAE,MAAM,6BAA6B,CAAC;AACnE,OAAO,EAAE,gBAAgB,EAAE,MAAM,wBAAwB,CAAC;AAU1D,MAAM,CAAC,MAAM,SAAS,GAAwB;IAC5C;QACE,IAAI,EAAE,WAAW;QACjB,SAAS,EAAE,IAAI,kBAAkB,EAAE;QACnC,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE;YACrB,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACxB,OAAO,CAAC,CAAC,CACP,QAAQ,CAAC,aAAa,CAAC,kDAAkD,CAAC;oBAC1E,QAAQ,CAAC,aAAa,CAAC,gCAAgC,CAAC;oBACxD,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,aAAa,CAAC;oBACzC,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,cAAc,CAAC;oBACzC,MAA4D,CAAC,wBAAwB,CACvF,CAAC;YACJ,CAAC,CAAC,CAAC;QACL,CAAC;QACD,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE;YAC3B,MAAM,OAAO,CAAC,GAAG,CAAC;gBAChB,IAAI,CAAC,gBAAgB,CAAC,aAAa,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,0CAA0C,CAAC,CAAC;gBAC1H,IAAI;qBACD,eAAe,CAAC,6DAA6D,EAAE;oBAC9E,OAAO,EAAE,IAAI;iBACd,CAAC;qBACD,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,yCAAyC,CAAC,CAAC;aACrE,CAAC,CAAC;YAEH,+BAA+B;YAC/B,MAAM,IAAI,CAAC,eAAe,CAAC,oBAAoB,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC,CAAC;YAE/G,mCAAmC;YACnC,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,kCAAkC,CAAC,CAAC,CAAC;gBAC1F,OAAO,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,CAAE,MAA4B,CAAC,KAAK,EAAE,CAAC,CAAC;YACrE,CAAC,CAAC,CAAC;YAEH,qCAAqC;YACrC,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,+CAA+C;YAC/C,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,yBAAyB,CAAC,CAAC,CAAC;gBACvF,aAAa,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,CAAE,MAA4B,CAAC,KAAK,EAAE,CAAC,CAAC;YAC3E,CAAC,CAAC,CAAC;YAEH,0DAA0D;YAC1D,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;YAEhC,8DAA8D;YAC9D,MAAM,IAAI,CAAC,QAAQ,CAAC,KAAK,IAAI,EAAE;gBAC7B,0BAA0B;gBAC1B,MAAM,MAAM,GAAG,QAAQ,CAAC,aAAa,CAAC,QAAQ,CAAsB,CAAC;gBACrE,MAAM,UAAU,GAAG,MAAM,EAAE,eAAe,IAAI,QAAQ,CAAC;gBAEvD,uCAAuC;gBACvC,MAAM,eAAe,GAAG,UAAU,CAAC,aAAa,CAAC,mCAAmC,CAAC,CAAC;gBACtF,IAAI,eAAe,EAAE,CAAC;oBACpB,eAAe,CAAC,QAAQ,CAAC,CAAC,EAAE,eAAe,CAAC,YAAY,CAAC,CAAC;oBAC1D,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC;oBACzD,eAAe,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;gBACjC,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,uDAAuD;YACvD,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACvB,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CACrF,GAAG,CAAC,WAAW,EAAE,WAAW,EAAE,CAAC,QAAQ,CAAC,WAAW,CAAC,CACrD,CAAC;gBACF,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,mBAAmB;YAChF,CAAC,CAAC,CAAC;YAEH,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,qDAAqD;YACrD,MAAM,IAAI,CAAC,QAAQ,CAAC,KAAK,IAAI,EAAE;gBAC7B,MAAM,MAAM,GAAG,QAAQ,CAAC,aAAa,CAAC,QAAQ,CAAsB,CAAC;gBACrE,MAAM,UAAU,GAAG,MAAM,EAAE,eAAe,IAAI,QAAQ,CAAC;gBAEvD,qCAAqC;gBACrC,MAAM,UAAU,GAAG,UAAU,CAAC,gBAAgB,CAC5C,sDAAsD,GAAG,mDAAmD,CAC7G,CAAC;gBACF,UAAU,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAE,GAAyB,CAAC,KAAK,EAAE,CAAC,CAAC;gBAEhE,+DAA+D;gBAC/D,MAAM,YAAY,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAC3E,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,WAAW,EAAE,QAAQ,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,WAAW,EAAE,QAAQ,CAAC,MAAM,CAAC,CAChF,CAAC;gBACF,KAAK,MAAM,GAAG,IAAI,YAAY,EAAE,CAAC;oBAC/B,GAAG,CAAC,KAAK,EAAE,CAAC;oBACZ,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC;gBAC3D,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,8CAA8C;YAC9C,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACzC,OAAO,QAAQ,CAAC,gBAAgB,CAAC,0EAA0E,CAAC,CAAC,MAAM,CAAC;YACtH,CAAC,CAAC,CAAC;YACH,GAAG,CAAC,KAAK,CAAC,SAAS,SAAS,gCAAgC,CAAC,CAAC;QAChE,CAAC;QACD,aAAa,EAAE;YACb,iBAAiB;YACjB,0BAA0B;YAC1B,2BAA2B;YAC3B,8BAA8B;YAC9B,2BAA2B;YAC3B,kBAAkB;SACnB;KACF;IACD;QACE,IAAI,EAAE,QAAQ;QACd,SAAS,EAAE,IAAI,oBAAoB,EAAE;QACrC,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE;YACrB,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE;gBACxB,OAAO,CACL,CAAC,MAAM,CAAC,QAAQ,CAAC,QAAQ,KAAK,WAAW,IAAI,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC;oBAC7F,QAAQ,CAAC,aAAa,CAAC,4CAA4C,CAAC,KAAK,IAAI,CAC9E,CAAC;YACJ,CAAC,CAAC,CAAC;QACL,CAAC;KACF;IACD;QACE,IAAI,EAAE,SAAS;QACf,SAAS,EAAE,IAAI,gBAAgB,EAAE;QACjC,MAAM,EAAE,KAAK,IAAI,EAAE,CAAC,IAAI;KACzB;CACF,CAAC"}