mcp-scraper 0.4.0 → 0.4.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (48) hide show
  1. package/dist/bin/api-server.cjs +934 -221
  2. package/dist/bin/api-server.cjs.map +1 -1
  3. package/dist/bin/api-server.js +2 -2
  4. package/dist/bin/mcp-scraper-cli.cjs +2 -1
  5. package/dist/bin/mcp-scraper-cli.cjs.map +1 -1
  6. package/dist/bin/mcp-scraper-cli.js +3 -3
  7. package/dist/bin/mcp-scraper-install.cjs +3 -2
  8. package/dist/bin/mcp-scraper-install.cjs.map +1 -1
  9. package/dist/bin/mcp-scraper-install.js +2 -2
  10. package/dist/bin/mcp-stdio-server.cjs +83 -22
  11. package/dist/bin/mcp-stdio-server.cjs.map +1 -1
  12. package/dist/bin/mcp-stdio-server.js +4 -4
  13. package/dist/bin/paa-harvest.cjs +1 -0
  14. package/dist/bin/paa-harvest.cjs.map +1 -1
  15. package/dist/bin/paa-harvest.js +2 -2
  16. package/dist/{chunk-6TDUY7CA.js → chunk-4MFYMUIF.js} +25 -3
  17. package/dist/chunk-4MFYMUIF.js.map +1 -0
  18. package/dist/{chunk-T3VZD2I4.js → chunk-5HNHYUJ5.js} +3 -2
  19. package/dist/chunk-5HNHYUJ5.js.map +1 -0
  20. package/dist/chunk-7HICT4GY.js +7 -0
  21. package/dist/chunk-7HICT4GY.js.map +1 -0
  22. package/dist/{chunk-NNEIXK5L.js → chunk-FUWZWKGO.js} +2 -2
  23. package/dist/{chunk-3NIPPJGP.js → chunk-GFY6XHZS.js} +80 -21
  24. package/dist/chunk-GFY6XHZS.js.map +1 -0
  25. package/dist/{chunk-HBC4C75S.js → chunk-NJK5BTUK.js} +2 -2
  26. package/dist/{chunk-JOEZJL4I.js → chunk-NVUBFCJK.js} +2 -2
  27. package/dist/{chunk-BP27CZ5Q.js → chunk-XGIPATLV.js} +2 -1
  28. package/dist/chunk-XGIPATLV.js.map +1 -0
  29. package/dist/index.cjs +1 -0
  30. package/dist/index.cjs.map +1 -1
  31. package/dist/index.js +2 -2
  32. package/dist/{server-Z4WKTE6F.js → server-6IYQQK6X.js} +793 -210
  33. package/dist/server-6IYQQK6X.js.map +1 -0
  34. package/dist/{site-extract-repository-2MQBAOX5.js → site-extract-repository-4NVO5KQA.js} +3 -3
  35. package/dist/{worker-RVZXIALX.js → worker-7XRZNLKB.js} +4 -4
  36. package/package.json +2 -1
  37. package/dist/chunk-3NIPPJGP.js.map +0 -1
  38. package/dist/chunk-6TDUY7CA.js.map +0 -1
  39. package/dist/chunk-BP27CZ5Q.js.map +0 -1
  40. package/dist/chunk-IXDYL7TE.js +0 -7
  41. package/dist/chunk-IXDYL7TE.js.map +0 -1
  42. package/dist/chunk-T3VZD2I4.js.map +0 -1
  43. package/dist/server-Z4WKTE6F.js.map +0 -1
  44. /package/dist/{chunk-NNEIXK5L.js.map → chunk-FUWZWKGO.js.map} +0 -0
  45. /package/dist/{chunk-HBC4C75S.js.map → chunk-NJK5BTUK.js.map} +0 -0
  46. /package/dist/{chunk-JOEZJL4I.js.map → chunk-NVUBFCJK.js.map} +0 -0
  47. /package/dist/{site-extract-repository-2MQBAOX5.js.map → site-extract-repository-4NVO5KQA.js.map} +0 -0
  48. /package/dist/{worker-RVZXIALX.js.map → worker-7XRZNLKB.js.map} +0 -0
@@ -9,7 +9,7 @@ import {
9
9
  DEFAULT_PROXY_MODE,
10
10
  HarvestOptionsSchema,
11
11
  RawPAAItemSchema
12
- } from "./chunk-BP27CZ5Q.js";
12
+ } from "./chunk-XGIPATLV.js";
13
13
  import {
14
14
  CaptchaError,
15
15
  ExtractionError,
@@ -3258,4 +3258,4 @@ export {
3258
3258
  resolveKernelProxyId,
3259
3259
  harvest
3260
3260
  };
3261
- //# sourceMappingURL=chunk-HBC4C75S.js.map
3261
+ //# sourceMappingURL=chunk-NJK5BTUK.js.map
@@ -1,6 +1,6 @@
1
1
  import {
2
2
