@crawlee/cheerio 3.13.11-beta.3 → 3.13.11-beta.4

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.
@@ -1,5 +1,5 @@
1
1
  import type { IncomingMessage } from 'node:http';
2
- import type { Configuration, EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler, RequestProvider, RouterRoutes, SkippedRequestCallback } from '@crawlee/http';
2
+ import type { BasicCrawlingContext, Configuration, EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler, RequestProvider, RouterRoutes, SkippedRequestCallback } from '@crawlee/http';
3
3
  import { HttpCrawler } from '@crawlee/http';
4
4
  import type { Dictionary } from '@crawlee/types';
5
5
  import { type CheerioRoot, type RobotsTxtFile } from '@crawlee/utils';
@@ -150,9 +150,16 @@ interface EnqueueLinksInternalOptions {
150
150
  originalRequestUrl: string;
151
151
  finalRequestUrl?: string;
152
152
  }
153
+ interface BoundEnqueueLinksInternalOptions {
154
+ enqueueLinks: BasicCrawlingContext['enqueueLinks'];
155
+ options?: EnqueueLinksOptions;
156
+ $: cheerio.CheerioAPI | null;
157
+ originalRequestUrl: string;
158
+ finalRequestUrl?: string;
159
+ }
153
160
  /** @internal */
154
161
  // @ts-ignore optional peer dependency or compatibility with es2022
155
- export declare function cheerioCrawlerEnqueueLinks({ options, $, requestQueue, robotsTxtFile, onSkippedRequest, originalRequestUrl, finalRequestUrl, }: EnqueueLinksInternalOptions): Promise<import("@crawlee/types").BatchAddRequestsResult>;
162
+ export declare function cheerioCrawlerEnqueueLinks(options: EnqueueLinksInternalOptions | BoundEnqueueLinksInternalOptions): Promise<import("@crawlee/types").BatchAddRequestsResult>;
156
163
  /**
157
164
  * Creates new {@link Router} instance that works based on request labels.
158
165
  * This instance can then serve as a `requestHandler` of your {@link CheerioCrawler}.
@@ -1 +1 @@
1
- {"version":3,"file":"cheerio-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/cheerio-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAGjD,OAAO,KAAK,EACR,aAAa,EACb,mBAAmB,EACnB,YAAY,EACZ,sBAAsB,EACtB,kBAAkB,EAClB,2BAA2B,EAC3B,gBAAgB,EAChB,cAAc,EACd,eAAe,EACf,YAAY,EACZ,sBAAsB,EACzB,MAAM,eAAe,CAAC;AACvB,OAAO,EAAgB,WAAW,EAAkD,MAAM,eAAe,CAAC;AAC1G,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,KAAK,WAAW,EAA0B,KAAK,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE9F,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AAInC,MAAM,MAAM,mBAAmB,CAC3B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,YAAY,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE7D,MAAM,WAAW,qBAAqB,CAClC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,kBAAkB,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;CAAG;AAE3E,MAAM,MAAM,WAAW,CACnB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,gBAAgB,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEjE,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,2BAA2B,CAAC,QAAQ,EAAE,QAAQ,EAAE,cAAc,CAAC;IACrE;;;OAGG;IACH,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC;IAEtB;;;;;;;;;;;OAWG;IACH,eAAe,CAAC,QAAQ,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAErE;;;;;;;;;;;;;OAaG;IACH,gBAAgB,CAAC,QAAQ,CAAC,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;CACjF;AAED,MAAM,MAAM,qBAAqB,CAC7B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,cAAc,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE/D;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4EG;AACH,qBAAa,cAAe,SAAQ,WAAW,CAAC,sBAAsB,CAAC;IACnE;;OAEG;gBAES,OAAO,CAAC,EAAE,qBAAqB,EAAE,MAAM,CAAC,EAAE,aAAa;cAI1C,UAAU,CAC/B,QAAQ,EAAE,eAAe,EACzB,KAAK,EAAE,OAAO,EACd,eAAe,EAAE,sBAAsB;;;;wCAiBG,mBAAmB;;cAejD,eAAe,CAAC,QAAQ,EAAE,eAAe,EAAE,KAAK,EAAE,OAAO;cAehD,kBAAkB,CAAC,OAAO,EAAE,sBAAsB;CAgB9E;AAED,UAAU,2BAA2B;IACjC,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,CAAC,EAAE,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC;IAC7B,YAAY,EAAE,eAAe,CAAC;IAC9B,aAAa,CAAC,EAAE,aAAa,CAAC;IAC9B,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IAC1C,kBAAkB,EAAE,MAAM,CAAC;IAC3B,eAAe,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,gBAAgB;AAChB,wBAAsB,0BAA0B,CAAC,EAC7C,OAAO,EACP,CAAC,EACD,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GAClB,EAAE,2BAA2B,4DA0B7B;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,mBAAmB,CAC/B,OAAO,SAAS,sBAAsB,GAAG,sBAAsB,EAC/D,QAAQ,SAAS,UAAU,GAAG,sBAAsB,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,EAC1E,MAAM,CAAC,EAAE,YAAY,CAAC,OAAO,EAAE,QAAQ,CAAC,kDAEzC"}
1
+ {"version":3,"file":"cheerio-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/cheerio-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAGjD,OAAO,KAAK,EACR,oBAAoB,EACpB,aAAa,EACb,mBAAmB,EACnB,YAAY,EACZ,sBAAsB,EACtB,kBAAkB,EAClB,2BAA2B,EAC3B,gBAAgB,EAChB,cAAc,EACd,eAAe,EACf,YAAY,EACZ,sBAAsB,EACzB,MAAM,eAAe,CAAC;AACvB,OAAO,EAAgB,WAAW,EAAkD,MAAM,eAAe,CAAC;AAC1G,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,KAAK,WAAW,EAA0B,KAAK,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE9F,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AAInC,MAAM,MAAM,mBAAmB,CAC3B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,YAAY,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE7D,MAAM,WAAW,qBAAqB,CAClC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,kBAAkB,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;CAAG;AAE3E,MAAM,MAAM,WAAW,CACnB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,gBAAgB,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEjE,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,2BAA2B,CAAC,QAAQ,EAAE,QAAQ,EAAE,cAAc,CAAC;IACrE;;;OAGG;IACH,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC;IAEtB;;;;;;;;;;;OAWG;IACH,eAAe,CAAC,QAAQ,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAErE;;;;;;;;;;;;;OAaG;IACH,gBAAgB,CAAC,QAAQ,CAAC,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;CACjF;AAED,MAAM,MAAM,qBAAqB,CAC7B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,cAAc,CAAC,sBAAsB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE/D;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4EG;AACH,qBAAa,cAAe,SAAQ,WAAW,CAAC,sBAAsB,CAAC;IACnE;;OAEG;gBAES,OAAO,CAAC,EAAE,qBAAqB,EAAE,MAAM,CAAC,EAAE,aAAa;cAI1C,UAAU,CAC/B,QAAQ,EAAE,eAAe,EACzB,KAAK,EAAE,OAAO,EACd,eAAe,EAAE,sBAAsB;;;;wCAmBG,mBAAmB;;cAgBjD,eAAe,CAAC,QAAQ,EAAE,eAAe,EAAE,KAAK,EAAE,OAAO;cAehD,kBAAkB,CAAC,OAAO,EAAE,sBAAsB;CAgB9E;AAED,UAAU,2BAA2B;IACjC,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,CAAC,EAAE,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC;IAC7B,YAAY,EAAE,eAAe,CAAC;IAC9B,aAAa,CAAC,EAAE,aAAa,CAAC;IAC9B,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IAC1C,kBAAkB,EAAE,MAAM,CAAC;IAC3B,eAAe,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,UAAU,gCAAgC;IACtC,YAAY,EAAE,oBAAoB,CAAC,cAAc,CAAC,CAAC;IACnD,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,CAAC,EAAE,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC;IAC7B,kBAAkB,EAAE,MAAM,CAAC;IAC3B,eAAe,CAAC,EAAE,MAAM,CAAC;CAC5B;AASD,gBAAgB;AAChB,wBAAsB,0BAA0B,CAC5C,OAAO,EAAE,2BAA2B,GAAG,gCAAgC,4DAmC1E;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,mBAAmB,CAC/B,OAAO,SAAS,sBAAsB,GAAG,sBAAsB,EAC/D,QAAQ,SAAS,UAAU,GAAG,sBAAsB,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,EAC1E,MAAM,CAAC,EAAE,YAAY,CAAC,OAAO,EAAE,QAAQ,CAAC,kDAEzC"}
@@ -105,6 +105,7 @@ class CheerioCrawler extends http_1.HttpCrawler {
105
105
  // Here we tell cheerio to use htmlparser2 for serialization, otherwise the conflict produces weird errors.
