@crawlee/linkedom 3.13.6-beta.0 → 4.0.0-beta.0

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
@@ -1,3 +1,3 @@
1
1
  export * from '@crawlee/http';
2
- export * from './internals/linkedom-crawler';
2
+ export * from './internals/linkedom-crawler.js';
3
3
  //# sourceMappingURL=index.d.ts.map
package/index.d.ts.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,8BAA8B,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,iCAAiC,CAAC"}
package/index.js CHANGED
@@ -1,6 +1,3 @@
1
- "use strict";
2
- Object.defineProperty(exports, "__esModule", { value: true });
3
- const tslib_1 = require("tslib");
4
- tslib_1.__exportStar(require("@crawlee/http"), exports);
5
- tslib_1.__exportStar(require("./internals/linkedom-crawler"), exports);
1
+ export * from '@crawlee/http';
2
+ export * from './internals/linkedom-crawler.js';
6
3
  //# sourceMappingURL=index.js.map
package/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;;AAAA,wDAA8B;AAC9B,uEAA6C"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,iCAAiC,CAAC"}
@@ -1 +1 @@
1
- {"version":3,"file":"linkedom-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/linkedom-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAEjD,OAAO,KAAK,EACR,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,EAEH,WAAW,EAId,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,KAAK,WAAW,EAAE,KAAK,aAAa,EAAS,MAAM,gBAAgB,CAAC;AAO7E,MAAM,MAAM,oBAAoB,CAC5B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,YAAY,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE9D,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,kBAAkB,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;CAAG;AAE5E,MAAM,WAAW,kCAAmC,SAAQ,IAAI,CAAC,mBAAmB,EAAE,MAAM,GAAG,cAAc,CAAC;CAAG;AAEjH,MAAM,MAAM,YAAY,CACpB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,gBAAgB,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAElE,MAAM,WAAW,uBAAuB,CACpC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,2BAA2B,CAAC,QAAQ,EAAE,QAAQ,EAAE,eAAe,CAAC;IACtE,MAAM,EAAE,MAAM,CAAC;IAMf,QAAQ,EAAE,QAAQ,CAAC;IAEnB;;;;;;;;;;;;OAYG;IACH,eAAe,CAAC,QAAQ,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAErE;;;;;;;;;;;OAWG;IACH,gBAAgB,CAAC,QAAQ,CAAC,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;CACjF;AAED,MAAM,MAAM,sBAAsB,CAC9B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,cAAc,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEhE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqEG;AAEH,qBAAa,eAAgB,SAAQ,WAAW,CAAC,uBAAuB,CAAC;IACrE,OAAO,CAAC,MAAM,CAAC,MAAM,CAAmB;cAEf,UAAU,CAC/B,QAAQ,EAAE,eAAe,EACzB,KAAK,EAAE,OAAO,EACd,eAAe,EAAE,uBAAuB;;;2BAaF,QAAQ;wCAEJ,kCAAkC;;IAcjE,kBAAkB,CAAC,OAAO,EAAE,uBAAuB;CA0BrE;AAED,UAAU,2BAA2B;IACjC,OAAO,CAAC,EAAE,kCAAkC,CAAC;IAC7C,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,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,2BAA2B,CAAC,EAC9C,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GAClB,EAAE,2BAA2B,4DA0B7B;AAmBD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,oBAAoB,CAChC,OAAO,SAAS,uBAAuB,GAAG,uBAAuB,EACjE,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":"linkedom-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/linkedom-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAEjD,OAAO,KAAK,EACR,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,EAEH,WAAW,EAId,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,KAAK,WAAW,EAAE,KAAK,aAAa,EAAS,MAAM,gBAAgB,CAAC;AAM7E,MAAM,MAAM,oBAAoB,CAC5B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,YAAY,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE9D,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,kBAAkB,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;CAAG;AAE5E,MAAM,WAAW,kCAAmC,SAAQ,IAAI,CAAC,mBAAmB,EAAE,MAAM,GAAG,cAAc,CAAC;CAAG;AAEjH,MAAM,MAAM,YAAY,CACpB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,gBAAgB,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAElE,MAAM,WAAW,uBAAuB,CACpC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,2BAA2B,CAAC,QAAQ,EAAE,QAAQ,EAAE,eAAe,CAAC;IACtE,MAAM,EAAE,MAAM,CAAC;IAMf,QAAQ,EAAE,QAAQ,CAAC;IAEnB;;;;;;;;;;;;OAYG;IACH,eAAe,CAAC,QAAQ,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAErE;;;;;;;;;;;OAWG;IACH,gBAAgB,CAAC,QAAQ,CAAC,EAAE,MAAM,EAAE,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;CACjF;AAED,MAAM,MAAM,sBAAsB,CAC9B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,cAAc,CAAC,uBAAuB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEhE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqEG;AAEH,qBAAa,eAAgB,SAAQ,WAAW,CAAC,uBAAuB,CAAC;IACrE,OAAO,CAAC,MAAM,CAAC,MAAM,CAAmB;cAEf,UAAU,CAC/B,QAAQ,EAAE,eAAe,EACzB,KAAK,EAAE,OAAO,EACd,eAAe,EAAE,uBAAuB;;;2BAaF,QAAQ;wCAEJ,kCAAkC;;IAcjE,kBAAkB,CAAC,OAAO,EAAE,uBAAuB;CA0BrE;AAED,UAAU,2BAA2B;IACjC,OAAO,CAAC,EAAE,kCAAkC,CAAC;IAC7C,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,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,2BAA2B,CAAC,EAC9C,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GAClB,EAAE,2BAA2B,4DA0B7B;AAmBD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,oBAAoB,CAChC,OAAO,SAAS,uBAAuB,GAAG,uBAAuB,EACjE,QAAQ,SAAS,UAAU,GAAG,sBAAsB,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,EAC1E,MAAM,CAAC,EAAE,YAAY,CAAC,OAAO,EAAE,QAAQ,CAAC,kDAEzC"}
@@ -1,15 +1,8 @@
1
- "use strict";
2
- Object.defineProperty(exports, "__esModule", { value: true });
3
- exports.LinkeDOMCrawler = void 0;
4
- exports.linkedomCrawlerEnqueueLinks = linkedomCrawlerEnqueueLinks;
5
- exports.createLinkeDOMRouter = createLinkeDOMRouter;
6
- const tslib_1 = require("tslib");
7
- const http_1 = require("@crawlee/http");
8
- const utils_1 = require("@crawlee/utils");
9
- const cheerio = tslib_1.__importStar(require("cheerio"));
10
- // @ts-expect-error This throws a compilation error due to TypeScript not inferring the module has CJS versions too
11
- const cached_1 = require("linkedom/cached");
12
- const utilities_1 = require("@apify/utilities");
1
+ import { enqueueLinks, HttpCrawler, resolveBaseUrlForEnqueueLinksFiltering, Router, tryAbsoluteURL, } from '@crawlee/http';
2
+ import { sleep } from '@crawlee/utils';
3
+ import * as cheerio from 'cheerio';
4
+ import { DOMParser } from 'linkedom/cached';
5
+ import { concatStreamToBuffer } from '@apify/utilities';
13
6
  /**
14
7
  * Provides a framework for the parallel crawling of web pages using plain HTTP requests and
15
8
  * [linkedom](https://www.npmjs.com/package/linkedom) LinkeDOM implementation.
@@ -80,9 +73,10 @@ const utilities_1 = require("@apify/utilities");
80
73
  * ```
81
74
  * @category Crawlers
82
75
  */
83
- class LinkeDOMCrawler extends http_1.HttpCrawler {
76
+ export class LinkeDOMCrawler extends HttpCrawler {
77
+ static parser = new DOMParser();
84
78
  async _parseHTML(response, isXml, crawlingContext) {
85
- const body = await (0, utilities_1.concatStreamToBuffer)(response);
79
