@crawlee/utils 3.9.3-beta.5 → 3.9.3-beta.50

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/index.d.ts CHANGED
@@ -7,7 +7,6 @@ export * from './internals/memory-info';
7
7
  export * from './internals/debug';
8
8
  export * as social from './internals/social';
9
9
  export * from './internals/typedefs';
10
- export * from './internals/error_tracker';
11
10
  export * from './internals/open_graph_parser';
12
11
  export * from './internals/gotScraping';
13
12
  export * from './internals/robots';
package/index.d.ts.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,qBAAqB,CAAC;AACpC,cAAc,qBAAqB,CAAC;AACpC,cAAc,mBAAmB,CAAC;AAClC,cAAc,0BAA0B,CAAC;AACzC,cAAc,qBAAqB,CAAC;AACpC,cAAc,yBAAyB,CAAC;AACxC,cAAc,mBAAmB,CAAC;AAClC,OAAO,KAAK,MAAM,MAAM,oBAAoB,CAAC;AAC7C,cAAc,sBAAsB,CAAC;AACrC,cAAc,2BAA2B,CAAC;AAC1C,cAAc,+BAA+B,CAAC;AAC9C,cAAc,yBAAyB,CAAC;AACxC,cAAc,oBAAoB,CAAC;AACnC,cAAc,qBAAqB,CAAC;AAEpC,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,gBAAgB,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,qBAAqB,CAAC;AACpC,cAAc,qBAAqB,CAAC;AACpC,cAAc,mBAAmB,CAAC;AAClC,cAAc,0BAA0B,CAAC;AACzC,cAAc,qBAAqB,CAAC;AACpC,cAAc,yBAAyB,CAAC;AACxC,cAAc,mBAAmB,CAAC;AAClC,OAAO,KAAK,MAAM,MAAM,oBAAoB,CAAC;AAC7C,cAAc,sBAAsB,CAAC;AACrC,cAAc,+BAA+B,CAAC;AAC9C,cAAc,yBAAyB,CAAC;AACxC,cAAc,oBAAoB,CAAC;AACnC,cAAc,qBAAqB,CAAC;AAEpC,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,gBAAgB,CAAC"}
package/index.js CHANGED
@@ -11,7 +11,6 @@ tslib_1.__exportStar(require("./internals/memory-info"), exports);
11
11
  tslib_1.__exportStar(require("./internals/debug"), exports);
12
12
  exports.social = tslib_1.__importStar(require("./internals/social"));
13
13
  tslib_1.__exportStar(require("./internals/typedefs"), exports);
14
- tslib_1.__exportStar(require("./internals/error_tracker"), exports);
15
14
  tslib_1.__exportStar(require("./internals/open_graph_parser"), exports);
16
15
  tslib_1.__exportStar(require("./internals/gotScraping"), exports);
17
16
  tslib_1.__exportStar(require("./internals/robots"), exports);
package/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;;;AAAA,8DAAoC;AACpC,8DAAoC;AACpC,4DAAkC;AAClC,mEAAyC;AACzC,8DAAoC;AACpC,kEAAwC;AACxC,4DAAkC;AAClC,qEAA6C;AAC7C,+DAAqC;AACrC,oEAA0C;AAC1C,wEAA8C;AAC9C,kEAAwC;AACxC,6DAAmC;AACnC,8DAAoC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;;;AAAA,8DAAoC;AACpC,8DAAoC;AACpC,4DAAkC;AAClC,mEAAyC;AACzC,8DAAoC;AACpC,kEAAwC;AACxC,4DAAkC;AAClC,qEAA6C;AAC7C,+DAAqC;AACrC,wEAA8C;AAC9C,kEAAwC;AACxC,6DAAmC;AACnC,8DAAoC"}
package/index.mjs CHANGED
@@ -2,7 +2,6 @@ import mod from "./index.js";
2
2
 
3
3
  export default mod;
4
4
  export const CLOUDFLARE_RETRY_CSS_SELECTORS = mod.CLOUDFLARE_RETRY_CSS_SELECTORS;
5
- export const ErrorTracker = mod.ErrorTracker;
6
5
  export const RETRY_CSS_SELECTORS = mod.RETRY_CSS_SELECTORS;
7
6
  export const ROTATE_PROXY_ERRORS = mod.ROTATE_PROXY_ERRORS;
8
7
  export const RobotsFile = mod.RobotsFile;
@@ -85,7 +85,7 @@ class RobotsFile {
85
85
  * @param [userAgent] relevant user agent, default to `*`
86
86
  */
87
87
  isAllowed(url, userAgent = '*') {
88
- return this.robots.isAllowed(url, userAgent) ?? false;
88
+ return this.robots.isAllowed(url, userAgent) ?? true; // `undefined` means that there is no explicit rule for the requested URL - assume it's allowed
89
89
  }
90
90
  /**
91
91
  * Get URLs of sitemaps referenced in the robots file.