106
106
  _useHtmlParser2: true,
107
107
  });
108
+ const originalEnqueueLinks = crawlingContext.enqueueLinks;
108
109
  return {
109
110
  dom,
110
111
  $,
@@ -118,6 +119,7 @@ class CheerioCrawler extends http_1.HttpCrawler {
118
119
  onSkippedRequest: this.handleSkippedRequest,
119
120
  originalRequestUrl: crawlingContext.request.url,
120
121
  finalRequestUrl: crawlingContext.request.loadedUrl,
122
+ enqueueLinks: originalEnqueueLinks,
121
123
  });
122
124
  },
123
125
  };
@@ -153,24 +155,36 @@ class CheerioCrawler extends http_1.HttpCrawler {
153
155
  }
154
156
  exports.CheerioCrawler = CheerioCrawler;
155
157
  /** @internal */
156
- async function cheerioCrawlerEnqueueLinks({ options, $, requestQueue, robotsTxtFile, onSkippedRequest, originalRequestUrl, finalRequestUrl, }) {
158
+ function containsEnqueueLinks(options) {
159
+ return !!options.enqueueLinks;
160
+ }
161
+ /** @internal */
162
+ async function cheerioCrawlerEnqueueLinks(options) {
163
+ const { options: enqueueLinksOptions, $, originalRequestUrl, finalRequestUrl } = options;
157
164
  if (!$) {
158
165
  throw new Error('Cannot enqueue links because the DOM is not available.');
159
166
  }
160
167
  const baseUrl = (0, http_1.resolveBaseUrlForEnqueueLinksFiltering)({
161
- enqueueStrategy: options?.strategy,
168
+ enqueueStrategy: enqueueLinksOptions?.strategy,
162
169
  finalRequestUrl,
163
170
  originalRequestUrl,
164
- userProvidedBaseUrl: options?.baseUrl,
171
+ userProvidedBaseUrl: enqueueLinksOptions?.baseUrl,
165
172
  });
166
- const urls = (0, utils_1.extractUrlsFromCheerio)($, options?.selector ?? 'a', options?.baseUrl ?? finalRequestUrl ?? originalRequestUrl);
173
+ const urls = (0, utils_1.extractUrlsFromCheerio)($, enqueueLinksOptions?.selector ?? 'a', enqueueLinksOptions?.baseUrl ?? finalRequestUrl ?? originalRequestUrl);
174
+ if (containsEnqueueLinks(options)) {
175
+ return options.enqueueLinks({
176
+ urls,
177
+ baseUrl,
178
+ ...enqueueLinksOptions,
179
+ });
180
+ }
167
181
  return (0, http_1.enqueueLinks)({
168
- requestQueue,
169
- robotsTxtFile,
170
- onSkippedRequest,
182
+ requestQueue: options.requestQueue,
183
+ robotsTxtFile: options.robotsTxtFile,
184
+ onSkippedRequest: options.onSkippedRequest,
171
185
  urls,
172
186
  baseUrl,
173
- ...options,
187
+ ...enqueueLinksOptions,
174
188
  });
175
189
  }
176
190
  /**
@@ -1 +1 @@
1