+ const body = await concatStreamToBuffer(response);
86
80
  const document = LinkeDOMCrawler.parser.parseFromString(body.toString(), isXml ? 'text/xml' : 'text/html');
87
81
  return {
88
82
  window: document.defaultView,
@@ -107,18 +101,18 @@ class LinkeDOMCrawler extends http_1.HttpCrawler {
107
101
  };
108
102
  }
109
103
  async _runRequestHandler(context) {
110
- context.waitForSelector = async (selector, timeoutMs = 5000) => {
104
+ context.waitForSelector = async (selector, timeoutMs = 5_000) => {
111
105
  const $ = cheerio.load(context.body);
112
106
  if ($(selector).get().length === 0) {
113
107
  if (timeoutMs) {
114
- await (0, utils_1.sleep)(50);
108
+ await sleep(50);
115
109
  await context.waitForSelector(selector, Math.max(timeoutMs - 50, 0));
116
110
  return;
117
111
  }
118
112
  throw new Error(`Selector '${selector}' not found.`);
119
113
  }
120
114
  };
121
- context.parseWithCheerio = async (selector, _timeoutMs = 5000) => {
115
+ context.parseWithCheerio = async (selector, _timeoutMs = 5_000) => {
122
116
  const $ = cheerio.load(context.body);
123
117
  if (selector && $(selector).get().length === 0) {
124
118
  throw new Error(`Selector '${selector}' not found.`);
@@ -128,26 +122,19 @@ class LinkeDOMCrawler extends http_1.HttpCrawler {
128
122
  await super._runRequestHandler(context);
129
123
  }
130
124
  }
131
- exports.LinkeDOMCrawler = LinkeDOMCrawler;
132
- Object.defineProperty(LinkeDOMCrawler, "parser", {
133
- enumerable: true,
134
- configurable: true,
135
- writable: true,
136
- value: new cached_1.DOMParser()
137
- });
138
125
  /** @internal */
139
- async function linkedomCrawlerEnqueueLinks({ options, window, requestQueue, robotsTxtFile, onSkippedRequest, originalRequestUrl, finalRequestUrl, }) {
126
+ export async function linkedomCrawlerEnqueueLinks({ options, window, requestQueue, robotsTxtFile, onSkippedRequest, originalRequestUrl, finalRequestUrl, }) {
140
127
  if (!window) {
141
128
  throw new Error('Cannot enqueue links because the DOM is not available.');
142
129
  }
143
- const baseUrl = (0, http_1.resolveBaseUrlForEnqueueLinksFiltering)({
130
+ const baseUrl = resolveBaseUrlForEnqueueLinksFiltering({
144
131
  enqueueStrategy: options?.strategy,
145
132
  finalRequestUrl,
146
133
  originalRequestUrl,
147
134
  userProvidedBaseUrl: options?.baseUrl,
148
135
  });
149
136
  const urls = extractUrlsFromWindow(window, options?.selector ?? 'a', options?.baseUrl ?? finalRequestUrl ?? originalRequestUrl);
150
- return (0, http_1.enqueueLinks)({
137
+ return enqueueLinks({
151
138
  requestQueue,
152
139
  robotsTxtFile,
153
140
  onSkippedRequest,
@@ -168,7 +155,7 @@ function extractUrlsFromWindow(window, selector, baseUrl) {
168
155
  if (href === undefined) {
169
156
  return undefined;
170
157
  }
171
- return (0, http_1.tryAbsoluteURL)(href, baseUrl);
158
+ return tryAbsoluteURL(href, baseUrl);
172
159
  })
173
160
  .filter((href) => href !== undefined && href !== '');
174
161
  }
@@ -196,7 +183,7 @@ function extractUrlsFromWindow(window, selector, baseUrl) {
196
183
  * await crawler.run();
197
184
  * ```
198
185
  */
199
- function createLinkeDOMRouter(routes) {
200
- return http_1.Router.create(routes);
186
+ export function createLinkeDOMRouter(routes) {
187
+ return Router.create(routes);
201
188
  }
202
189