@@ -1 +1 @@
1
- {"version":3,"file":"robots.js","sourceRoot":"","sources":["../../src/internals/robots.ts"],"names":[],"mappings":";;;;AAGA,0EAAyC;AAEzC,+CAA4C;AAC5C,uCAAoC;AAEpC,IAAI,SAAgC,CAAC;AAErC;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAa,UAAU;IACnB,YACY,MAAgD,EAChD,QAAiB;QADzB;;;;mBAAQ,MAAM;WAA0C;QACxD;;;;mBAAQ,QAAQ;WAAS;IAC1B,CAAC;IAEJ;;;;OAIG;IACH,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAW,EAAE,QAAiB;QAC5C,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QACnC,aAAa,CAAC,QAAQ,GAAG,aAAa,CAAC;QACvC,aAAa,CAAC,MAAM,GAAG,EAAE,CAAC;QAE1B,OAAO,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,EAAE,QAAQ,CAAC,CAAC;IAC/D,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,IAAI,CAAC,GAAW,EAAE,OAAe,EAAE,QAAiB;QACvD,OAAO,IAAI,UAAU,CAAC,IAAA,uBAAY,EAAC,GAAG,EAAE,OAAO,CAAC,EAAE,QAAQ,CAAC,CAAC;IAChE,CAAC;IAES,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAW,EAAE,QAAiB;QACtD,IAAI,CAAC,SAAS,EAAE,CAAC;YACb,SAAS,GAAG,CAAC,MAAM,MAAM,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC;QACzD,CAAC;QAED,IAAI,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,IAAA,yBAAW,EAAC;gBAC/B,GAAG;gBACH,QAAQ;gBACR,MAAM,EAAE,KAAK;gBACb,YAAY,EAAE,MAAM;aACvB,CAAC,CAAC;YAEH,OAAO,IAAI,UAAU,CAAC,IAAA,uBAAY,EAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,QAAQ,CAAC,CAAC;QACjF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACT,IAAI,CAAC,YAAY,SAAS,IAAI,CAAC,CAAC,QAAQ,CAAC,UAAU,KAAK,GAAG,EAAE,CAAC;gBAC1D,OAAO,IAAI,UAAU,CAAC,EAAE,SAAS,KAAK,OAAO,IAAI,CAAC,CAAC,CAAC,EAAE,WAAW,KAAK,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;YACpG,CAAC;YACD,MAAM,CAAC,CAAC;QACZ,CAAC;IACL,CAAC;IAED;;;;OAIG;IACH,SAAS,CAAC,GAAW,EAAE,SAAS,GAAG,GAAG;QAClC,OAAO,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,GAAG,EAAE,SAAS,CAAC,IAAI,KAAK,CAAC;IAC1D,CAAC;IAED;;OAEG;IACH,WAAW;QACP,OAAO,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;IACrC,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,aAAa;QACf,OAAO,iBAAO,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC;IAClE,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,qBAAqB;QACvB,OAAO,CAAC,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC;IAC7C,CAAC;CACJ;AAhFD,gCAgFC"}
1
+ {"version":3,"file":"robots.js","sourceRoot":"","sources":["../../src/internals/robots.ts"],"names":[],"mappings":";;;;AAGA,0EAAyC;AAEzC,+CAA4C;AAC5C,uCAAoC;AAEpC,IAAI,SAAgC,CAAC;AAErC;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAa,UAAU;IACnB,YACY,MAAgD,EAChD,QAAiB;QADzB;;;;mBAAQ,MAAM;WAA0C;QACxD;;;;mBAAQ,QAAQ;WAAS;IAC1B,CAAC;IAEJ;;;;OAIG;IACH,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAW,EAAE,QAAiB;QAC5C,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QACnC,aAAa,CAAC,QAAQ,GAAG,aAAa,CAAC;QACvC,aAAa,CAAC,MAAM,GAAG,EAAE,CAAC;QAE1B,OAAO,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,EAAE,QAAQ,CAAC,CAAC;IAC/D,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,IAAI,CAAC,GAAW,EAAE,OAAe,EAAE,QAAiB;QACvD,OAAO,IAAI,UAAU,CAAC,IAAA,uBAAY,EAAC,GAAG,EAAE,OAAO,CAAC,EAAE,QAAQ,CAAC,CAAC;IAChE,CAAC;IAES,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAW,EAAE,QAAiB;QACtD,IAAI,CAAC,SAAS,EAAE,CAAC;YACb,SAAS,GAAG,CAAC,MAAM,MAAM,CAAC,cAAc,CAAC,CAAC,CAAC,SAAS,CAAC;QACzD,CAAC;QAED,IAAI,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,IAAA,yBAAW,EAAC;gBAC/B,GAAG;gBACH,QAAQ;gBACR,MAAM,EAAE,KAAK;gBACb,YAAY,EAAE,MAAM;aACvB,CAAC,CAAC;YAEH,OAAO,IAAI,UAAU,CAAC,IAAA,uBAAY,EAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,QAAQ,CAAC,CAAC;QACjF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACT,IAAI,CAAC,YAAY,SAAS,IAAI,CAAC,CAAC,QAAQ,CAAC,UAAU,KAAK,GAAG,EAAE,CAAC;gBAC1D,OAAO,IAAI,UAAU,CAAC,EAAE,SAAS,KAAK,OAAO,IAAI,CAAC,CAAC,CAAC,EAAE,WAAW,KAAK,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;YACpG,CAAC;YACD,MAAM,CAAC,CAAC;QACZ,CAAC;IACL,CAAC;IAED;;;;OAIG;IACH,SAAS,CAAC,GAAW,EAAE,SAAS,GAAG,GAAG;QAClC,OAAO,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,GAAG,EAAE,SAAS,CAAC,IAAI,IAAI,CAAC,CAAC,+FAA+F;IACzJ,CAAC;IAED;;OAEG;IACH,WAAW;QACP,OAAO,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;IACrC,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,aAAa;QACf,OAAO,iBAAO,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC;IAClE,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,qBAAqB;QACvB,OAAO,CAAC,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC;IAC7C,CAAC;CACJ;AAhFD,gCAgFC"}
@@ -1 +1 @@
1
- {"version":3,"file":"sitemap.d.ts","sourceRoot":"","sources":["../../src/internals/sitemap.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,KAAK,CAAC;AAGrC,cAAM,YAAY;IACd,WAAW,EAAE,MAAM,EAAE,CAAM;IAC3B,IAAI,EAAE,MAAM,EAAE,CAAM;IACpB,kBAAkB,EAAE,MAAM,EAAE,CAAM;IAClC,OAAO,CAAC,EAAE,cAAc,GAAG,QAAQ,CAAC;IACpC,GAAG,UAAS;IAEZ,YAAY;CAIf;AAyCD;;;;;;;;;;;GAWG;AACH,qBAAa,OAAO;IACJ,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE;gBAAd,IAAI,EAAE,MAAM,EAAE;IAEnC,SAAS,CAAC,MAAM,CAAC,eAAe,CAAC,YAAY,EAAE,YAAY,EAAE,KAAK,EAAE,MAAM,IAAI,EAAE,OAAO,EAAE,CAAC,KAAK,EAAE,KAAK,KAAK,IAAI,GAAG,SAAS;IA0C3H;;;;;OAKG;WACU,cAAc,CAAC,GAAG,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;IAe7E;;;;OAIG;WACU,IAAI,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;CAkDlF"}
1
+ {"version":3,"file":"sitemap.d.ts","sourceRoot":"","sources":["../../src/internals/sitemap.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,KAAK,CAAC;AAIrC,cAAM,YAAY;IACd,WAAW,EAAE,MAAM,EAAE,CAAM;IAC3B,IAAI,EAAE,MAAM,EAAE,CAAM;IACpB,kBAAkB,EAAE,MAAM,EAAE,CAAM;IAClC,OAAO,CAAC,EAAE,cAAc,GAAG,QAAQ,CAAC;IACpC,GAAG,UAAS;IAEZ,YAAY;CAIf;AAyCD;;;;;;;;;;;GAWG;AACH,qBAAa,OAAO;IACJ,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE;gBAAd,IAAI,EAAE,MAAM,EAAE;IAEnC,SAAS,CAAC,MAAM,CAAC,eAAe,CAAC,YAAY,EAAE,YAAY,EAAE,KAAK,EAAE,MAAM,IAAI,EAAE,OAAO,EAAE,CAAC,KAAK,EAAE,KAAK,KAAK,IAAI,GAAG,SAAS;IA0C3H;;;;;OAKG;WACU,cAAc,CAAC,GAAG,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;IAe7E;;;;OAIG;WACU,IAAI,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;CAyDlF"}