  LedgerOperation
3
- } from "./chunk-6TDUY7CA.js";
3
+ } from "./chunk-4MFYMUIF.js";
4
4
  import {
5
5
  sanitizeVendorName
6
6
  } from "./chunk-M2S27J6Z.js";
@@ -127,4 +127,4 @@ export {
127
127
  failExtractJob,
128
128
  settleExtractJob
129
129
  };
130
- //# sourceMappingURL=chunk-JOEZJL4I.js.map
130
+ //# sourceMappingURL=chunk-NVUBFCJK.js.map
@@ -32,6 +32,7 @@ var MapsPlaceOptionsSchema = z.object({
32
32
  hl: z.string().length(2).default("en"),
33
33
  includeReviews: z.boolean().default(false),
34
34
  maxReviews: z.number().int().min(1).max(500).default(50),
35
+ includeServices: z.boolean().default(false),
35
36
  kernelApiKey: z.string().optional(),
36
37
  kernelProxyId: z.string().optional(),
37
38
  headless: z.boolean().default(true)
@@ -109,4 +110,4 @@ export {
109
110
  RawMapsReviewStatsSchema,
110
111
  RawMapsAboutAttributeSchema
111
112
  };
112
- //# sourceMappingURL=chunk-BP27CZ5Q.js.map
113
+ //# sourceMappingURL=chunk-XGIPATLV.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["../src/schemas.ts"],"sourcesContent":["import { z } from 'zod'\n\nexport const DEFAULT_PROXY_MODE = 'none' as const\nexport const DEFAULT_MAPS_PROXY_MODE = 'location' as const\n\nexport const HarvestOptionsSchema = z.object({\n query: z.string().min(1),\n location: z.string().optional(),\n gl: z.string().length(2).default('us'),\n hl: z.string().length(2).default('en'),\n device: z.enum(['desktop', 'mobile']).default('desktop'),\n proxyMode: z.enum(['location', 'configured', 'none']).default(DEFAULT_PROXY_MODE),\n proxyZip: z.string().regex(/^\\d{5}$/).optional(),\n debug: z.boolean().default(false),\n depth: z.number().int().min(1).max(30).default(3),\n maxQuestions: z.number().int().min(1).max(1000).default(100),\n headless: z.boolean().default(false),\n profileDir: z.string().optional(),\n proxy: z.string().url().optional(),\n kernelApiKey: z.string().optional(),\n kernelProxyId: z.string().optional(),\n kernelProxyResolution: z.unknown().optional(),\n outputDir: z.string().default('./paa-output'),\n format: z.enum(['json', 'csv', 'both']).default('both'),\n serpOnly: z.boolean().default(false),\n pages: z.number().int().min(1).max(2).default(1),\n softDeadlineMs: z.number().optional(),\n})\n\nexport const MapsPlaceOptionsSchema = z.object({\n businessName: z.string().min(1),\n location: z.string().min(1),\n gl: z.string().length(2).default('us'),\n hl: z.string().length(2).default('en'),\n includeReviews: z.boolean().default(false),\n maxReviews: z.number().int().min(1).max(500).default(50),\n includeServices: z.boolean().default(false),\n kernelApiKey: z.string().optional(),\n kernelProxyId: z.string().optional(),\n headless: z.boolean().default(true),\n})\n\nexport const MapsSearchOptionsSchema = z.object({\n query: z.string().min(1),\n location: z.string().optional(),\n gl: z.string().length(2).default('us'),\n hl: z.string().length(2).default('en'),\n maxResults: z.number().int().min(1).max(50).default(10),\n proxyMode: z.enum(['location', 'configured', 'none']).default(DEFAULT_MAPS_PROXY_MODE),\n proxyZip: z.string().regex(/^\\d{5}$/).optional(),\n debug: z.boolean().default(false),\n kernelApiKey: z.string().optional(),\n kernelProxyId: z.string().optional(),\n kernelProxyResolution: z.unknown().optional(),\n headless: z.boolean().default(true),\n})\n\nexport const RawPAAItemSchema = z.object({\n question: z.string().min(1),\n answer: z.string().optional(),\n sourceTitle: z.string().optional(),\n sourceSite: z.string().optional(),\n sourceCite: z.string().optional(),\n})\n\nexport const RawMapsOverviewSchema = z.object({\n name: z.string().nullable(),\n rating: z.string().nullable(),\n reviewCount: z.string().nullable(),\n category: z.string().nullable(),\n address: z.string().nullable(),\n hoursSummary: z.string().nullable(),\n phone: z.string().nullable(),\n phoneDisplay: z.string().nullable(),\n website: z.string().nullable(),\n plusCode: z.string().nullable(),\n bookingUrl: z.string().nullable(),\n})\n\nexport const RawMapsHoursRowSchema = z.object({\n day: z.string(),\n hours: z.string(),\n})\n\nexport const RawMapsReviewStatsSchema = z.object({\n reviewHistogram: z.array(z.object({\n stars: z.number(),\n count: z.string(),\n })),\n reviewTopics: z.array(z.object({\n label: z.string(),\n count: z.string(),\n })),\n})\n\nexport const RawMapsReviewCardSchema = z.object({\n reviewId: z.string(),\n author: z.string().nullable(),\n stars: z.string().nullable(),\n date: z.string().nullable(),\n text: z.string().nullable(),\n ownerResponse: z.string().nullable(),\n})\n\nexport const RawMapsAboutAttributeSchema = z.object({\n section: z.string(),\n attribute: z.string(),\n})\n"],"mappings":";AAAA,SAAS,SAAS;AAEX,IAAM,qBAAqB;AAC3B,IAAM,0BAA0B;AAEhC,IAAM,uBAAuB,EAAE,OAAO;AAAA,EAC3C,OAAc,EAAE,OAAO,EAAE,IAAI,CAAC;AAAA,EAC9B,UAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,IAAc,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAC/C,IAAc,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAC/C,QAAc,EAAE,KAAK,CAAC,WAAW,QAAQ,CAAC,EAAE,QAAQ,SAAS;AAAA,EAC7D,WAAc,EAAE,KAAK,CAAC,YAAY,cAAc,MAAM,CAAC,EAAE,QAAQ,kBAAkB;AAAA,EACnF,UAAc,EAAE,OAAO,EAAE,MAAM,SAAS,EAAE,SAAS;AAAA,EACnD,OAAc,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EACvC,OAAc,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,EAAE,QAAQ,CAAC;AAAA,EACvD,cAAc,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,GAAI,EAAE,QAAQ,GAAG;AAAA,EAC3D,UAAc,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EACvC,YAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,OAAc,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS;AAAA,EACxC,cAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,eAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,uBAAuB,EAAE,QAAQ,EAAE,SAAS;AAAA,EAC5C,WAAc,EAAE,OAAO,EAAE,QAAQ,cAAc;AAAA,EAC/C,QAAc,EAAE,KAAK,CAAC,QAAQ,OAAO,MAAM,CAAC,EAAE,QAAQ,MAAM;AAAA,EAC5D,UAAc,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EACvC,OAAc,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,CAAC,EAAE,QAAQ,CAAC;AAAA,EACtD,gBAAgB,EAAE,OAAO,EAAE,SAAS;AACtC,CAAC;AAEM,IAAM,yBAAyB,EAAE,OAAO;AAAA,EAC7C,cAAe,EAAE,OAAO,EAAE,IAAI,CAAC;AAAA,EAC/B,UAAe,EAAE,OAAO,EAAE,IAAI,CAAC;AAAA,EAC/B,IAAe,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAChD,IAAe,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAChD,gBAAgB,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EACzC,YAAe,