  //# sourceMappingURL=linkedom-crawler.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"linkedom-crawler.js","sourceRoot":"","sources":["../../src/internals/linkedom-crawler.ts"],"names":[],"mappings":";;;AA8OA,kEAkCC;AA2CD,oDAKC;;AAlTD,wCAMuB;AAEvB,0CAA6E;AAC7E,yDAAmC;AACnC,mHAAmH;AACnH,4CAA4C;AAE5C,gDAAwD;AAkExD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqEG;AAEH,MAAa,eAAgB,SAAQ,kBAAoC;IAGlD,KAAK,CAAC,UAAU,CAC/B,QAAyB,EACzB,KAAc,EACd,eAAwC;QAExC,MAAM,IAAI,GAAG,MAAM,IAAA,gCAAoB,EAAC,QAAQ,CAAC,CAAC;QAElD,MAAM,QAAQ,GAAG,eAAe,CAAC,MAAM,CAAC,eAAe,CAAC,IAAI,CAAC,QAAQ,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC;QAE3G,OAAO;YACH,MAAM,EAAE,QAAQ,CAAC,WAAW;YAC5B,IAAI,IAAI;gBACJ,OAAO,QAAQ,CAAC,eAAe,CAAC,SAAS,CAAC;YAC9C,CAAC;YACD,IAAI,QAAQ;gBACR,iEAAiE;gBACjE,OAAO,QAA+B,CAAC;YAC3C,CAAC;YACD,YAAY,EAAE,KAAK,EAAE,cAAmD,EAAE,EAAE;gBACxE,OAAO,2BAA2B,CAAC;oBAC/B,OAAO,EAAE,cAAc;oBACvB,MAAM,EAAE,QAAQ,CAAC,WAAW;oBAC5B,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,gBAAgB;oBACvC,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;IAEQ,KAAK,CAAC,kBAAkB,CAAC,OAAgC;QAC9D,OAAO,CAAC,eAAe,GAAG,KAAK,EAAE,QAAgB,EAAE,SAAS,GAAG,IAAK,EAAE,EAAE;YACpE,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAErC,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACjC,IAAI,SAAS,EAAE,CAAC;oBACZ,MAAM,IAAA,aAAK,EAAC,EAAE,CAAC,CAAC;oBAChB,MAAM,OAAO,CAAC,eAAe,CAAC,QAAQ,EAAE,IAAI,CAAC,GAAG,CAAC,SAAS,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;oBACrE,OAAO;gBACX,CAAC;gBAED,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;QACL,CAAC,CAAC;QACF,OAAO,CAAC,gBAAgB,GAAG,KAAK,EAAE,QAAiB,EAAE,UAAU,GAAG,IAAK,EAAE,EAAE;YACvE,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAErC,IAAI,QAAQ,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC7C,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;YAED,OAAO,CAAC,CAAC;QACb,CAAC,CAAC;QAEF,MAAM,KAAK,CAAC,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAC5C,CAAC;;AA5DL,0CA6DC;AA5DkB;;;;WAAS,IAAI,kBAAS,EAAE;GAAC;AAwE5C,gBAAgB;AACT,KAAK,UAAU,2BAA2B,CAAC,EAC9C,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GACW;IAC1B,IAAI,CAAC,MAAM,EAAE,CAAC;QACV,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,qBAAqB,CAC9B,MAAM,EACN,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;;;GAGG;AACH,SAAS,qBAAqB,CAAC,MAAc,EAAE,QAAgB,EAAE,OAAe;IAC5E,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC;SACxD,GAAG,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;SACvB,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,EAAE,CAAC;SACnD,GAAG,CAAC,CAAC,IAAwB,EAAE,EAAE;QAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACrB,OAAO,SAAS,CAAC;QACrB,CAAC;QACD,OAAO,IAAA,qBAAc,EAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IACzC,CAAC,CAAC;SACD,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,EAAE,CAAa,CAAC;AACzE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,SAAgB,oBAAoB,CAGlC,MAAwC;IACtC,OAAO,aAAM,CAAC,MAAM,CAAU,MAAM,CAAC,CAAC;AAC1C,CAAC"}
1
+ {"version":3,"file":"linkedom-crawler.js","sourceRoot":"","sources":["../../src/internals/linkedom-crawler.ts"],"names":[],"mappings":"AAcA,OAAO,EACH,YAAY,EACZ,WAAW,EACX,sCAAsC,EACtC,MAAM,EACN,cAAc,GACjB,MAAM,eAAe,CAAC;AAEvB,OAAO,EAAwC,KAAK,EAAE,MAAM,gBAAgB,CAAC;AAC7E,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AACnC,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAE5C,OAAO,EAAE,oBAAoB,EAAE,MAAM,kBAAkB,CAAC;AAkExD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqEG;AAEH,MAAM,OAAO,eAAgB,SAAQ,WAAoC;IAC7D,MAAM,CAAC,