- {"version":3,"file":"cheerio-crawler.js","sourceRoot":"","sources":["../../src/internals/cheerio-crawler.ts"],"names":[],"mappings":";;;AA0PA,gEAkCC;AA0BD,kDAKC;;AA1TD,qDAAmE;AAenE,wCAA0G;AAE1G,0CAA8F;AAE9F,yDAAmC;AACnC,6CAAwD;AACxD,mEAAgE;AA+DhE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4EG;AACH,MAAa,cAAe,SAAQ,kBAAmC;IACnE;;OAEG;IACH,qEAAqE;IACrE,YAAY,OAA+B,EAAE,MAAsB;QAC/D,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IAC3B,CAAC;IAEkB,KAAK,CAAC,UAAU,CAC/B,QAAyB,EACzB,KAAc,EACd,eAAuC;QAEvC,MAAM,IAAI,GAAG,MAAM,IAAA,gBAAkB,EAAC,QAAQ,CAAC,CAAC;QAChD,MAAM,GAAG,GAAG,IAAA,2BAAa,EAAC,IAAI,EAAE,EAAE,cAAc,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QAE1E,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,EAAE;YACzB,OAAO,EAAE,KAAK;YACd,yGAAyG;YACzG,gFAAgF;YAChF,2GAA2G;YAC3G,eAAe,EAAE,IAAI;SACN,CAAC,CAAC;QAErB,OAAO;YACH,GAAG;YACH,CAAC;YACD,IAAI;YACJ,YAAY,EAAE,KAAK,EAAE,cAAoC,EAAE,EAAE;gBACzD,OAAO,0BAA0B,CAAC;oBAC9B,OAAO,EAAE,EAAE,GAAG,cAAc,EAAE,KAAK,EAAE,IAAI,CAAC,6BAA6B,CAAC,cAAc,EAAE,KAAK,CAAC,EAAE;oBAChG,CAAC;oBACD,YAAY,EAAE,MAAM,IAAI,CAAC,eAAe,EAAE;oBAC1C,aAAa,EAAE,MAAM,IAAI,CAAC,sBAAsB,CAAC,eAAe,CAAC,OAAO,CAAC,GAAG,CAAC;oBAC7E,gBAAgB,EAAE,IAAI,CAAC,oBAAoB;oBAC3C,kBAAkB,EAAE,eAAe,CAAC,OAAO,CAAC,GAAG;oBAC/C,eAAe,EAAE,eAAe,CAAC,OAAO,CAAC,SAAS;iBACrD,CAAC,CAAC;YACP,CAAC;SACJ,CAAC;IACN,CAAC;IAED,oCAAoC;IAC1B,KAAK,CAAC,eAAe,CAAC,QAAyB,EAAE,KAAc;QACrE,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACnC,MAAM,UAAU,GAAG,IAAI,wBAAU,CAC7B,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE;gBACT,IAAI,GAAG;oBAAE,MAAM,CAAC,GAAG,CAAC,CAAC;;oBAChB,OAAO,CAAC,GAAG,CAAC,CAAC;YACtB,CAAC,EACD,EAAE,OAAO,EAAE,KAAK,EAAE,CACrB,CAAC;YACF,MAAM,MAAM,GAAG,IAAI,+BAAc,CAAC,UAAU,EAAE,EAAE,cAAc,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;YACxF,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC3B,QAAQ,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC9C,CAAC,CAAC,CAAC;IACP,CAAC;IAEkB,KAAK,CAAC,kBAAkB,CAAC,OAA+B;QACvE,OAAO,CAAC,eAAe,GAAG,KAAK,EAAE,QAAiB,EAAE,UAAmB,EAAE,EAAE;YACvE,IAAI,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACzC,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;QACL,CAAC,CAAC;QACF,OAAO,CAAC,gBAAgB,GAAG,KAAK,EAAE,QAAiB,EAAE,SAAkB,EAAE,EAAE;YACvE,IAAI,QAAQ,EAAE,CAAC;gBACX,MAAM,OAAO,CAAC,eAAe,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;YACvD,CAAC;YAED,OAAO,OAAO,CAAC,CAAC,CAAC;QACrB,CAAC,CAAC;QAEF,MAAM,KAAK,CAAC,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAC5C,CAAC;CACJ;AA3ED,wCA2EC;AAYD,gBAAgB;AACT,KAAK,UAAU,0BAA0B,CAAC,EAC7C,OAAO,EACP,CAAC,EACD,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GACW;IAC1B,IAAI,CAAC,CAAC,EAAE,CAAC;QACL,MAAM,IAAI,KAAK,CAAC,wDAAwD,CAAC,CAAC;IAC9E,CAAC;IAED,MAAM,OAAO,GAAG,IAAA,6CAAsC,EAAC;QACnD,eAAe,EAAE,OAAO,EAAE,QAAQ;QAClC,eAAe;QACf,kBAAkB;QAClB,mBAAmB,EAAE,OAAO,EAAE,OAAO;KACxC,CAAC,CAAC;IAEH,MAAM,IAAI,GAAG,IAAA,8BAAsB,EAC/B,CAAC,EACD,OAAO,EAAE,QAAQ,IAAI,GAAG,EACxB,OAAO,EAAE,OAAO,IAAI,eAAe,IAAI,kBAAkB,CAC5D,CAAC;IAEF,OAAO,IAAA,mBAAY,EAAC;QAChB,YAAY;QACZ,aAAa;QACb,gBAAgB;QAChB,IAAI;QACJ,OAAO;QACP,GAAG,OAAO;KACb,CAAC,CAAC;AACP,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,SAAgB,mBAAmB,CAGjC,MAAwC;IACtC,OAAO,aAAM,CAAC,MAAM,CAAU,MAAM,CAAC,CAAC;AAC1C,CAAC"}
1
+ {"version":3,"file":"cheerio-crawler.js","sourceRoot":"","sources":["../../src/internals/cheerio-crawler.ts"],"names":[],"mappings":";;;AA6QA,gEAoCC;AA0BD,kDAKC;;AA/UD,qDAAmE;AAgBnE,wCAA0G;AAE1G,0CAA8F;AAE9F,yDAAmC;AACnC,6CAAwD;AACxD,mEAAgE;AA+DhE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4EG;AACH,MAAa,cAAe,SAAQ,kBAAmC;IACnE;;OAEG;IACH,qEAAqE;IACrE,YAAY,OAA+B,EAAE,MAAsB;QAC/D,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IAC3B,CAAC;IAEkB,KAAK,CAAC,UAAU,CAC/B,QAAyB,EACzB,KAAc,EACd,eAAuC;QAEvC,MAAM,IAAI,GAAG,MAAM,IAAA,gBAAkB,EAAC,QAAQ,CAAC,CAAC;QAChD,MAAM,GAAG,GAAG,IAAA,2BAAa,EAAC,IAAI,EAAE,EAAE,cAAc,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QAE1E,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,EAAE;YACzB,OAAO,EAAE,KAAK;YACd,yGAAyG;YACzG,gFAAgF;YAChF,2GAA2G;YAC3G,eAAe,EAAE,IAAI;SACN,CAAC,CAAC;QAErB,MAAM,oBAAoB,GAAG,eAAe,CAAC,YAAY,CAAC;QAE1D,OAAO;YACH,GAAG;YACH,CAAC;YACD,IAAI;YACJ,YAAY,EAAE,KAAK,EAAE,cAAoC,EAAE,EAAE;gBACzD,OAAO,0BAA0B,CAAC;oBAC9B,OAAO,EAAE,EAAE,GAAG,cAAc,EAAE,KAAK,EAAE,IAAI,CAAC,6BAA6B,CAAC,cAAc,EAAE,KAAK,CAAC,EAAE;oBAChG,CAAC;oBACD,YAAY,EAAE,MAAM,IAAI,CAAC,eAAe,EAAE;oBAC1C,aAAa,EAAE,MAAM,IAAI,CAAC,sBAAsB,CAAC,eAAe,CAAC,OAAO,CAAC,GAAG,CAAC;oBAC7E,gBAAgB,EAAE,IAAI,CAAC,oBAAoB;oBAC3C,kBAAkB,EAAE,eAAe,CAAC,OAAO,CAAC,GAAG;oBAC/C,eAAe,EAAE,eAAe,CAAC,OAAO,CAAC,SAAS;oBAClD,YAAY,EAAE,oBAAoB;iBACrC,CAAC,CAAC;YACP,CAAC;SACJ,CAAC;IACN,CAAC;IAED,oCAAoC;IAC1B,KAAK,CAAC,eAAe,CAAC,QAAyB,EAAE,KAAc;QACrE,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACnC,MAAM,UAAU,GAAG,IAAI,wBAAU,CAC7B,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE;gBACT,IAAI,GAAG;oBAAE,MAAM,CAAC,GAAG,CAAC,CAAC;;oBAChB,OAAO,CAAC,GAAG,CAAC,CAAC;YACtB,CAAC,EACD,EAAE,OAAO,EAAE,KAAK,EAAE,CACrB,CAAC;YACF,MAAM,MAAM,GAAG,IAAI,+BAAc,CAAC,UAAU,EAAE,EAAE,cAAc,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;YACxF,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC3B,QAAQ,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC9C,CAAC,CAAC,CAAC;IACP,CAAC;IAEkB,KAAK,CAAC,kBAAkB,CAAC,OAA+B;QACvE,OAAO,CAAC,eAAe,GAAG,KAAK,EAAE,QAAiB,EAAE,UAAmB,EAAE,EAAE;YACvE,IAAI,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACzC,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;QACL,CAAC,CAAC;QACF,OAAO,CAAC,gBAAgB,GAAG,KAAK,EAAE,QAAiB,EAAE,SAAkB,EAAE,EAAE;YACvE,IAAI,QAAQ,EAAE,CAAC;gBACX,MAAM,OAAO,CAAC,eAAe,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;YACvD,CAAC;YAED,OAAO,OAAO,CAAC,CAAC,CAAC;QACrB,CAAC,CAAC;QAEF,MAAM,KAAK,CAAC,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAC5C,CAAC;CACJ;AA9ED,wCA8EC;AAoBD,gBAAgB;AAChB,SAAS,oBAAoB,CACzB,OAAuE;IAEvE,OAAO,CAAC,CAAE,OAA4C,CAAC,YAAY,CAAC;AACxE,CAAC;AAED,gBAAgB;AACT,KAAK,UAAU,0BAA0B,CAC5C,OAAuE;IAEvE,MAAM,EAAE,OAAO,EAAE,mBAAmB,EAAE,CAAC,EAAE,kBAAkB,EAAE,eAAe,EAAE,GAAG,OAAO,CAAC;IACzF,IAAI,CAAC,CAAC,EAAE,CAAC;QACL,MAAM,IAAI,KAAK,CAAC,wDAAwD,CAAC,CAAC;IAC9E,CAAC;IAED,MAAM,OAAO,GAAG,IAAA,6CAAsC,EAAC;QACnD,eAAe,EAAE,mBAAmB,EAAE,QAAQ;QAC9C,eAAe;QACf,kBAAkB;QAClB,mBAAmB,EAAE,mBAAmB,EAAE,OAAO;KACpD,CAAC,CAAC;IAEH,MAAM,IAAI,GAAG,IAAA,8BAAsB,EAC/B,CAAC,EACD,mBAAmB,EAAE,QAAQ,IAAI,GAAG,EACpC,mBAAmB,EAAE,OAAO,IAAI,eAAe,IAAI,kBAAkB,CACxE,CAAC;IAEF,IAAI,oBAAoB,CAAC,OAAO,CAAC,EAAE,CAAC;QAChC,OAAO,OAAO,CAAC,YAAY,CAAC;YACxB,IAAI;YACJ,OAAO;YACP,GAAG,mBAAmB;SACzB,CAAC,CAAC;IACP,CAAC;IACD,OAAO,IAAA,mBAAY,EAAC;QAChB,YAAY,EAAE,OAAO,CAAC,YAAY;QAClC,aAAa,EAAE,OAAO,CAAC,aAAa;QACpC,gBAAgB,EAAE,OAAO,CAAC,gBAAgB;QAC1C,IAAI;QACJ,OAAO;QACP,GAAG,mBAAmB;KACzB,CAAC,CAAC;AACP,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,SAAgB,mBAAmB,CAGjC,MAAwC;IACtC,OAAO,aAAM,CAAC,MAAM,CAAU,MAAM,CAAC,CAAC;AAC1C,CAAC"}
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@crawlee/cheerio",
3
- "version": "3.13.11-beta.3",
3
+ "version": "3.13.11-beta.4",
4
4
  "description": "The scalable web crawling and scraping library for JavaScript/Node.js. Enables development of data extraction and web automation jobs (not only) with headless Chrome and Puppeteer.",
5
5
  "engines": {
6
6
  "node": ">=16.0.0"
@@ -53,9 +53,9 @@
53
53
  "access": "public"
54
54
  },
55
55
  "dependencies": {
56
- "@crawlee/http": "3.13.11-beta.3",
57
- "@crawlee/types": "3.13.11-beta.3",
58
- "@crawlee/utils": "3.13.11-beta.3",
56
+ "@crawlee/http": "3.13.11-beta.4",
57
+ "@crawlee/types": "3.13.11-beta.4",
58
+ "@crawlee/utils": "3.13.11-beta.4",
59
59
  "cheerio": "1.0.0-rc.12",
60
60
  "htmlparser2": "^9.0.0",
61
61
  "tslib": "^2.4.0"
@@ -67,5 +67,5 @@
67
67
  }
68
68
  }
69
69
  },
70
- "gitHead": "df9a4c484593a7fc01e49b1341f4d780cdc6f274"
70
+ "gitHead": "cdf793e1445a3f4f5999bfa40f447815227c7828"
71
71
  }