mcp-scraper 0.69.1 → 0.72.1

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 (45) hide show
  1. package/CHANGELOG.md +63 -2
  2. package/README.md +6 -2
  3. package/dist/{analytics-repository-62CAZAIP.js → analytics-repository-6YLRSWYO.js} +3 -2
  4. package/dist/bin/api-server.cjs +18989 -13102
  5. package/dist/bin/api-server.js +4 -4
  6. package/dist/bin/mcp-scraper-cli.cjs +1 -1
  7. package/dist/bin/mcp-scraper-cli.js +1 -1
  8. package/dist/bin/mcp-scraper-install.cjs +3 -3
  9. package/dist/bin/mcp-scraper-install.js +2 -2
  10. package/dist/bin/mcp-stdio-server.cjs +10493 -9135
  11. package/dist/bin/mcp-stdio-server.js +9 -7
  12. package/dist/bin/paa-harvest.cjs +308 -41
  13. package/dist/bin/paa-harvest.js +4 -3
  14. package/dist/{chunk-RODSHBN7.js → chunk-57F5PKS5.js} +1 -1
  15. package/dist/chunk-7RQULQF3.js +263 -0
  16. package/dist/{chunk-UL4ZKAWZ.js → chunk-A66DGFOU.js} +1 -1
  17. package/dist/{chunk-ES25GP6C.js → chunk-CXY5WV45.js} +24 -34
  18. package/dist/{chunk-Y46YNQMM.js → chunk-F5TK4KMT.js} +1 -1
  19. package/dist/{chunk-62LEXS5O.js → chunk-F6MUGMRN.js} +306 -44
  20. package/dist/{chunk-5RFQGIYC.js → chunk-FVL4GUTP.js} +2 -2
  21. package/dist/{chunk-OZIUFN6B.js → chunk-I45JV4EU.js} +2 -2
  22. package/dist/{chunk-2LXTOAKS.js → chunk-KIPIBPUB.js} +9 -1
  23. package/dist/chunk-MB72PA6S.js +2003 -0
  24. package/dist/chunk-OM7HVEJ3.js +26 -0
  25. package/dist/chunk-PJEEKOUM.js +404 -0
  26. package/dist/{chunk-RX2QAHML.js → chunk-RAFQEPJ4.js} +2 -2
  27. package/dist/{chunk-KO5CK5ZT.js → chunk-SR7GSLEA.js} +202 -441
  28. package/dist/chunk-T3MZISOF.js +240 -0
  29. package/dist/{chunk-65FTMB7G.js → chunk-WEFPBAAG.js} +180 -9
  30. package/dist/{chunk-LN6N3YLK.js → chunk-WFQ2E4WE.js} +1 -1
  31. package/dist/{chunk-2LMMZUU5.js → chunk-XYRD4KLT.js} +3173 -2019
  32. package/dist/{db-ZGZAKYBW.js → db-566MOHHD.js} +7 -1
  33. package/dist/{extract-bundle-SXSJCPQI.js → extract-bundle-337ZDIIQ.js} +4 -4
  34. package/dist/gmail-service-E276SBJU.js +25 -0
  35. package/dist/index.cjs +322 -49
  36. package/dist/index.d.cts +15 -2
  37. package/dist/index.d.ts +15 -2
  38. package/dist/index.js +4 -3
  39. package/dist/{lead-list-enrichment-repository-CXLVIV3R.js → lead-list-enrichment-repository-WWEMJ7EN.js} +2 -2
  40. package/dist/{location-data-repository-G72M7DTN.js → location-data-repository-H4B2BWGY.js} +2 -2
  41. package/dist/{server-VDJBV4V4.js → server-QY2AZ2OW.js} +8377 -6637
  42. package/dist/{site-extract-repository-RLB3TFJ3.js → site-extract-repository-ACNAVNDN.js} +3 -3
  43. package/dist/{worker-KHSJG365.js → worker-JQORPCFQ.js} +9 -7
  44. package/package.json +1 -1
  45. package/dist/chunk-2L4C4DAZ.js +0 -103
@@ -17,6 +17,7 @@ import {
17
17
  claimStripeEvent,
18
18
  completeJob,
19
19
  completeKpoJob,
20
+ completeRunningJob,
20
21
  completeStripeEvent,
21
22
  completeWorkflowRunRecord,
22
23
  consumeCode,