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,GAAG,EAAE,QAAQ,EAAE;AAAA,EAC1D,iBAAiB,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EAC1C,cAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,eAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,UAAe,EAAE,QAAQ,EAAE,QAAQ,IAAI;AACzC,CAAC;AAEM,IAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,OAAe,EAAE,OAAO,EAAE,IAAI,CAAC;AAAA,EAC/B,UAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,IAAe,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAChD,IAAe,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,QAAQ,IAAI;AAAA,EAChD,YAAe,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,EAAE,QAAQ,EAAE;AAAA,EACzD,WAAe,EAAE,KAAK,CAAC,YAAY,cAAc,MAAM,CAAC,EAAE,QAAQ,uBAAuB;AAAA,EACzF,UAAe,EAAE,OAAO,EAAE,MAAM,SAAS,EAAE,SAAS;AAAA,EACpD,OAAe,EAAE,QAAQ,EAAE,QAAQ,KAAK;AAAA,EACxC,cAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,eAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,uBAAuB,EAAE,QAAQ,EAAE,SAAS;AAAA,EAC5C,UAAe,EAAE,QAAQ,EAAE,QAAQ,IAAI;AACzC,CAAC;AAEM,IAAM,mBAAmB,EAAE,OAAO;AAAA,EACvC,UAAa,EAAE,OAAO,EAAE,IAAI,CAAC;AAAA,EAC7B,QAAa,EAAE,OAAO,EAAE,SAAS;AAAA,EACjC,aAAa,EAAE,OAAO,EAAE,SAAS;AAAA,EACjC,YAAa,EAAE,OAAO,EAAE,SAAS;AAAA,EACjC,YAAa,EAAE,OAAO,EAAE,SAAS;AACnC,CAAC;AAEM,IAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,MAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,QAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,aAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,UAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,SAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,cAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,OAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,cAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,SAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,UAAc,EAAE,OAAO,EAAE,SAAS;AAAA,EAClC,YAAc,EAAE,OAAO,EAAE,SAAS;AACpC,CAAC;AAEM,IAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,KAAO,EAAE,OAAO;AAAA,EAChB,OAAO,EAAE,OAAO;AAClB,CAAC;AAEM,IAAM,2BAA2B,EAAE,OAAO;AAAA,EAC/C,iBAAiB,EAAE,MAAM,EAAE,OAAO;AAAA,IAChC,OAAO,EAAE,OAAO;AAAA,IAChB,OAAO,EAAE,OAAO;AAAA,EAClB,CAAC,CAAC;AAAA,EACF,cAAc,EAAE,MAAM,EAAE,OAAO;AAAA,IAC7B,OAAO,EAAE,OAAO;AAAA,IAChB,OAAO,EAAE,OAAO;AAAA,EAClB,CAAC,CAAC;AACJ,CAAC;AAEM,IAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,UAAe,EAAE,OAAO;AAAA,EACxB,QAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,OAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,MAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,MAAe,EAAE,OAAO,EAAE,SAAS;AAAA,EACnC,eAAe,EAAE,OAAO,EAAE,SAAS;AACrC,CAAC;AAEM,IAAM,8BAA8B,EAAE,OAAO;AAAA,EAClD,SAAW,EAAE,OAAO;AAAA,EACpB,WAAW,EAAE,OAAO;AACtB,CAAC;","names":[]}
package/dist/index.cjs CHANGED
@@ -70,6 +70,7 @@ var MapsPlaceOptionsSchema = import_zod.z.object({
70
70
  hl: import_zod.z.string().length(2).default("en"),
71
71
  includeReviews: import_zod.z.boolean().default(false),
72
72
  maxReviews: import_zod.z.number().int().min(1).max(500).default(50),
73
+ includeServices: import_zod.z.boolean().default(false),
73
74
  kernelApiKey: import_zod.z.string().optional(),
74
75
  kernelProxyId: import_zod.z.string().optional(),
75
76
  headless: import_zod.z.boolean().default(true)