@@ -7,6 +7,7 @@ const node_string_decoder_1 = require("node:string_decoder");
7
7
  const node_zlib_1 = require("node:zlib");
8
8
  const log_1 = tslib_1.__importDefault(require("@apify/log"));
9
9
  const sax_1 = tslib_1.__importDefault(require("sax"));
10
+ const whatwg_mimetype_1 = tslib_1.__importDefault(require("whatwg-mimetype"));
10
11
  class ParsingState {
11
12
  constructor() {
12
13
  Object.defineProperty(this, "sitemapUrls", {
@@ -201,10 +202,17 @@ class Sitemap {
201
202
  }
202
203
  const parser = (() => {
203
204
  const contentType = sitemapStream.response.headers['content-type'];
204
- if (['text/xml', 'application/xml'].includes(contentType ?? '') || sitemapUrl.pathname.endsWith('.xml')) {
205
+ let mimeType;
206
+ try {
207
+ mimeType = new whatwg_mimetype_1.default(contentType ?? '');
208
+ }
209
+ catch (e) {
210
+ mimeType = null;
211
+ }
212
+ if (mimeType?.isXML() || sitemapUrl.pathname.endsWith('.xml')) {
205
213
  return Sitemap.createXmlParser(parsingState, () => resolve(undefined), reject);
206
214
  }
207
- if (contentType === 'text/plain' || sitemapUrl.pathname.endsWith('.txt')) {
215
+ if (mimeType?.essence === 'text/plain' || sitemapUrl.pathname.endsWith('.txt')) {
208
216
  return new SitemapTxtParser(parsingState, () => resolve(undefined));
209
217
  }
210
218
  throw new Error('Unsupported sitemap content type');
@@ -1 +1 @@
1
- {"version":3,"file":"sitemap.js","sourceRoot":"","sources":["../../src/internals/sitemap.ts"],"names":[],"mappings":";;;;AACA,6CAAuC;AACvC,6DAAoD;AACpD,yCAAyC;AAEzC,6DAA6B;AAE7B,sDAAsB;AAEtB,MAAM,YAAY;IAAlB;QACI;;;;mBAAwB,EAAE;WAAC;QAC3B;;;;mBAAiB,EAAE;WAAC;QACpB;;;;mBAA+B,EAAE;WAAC;QAClC;;;;;WAAoC;QACpC;;;;mBAAM,KAAK;WAAC;IAMhB,CAAC;IAJG,YAAY;QACR,IAAI,CAAC,OAAO,GAAG,SAAS,CAAC;QACzB,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC;IACrB,CAAC;CACJ;AAED,MAAM,gBAAiB,SAAQ,sBAAQ;IAInC,YAAoB,YAA0B,EAAU,KAAiB;QACrE,KAAK,EAAE,CAAC;QADA;;;;mBAAQ,YAAY;WAAc;QAAE;;;;mBAAQ,KAAK;WAAY;QAHjE;;;;mBAAyB,IAAI,mCAAa,CAAC,MAAM,CAAC;WAAC;QACnD;;;;mBAAiB,EAAE;WAAC;IAI5B,CAAC;IAEO,aAAa,CAAC,KAAa,EAAE,QAAiB;QAClD,IAAI,CAAC,MAAM,IAAI,KAAK,CAAC;QAErB,IAAI,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YACzC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM;iBACpB,KAAK,CAAC,IAAI,CAAC;iBACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;iBAC1B,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAEvC,IAAI,QAAQ,EAAE,CAAC;gBACX,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;gBACtC,IAAI,CAAC,MAAM,GAAG,EAAE,CAAC;YACrB,CAAC;iBAAM,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC1B,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnD,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC;YAChC,CAAC;QACL,CAAC;IACL,CAAC;IAEQ,MAAM,CAAC,KAAU,EAAE,SAAyB,EAAE,QAAoD;QACvG,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,KAAK,CAAC,CAAC;QACrD,QAAQ,EAAE,CAAC;IACf,CAAC;IAEQ,MAAM,CAAC,QAAoD;QAChE,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,IAAI,CAAC,CAAC;QAC7C,QAAQ,EAAE,CAAC;QACX,IAAI,CAAC,KAAK,EAAE,CAAC;IACjB,CAAC;CACJ;AAED;;;;;;;;;;;GAWG;AACH,MAAa,OAAO;IAChB,YAAqB,IAAc;QAAvB;;;;mBAAS,IAAI;WAAU;IAAG,CAAC;IAE7B,MAAM,CAAC,eAAe,CAAC,YAA0B,EAAE,KAAiB,EAAE,OAA+B;QAC3G,MAAM,MAAM,GAAG,aAAG,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAEtC,MAAM,CAAC,EAAE,CAAC,SAAS,EAAE,CAAC,IAAI,EAAE,EAAE;YAC1B,IAAI,IAAI,CAAC,IAAI,KAAK,KAAK,IAAI,YAAY,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;gBAC5D,YAAY,CAAC,GAAG,GAAG,IAAI,CAAC;YAC5B,CAAC;YACD,IAAI,IAAI,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBACzB,YAAY,CAAC,OAAO,GAAG,QAAQ,CAAC;YACpC,CAAC;YACD,IAAI,IAAI,CAAC,IAAI,KAAK,cAAc,EAAE,CAAC;gBAC/B,YAAY,CAAC,OAAO,GAAG,cAAc,CAAC;YAC1C,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,MAAM,CAAC,EAAE,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,EAAE;YAC3B,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;gBACjB,YAAY,CAAC,GAAG,GAAG,KAAK,CAAC;YAC7B,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,EAAE;YAC5B,IAAI,YAAY,CAAC,GAAG,EAAE,CAAC;gBACnB,IAAI,YAAY,CAAC,OAAO,KAAK,cAAc,EAAE,CAAC;oBAC1C,IAAI,CAAC,YAAY,CAAC,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;wBAClD,YAAY,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACxC,CAAC;gBACL,CAAC;gBACD,IAAI,YAAY,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;oBACpC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACjC,CAAC;YACL,CAAC;QACL,CAAC,CAAC;QACF,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAC1B,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;QAE3B,MAAM,CAAC,EAAE,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;QACxB,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QAE5B,OAAO,MAAM,CAAC;IAClB,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,GAAW,EAAE,QAAiB;QACtD,MAAM,WAAW,GAAa,EAAE,CAAC;QAEjC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QAChC,UAAU,CAAC,MAAM,GAAG,EAAE,CAAC;QAEvB,UAAU,CAAC,QAAQ,GAAG,cAAc,CAAC;QACrC,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;QAExC,UAAU,CAAC,QAAQ,GAAG,cAAc,CAAC;QACrC,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;QAExC,OAAO,OAAO,CAAC,IAAI,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;IAC/C,CAAC;IAED;;;;OAIG;IACH,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAuB,EAAE,QAAiB;QACxD,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,MAAM,CAAC,cAAc,CAAC,CAAC;QAErD,MAAM,YAAY,GAAG,IAAI,YAAY,EAAE,CAAC;QACxC,YAAY,CAAC,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAE/D,OAAO,YAAY,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,WAAW,CAAC,GAAG,EAAG,CAAC,CAAC;YAC5D,YAAY,CAAC,kBAAkB,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;YAC5D,YAAY,CAAC,YAAY,EAAE,CAAC;YAE5B,IAAI,CAAC;gBACD,MAAM,aAAa,GAAG,MAAM,IAAI,OAAO,CAAwC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;oBAC/F,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;oBACjF,OAAO,CAAC,EAAE,CAAC,UAAU,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC;oBAC/C,OAAO,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;gBAChC,CAAC,CAAC,CAAC;gBAEH,IAAI,aAAa,CAAC,QAAS,CAAC,UAAU,KAAK,GAAG,EAAE,CAAC;oBAC7C,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;wBAClC,IAAI,MAAM,GAAW,aAAa,CAAC;wBACnC,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;4BACtC,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,IAAA,wBAAY,GAAE,CAAC,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;4BACzD,UAAU,CAAC,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,UAAU,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;wBAC3F,CAAC;wBAED,MAAM,MAAM,GAAG,CAAC,GAAG,EAAE;4BACjB,MAAM,WAAW,GAAG,aAAa,CAAC,QAAS,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC;4BAEpE,IAAI,CAAC,UAAU,EAAE,iBAAiB,CAAC,CAAC,QAAQ,CAAC,WAAW,IAAI,EAAE,CAAC,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gCACtG,OAAO,OAAO,CAAC,eAAe,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC,CAAC;4BACnF,CAAC;4BAED,IAAI,WAAW,KAAK,YAAY,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gCACvE,OAAO,IAAI,gBAAgB,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC;4BACxE,CAAC;4BAED,MAAM,IAAI,KAAK,CAAC,kCAAkC,CAAC,CAAC;wBACxD,CAAC,CAAC,EAAE,CAAC;wBAEL,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;oBACxB,CAAC,CAAC,CAAC;gBACP,CAAC;YACL,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACT,aAAG,CAAC,OAAO,CAAC,8BAA8B,UAAU,EAAE,CAAC,CAAC;YAC5D,CAAC;QACL,CAAC;QAED,OAAO,IAAI,OAAO,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;IAC1C,CAAC;CACJ;AAzHD,0BAyHC"}
1
+ {"version":3,"file":"sitemap.js","sourceRoot":"","sources":["../../src/internals/sitemap.ts"],"names":[],"mappings":";;;;AACA,6CAAuC;AACvC,6DAAoD;AACpD,yCAAyC;AAEzC,6DAA6B;AAE7B,sDAAsB;AACtB,8EAAuC;AAEvC,MAAM,YAAY;IAAlB;QACI;;;;mBAAwB,EAAE;WAAC;QAC3B;;;;mBAAiB,EAAE;WAAC;QACpB;;;;mBAA+B,EAAE;WAAC;QAClC;;;;;WAAoC;QACpC;;;;mBAAM,KAAK;WAAC;IAMhB,CAAC;IAJG,YAAY;QACR,IAAI,CAAC,OAAO,GAAG,SAAS,CAAC;QACzB,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC;IACrB,CAAC;CACJ;AAED,MAAM,gBAAiB,SAAQ,sBAAQ;IAInC,YAAoB,YAA0B,EAAU,KAAiB;QACrE,KAAK,EAAE,CAAC;QADA;;;;mBAAQ,YAAY;WAAc;QAAE;;;;mBAAQ,KAAK;WAAY;QAHjE;;;;mBAAyB,IAAI,mCAAa,CAAC,MAAM,CAAC;WAAC;QACnD;;;;mBAAiB,EAAE;WAAC;IAI5B,CAAC;IAEO,aAAa,CAAC,KAAa,EAAE,QAAiB;QAClD,IAAI,CAAC,MAAM,IAAI,KAAK,CAAC;QAErB,IAAI,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YACzC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM;iBACpB,KAAK,CAAC,IAAI,CAAC;iBACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;iBAC1B,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAEvC,IAAI,QAAQ,EAAE,CAAC;gBACX,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;gBACtC,IAAI,CAAC,MAAM,GAAG,EAAE,CAAC;YACrB,CAAC;iBAAM,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC1B,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnD,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC;YAChC,CAAC;QACL,CAAC;IACL,CAAC;IAEQ,MAAM,CAAC,KAAU,EAAE,SAAyB,EAAE,QAAoD;QACvG,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,KAAK,CAAC,CAAC;QACrD,QAAQ,EAAE,CAAC;IACf,CAAC;IAEQ,MAAM,CAAC,QAAoD;QAChE,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,IAAI,CAAC,CAAC;QAC7C,QAAQ,EAAE,CAAC;QACX,IAAI,CAAC,KAAK,EAAE,CAAC;IACjB,CAAC;CACJ;AAED;;;;;;;;;;;GAWG;AACH,MAAa,OAAO;IAChB,YAAqB,IAAc;QAAvB;;;;mBAAS,IAAI;WAAU;IAAG,CAAC;IAE7B,MAAM,CAAC,eAAe,CAAC,YAA0B,EAAE,KAAiB,EAAE,OAA+B;QAC3G,MAAM,MAAM,GAAG,aAAG,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAEtC,MAAM,CAAC,EAAE,CAAC,SAAS,EAAE,CAAC,IAAI,EAAE,EAAE;YAC1B,IAAI,IAAI,CAAC,IAAI,KAAK,KAAK,IAAI,YAAY,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;gBAC5D,YAAY,CAAC,GAAG,GAAG,IAAI,CAAC;YAC5B,CAAC;YACD,IAAI,IAAI,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBACzB,YAAY,CAAC,OAAO,GAAG,QAAQ,CAAC;YACpC,CAAC;YACD,IAAI,IAAI,CAAC,IAAI,KAAK,cAAc,EAAE,CAAC;gBAC/B,YAAY,CAAC,OAAO,GAAG,cAAc,CAAC;YAC1C,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,MAAM,CAAC,EAAE,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,EAAE;YAC3B,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;gBACjB,YAAY,CAAC,GAAG,GAAG,KAAK,CAAC;YAC7B,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,EAAE;YAC5B,IAAI,YAAY,CAAC,GAAG,EAAE,CAAC;gBACnB,IAAI,YAAY,CAAC,OAAO,KAAK,cAAc,EAAE,CAAC;oBAC1C,IAAI,CAAC,YAAY,CAAC,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;wBAClD,YAAY,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACxC,CAAC;gBACL,CAAC;gBACD,IAAI,YAAY,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;oBACpC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACjC,CAAC;YACL,CAAC;QACL,CAAC,CAAC;QACF,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAC1B,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;QAE3B,MAAM,CAAC,EAAE,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;QACxB,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QAE5B,OAAO,MAAM,CAAC;IAClB,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,GAAW,EAAE,QAAiB;QACtD,MAAM,WAAW,GAAa,EAAE,CAAC;QAEjC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QAChC,UAAU,CAAC,MAAM,GAAG,EAAE,CAAC;QAEvB,UAAU,CAAC,QAAQ,GAAG,cAAc,CAAC;QACrC,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;QAExC,UAAU,CAAC,QAAQ,GAAG,cAAc,CAAC;QACrC,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;QAExC,OAAO,OAAO,CAAC,IAAI,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;IAC/C,CAAC;IAED;;;;OAIG;IACH,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAuB,EAAE,QAAiB;QACxD,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,MAAM,CAAC,cAAc,CAAC,CAAC;QAErD,MAAM,YAAY,GAAG,IAAI,YAAY,EAAE,CAAC;QACxC,YAAY,CAAC,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAE/D,OAAO,YAAY,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,WAAW,CAAC,GAAG,EAAG,CAAC,CAAC;YAC5D,YAAY,CAAC,kBAAkB,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;YAC5D,YAAY,CAAC,YAAY,EAAE,CAAC;YAE5B,IAAI,CAAC;gBACD,MAAM,aAAa,GAAG,MAAM,IAAI,OAAO,CAAwC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;oBAC/F,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;oBACjF,OAAO,CAAC,EAAE,CAAC,UAAU,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC;oBAC/C,OAAO,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;gBAChC,CAAC,CAAC,CAAC;gBAEH,IAAI,aAAa,CAAC,QAAS,CAAC,UAAU,KAAK,GAAG,EAAE,CAAC;oBAC7C,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;wBAClC,IAAI,MAAM,GAAW,aAAa,CAAC;wBACnC,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;4BACtC,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,IAAA,wBAAY,GAAE,CAAC,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;4BACzD,UAAU,CAAC,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,UAAU,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;wBAC3F,CAAC;wBAED,MAAM,MAAM,GAAG,CAAC,GAAG,EAAE;4BACjB,MAAM,WAAW,GAAG,aAAa,CAAC,QAAS,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC;4BACpE,IAAI,QAAyB,CAAC;4BAE9B,IAAI,CAAC;gCACD,QAAQ,GAAG,IAAI,yBAAQ,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;4BAC/C,CAAC;4BAAC,OAAO,CAAC,EAAE,CAAC;gCACT,QAAQ,GAAG,IAAI,CAAC;4BACpB,CAAC;4BAED,IAAI,QAAQ,EAAE,KAAK,EAAE,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gCAC5D,OAAO,OAAO,CAAC,eAAe,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC,CAAC;4BACnF,CAAC;4BAED,IAAI,QAAQ,EAAE,OAAO,KAAK,YAAY,IAAI,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gCAC7E,OAAO,IAAI,gBAAgB,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC;4BACxE,CAAC;4BAED,MAAM,IAAI,KAAK,CAAC,kCAAkC,CAAC,CAAC;wBACxD,CAAC,CAAC,EAAE,CAAC;wBAEL,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;oBACxB,CAAC,CAAC,CAAC;gBACP,CAAC;YACL,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACT,aAAG,CAAC,OAAO,CAAC,8BAA8B,UAAU,EAAE,CAAC,CAAC;YAC5D,CAAC;QACL,CAAC;QAED,OAAO,IAAI,OAAO,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;IAC1C,CAAC;CACJ;AAhID,0BAgIC"}
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@crawlee/utils",
3
- "version": "3.9.3-beta.5",
3
+ "version": "3.9.3-beta.50",
4
4
  "description": "A set of shared utilities that can be used by crawlers",
5
5
  "engines": {
6
6
  "node": ">=16.0.0"
@@ -49,14 +49,18 @@
49
49
  "dependencies": {
50
50
  "@apify/log": "^2.4.0",
51
51
  "@apify/ps-tree": "^1.2.0",
52
- "@crawlee/types": "3.9.3-beta.5",
52
+ "@crawlee/types": "3.9.3-beta.50",
53
53
  "@types/sax": "^1.2.7",
54
54
  "cheerio": "^1.0.0-rc.12",
55
55
  "got-scraping": "^4.0.3",
56
56
  "ow": "^0.28.1",
57
57
  "robots-parser": "^3.0.1",
58
58
  "sax": "^1.3.0",
59
- "tslib": "^2.4.0"
59
+ "tslib": "^2.4.0",
60
+ "whatwg-mimetype": "^4.0.0"
61
+ },
62
+ "devDependencies": {
63
+ "@types/whatwg-mimetype": "^3.0.2"
60
64
  },
61
65
  "lerna": {
62
66
  "command": {
@@ -65,5 +69,5 @@
65
69
  }
66
70
  }
67
71
  },
68
- "gitHead": "e5671d22a9e95b4d8a8fc838526a92ce8956bac2"
72
+ "gitHead": "b4d36bdb7cb3de644b5c286ff634b110cf9ab580"
69
73
  }