@@ -77,6 +78,7 @@ import {
77
78
  listActiveUserIds,
78
79
  listDueWorkflowSchedules,
79
80
  listHarvestAttempts,
81
+ listHarvestAttemptsPendingProviderTelemetry,
80
82
  listInboxMessages,
81
83
  listJobs,
82
84
  listKpoJobs,
@@ -123,10 +125,11 @@ import {
123
125
  softDeleteInboxMessage,
124
126
  startHarvestAttempt,
125
127
  stripeEventAlreadyProcessed,
128
+ updateHarvestAttemptProviderTelemetry,
126
129
  updateKpoJobState,
127
130
  upsertPageSnapshot,
128
131
  verifyPassword
129
- } from "./chunk-65FTMB7G.js";
132
+ } from "./chunk-WEFPBAAG.js";
130
133
  export {
131
134
  CORE_SCHEMA_VERSION,
132
135
  CREDIT_LOTS_EPOCH,
@@ -146,6 +149,7 @@ export {
146
149
  claimStripeEvent,
147
150
  completeJob,
148
151
  completeKpoJob,
152
+ completeRunningJob,
149
153
  completeStripeEvent,
150
154
  completeWorkflowRunRecord,
151
155
  consumeCode,
@@ -206,6 +210,7 @@ export {
206
210
  listActiveUserIds,
207
211
  listDueWorkflowSchedules,
208
212
  listHarvestAttempts,
213
+ listHarvestAttemptsPendingProviderTelemetry,
209
214
  listInboxMessages,
210
215
  listJobs,
211
216
  listKpoJobs,
@@ -252,6 +257,7 @@ export {
252
257
  softDeleteInboxMessage,
253
258
  startHarvestAttempt,
254
259
  stripeEventAlreadyProcessed,
260
+ updateHarvestAttemptProviderTelemetry,
255
261
  updateKpoJobState,
256
262
  upsertPageSnapshot,
257
263
  verifyPassword
@@ -4,7 +4,7 @@ import {
4
4
  downloadAsset,
5
5
  embedCommonsTexts,
6
6
  publicizeExtractionFailure
7
- } from "./chunk-LN6N3YLK.js";
7
+ } from "./chunk-WFQ2E4WE.js";
8
8
  import "./chunk-OZJMVCDK.js";
9
9
  import {
10
10
  computeIssues,
@@ -18,12 +18,12 @@ import {
18
18
  createSiteExtractContentReader,
19
19
  createSiteExtractImageArtifact,
20
20
  extractJobLimitInfo
21
- } from "./chunk-OZIUFN6B.js";
21
+ } from "./chunk-I45JV4EU.js";
22
22
  import "./chunk-DNM65UCK.js";
23
- import "./chunk-5RFQGIYC.js";
23
+ import "./chunk-FVL4GUTP.js";
24
24
  import {
25
25
  getDb
26
- } from "./chunk-65FTMB7G.js";
26
+ } from "./chunk-WEFPBAAG.js";
27
27
 
28
28
  // src/api/extract-bundle.ts
29
29
  import { createWriteStream, mkdirSync, readFileSync, readdirSync, rmSync, statSync, writeFileSync } from "fs";
@@ -0,0 +1,25 @@
1
+ import {
2
+ GMAIL_ATTACHMENT_REF_TTL_MS,
3
+ GMAIL_INLINE_BODY_BYTES,
4
+ GMAIL_MAIN_CONTRACT_HASH,
5
+ GMAIL_MAIN_CONTRACT_VERSION,
6
+ GmailService,
7
+ GmailServiceError,
8
+ normalizeGmailMessage,
9
+ parseGmailAddresses
10
+ } from "./chunk-MB72PA6S.js";
11
+ import "./chunk-PJEEKOUM.js";
12
+ import "./chunk-T3MZISOF.js";
13
+ import "./chunk-CXY5WV45.js";
14
+ import "./chunk-FVL4GUTP.js";
15
+ import "./chunk-WEFPBAAG.js";
16
+ export {
17
+ GMAIL_ATTACHMENT_REF_TTL_MS,
18
+ GMAIL_INLINE_BODY_BYTES,
19
+ GMAIL_MAIN_CONTRACT_HASH,
20
+ GMAIL_MAIN_CONTRACT_VERSION,
21
+ GmailService,
22
+ GmailServiceError,
23
+ normalizeGmailMessage,
24
+ parseGmailAddresses
25
+ };