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;IAErB,KAAK,CAAC,UAAU,CAC/B,QAAyB,EACzB,KAAc,EACd,eAAwC;QAExC,MAAM,IAAI,GAAG,MAAM,oBAAoB,CAAC,QAAQ,CAAC,CAAC;QAElD,MAAM,QAAQ,GAAG,eAAe,CAAC,MAAM,CAAC,eAAe,CAAC,IAAI,CAAC,QAAQ,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC;QAE3G,OAAO;YACH,MAAM,EAAE,QAAQ,CAAC,WAAW;YAC5B,IAAI,IAAI;gBACJ,OAAO,QAAQ,CAAC,eAAe,CAAC,SAAS,CAAC;YAC9C,CAAC;YACD,IAAI,QAAQ;gBACR,iEAAiE;gBACjE,OAAO,QAA+B,CAAC;YAC3C,CAAC;YACD,YAAY,EAAE,KAAK,EAAE,cAAmD,EAAE,EAAE;gBACxE,OAAO,2BAA2B,CAAC;oBAC/B,OAAO,EAAE,cAAc;oBACvB,MAAM,EAAE,QAAQ,CAAC,WAAW;oBAC5B,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,gBAAgB;oBACvC,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;IAEQ,KAAK,CAAC,kBAAkB,CAAC,OAAgC;QAC9D,OAAO,CAAC,eAAe,GAAG,KAAK,EAAE,QAAgB,EAAE,SAAS,GAAG,KAAK,EAAE,EAAE;YACpE,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAErC,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACjC,IAAI,SAAS,EAAE,CAAC;oBACZ,MAAM,KAAK,CAAC,EAAE,CAAC,CAAC;oBAChB,MAAM,OAAO,CAAC,eAAe,CAAC,QAAQ,EAAE,IAAI,CAAC,GAAG,CAAC,SAAS,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;oBACrE,OAAO;gBACX,CAAC;gBAED,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;QACL,CAAC,CAAC;QACF,OAAO,CAAC,gBAAgB,GAAG,KAAK,EAAE,QAAiB,EAAE,UAAU,GAAG,KAAK,EAAE,EAAE;YACvE,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAErC,IAAI,QAAQ,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC7C,MAAM,IAAI,KAAK,CAAC,aAAa,QAAQ,cAAc,CAAC,CAAC;YACzD,CAAC;YAED,OAAO,CAAC,CAAC;QACb,CAAC,CAAC;QAEF,MAAM,KAAK,CAAC,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAC5C,CAAC;;AAaL,gBAAgB;AAChB,MAAM,CAAC,KAAK,UAAU,2BAA2B,CAAC,EAC9C,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,kBAAkB,EAClB,eAAe,GACW;IAC1B,IAAI,CAAC,MAAM,EAAE,CAAC;QACV,MAAM,IAAI,KAAK,CAAC,wDAAwD,CAAC,CAAC;IAC9E,CAAC;IAED,MAAM,OAAO,GAAG,sCAAsC,CAAC;QACnD,eAAe,EAAE,OAAO,EAAE,QAAQ;QAClC,eAAe;QACf,kBAAkB;QAClB,mBAAmB,EAAE,OAAO,EAAE,OAAO;KACxC,CAAC,CAAC;IAEH,MAAM,IAAI,GAAG,qBAAqB,CAC9B,MAAM,EACN,OAAO,EAAE,QAAQ,IAAI,GAAG,EACxB,OAAO,EAAE,OAAO,IAAI,eAAe,IAAI,kBAAkB,CAC5D,CAAC;IAEF,OAAO,YAAY,CAAC;QAChB,YAAY;QACZ,aAAa;QACb,gBAAgB;QAChB,IAAI;QACJ,OAAO;QACP,GAAG,OAAO;KACb,CAAC,CAAC;AACP,CAAC;AAED;;;GAGG;AACH,SAAS,qBAAqB,CAAC,MAAc,EAAE,QAAgB,EAAE,OAAe;IAC5E,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC;SACxD,GAAG,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;SACvB,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,EAAE,CAAC;SACnD,GAAG,CAAC,CAAC,IAAwB,EAAE,EAAE;QAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACrB,OAAO,SAAS,CAAC;QACrB,CAAC;QACD,OAAO,cAAc,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IACzC,CAAC,CAAC;SACD,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,EAAE,CAAa,CAAC;AACzE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,MAAM,UAAU,oBAAoB,CAGlC,MAAwC;IACtC,OAAO,MAAM,CAAC,MAAM,CAAU,MAAM,CAAC,CAAC;AAC1C,CAAC"}
package/package.json CHANGED
@@ -1,19 +1,13 @@
1
1
  {
2
2
  "name": "@crawlee/linkedom",
3
- "version": "3.13.6-beta.0",
3
+ "version": "4.0.0-beta.0",
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
- "node": ">=16.0.0"
6
+ "node": ">=22.0.0"
7
7
  },
8
- "main": "./index.js",
9
- "module": "./index.mjs",
10
- "types": "./index.d.ts",
8
+ "type": "module",
11
9
  "exports": {
12
- ".": {
13
- "import": "./index.mjs",
14
- "require": "./index.js",
15
- "types": "./index.d.ts"
16
- },
10
+ ".": "./index.js",
17
11
  "./package.json": "./package.json"
18
12
  },
19
13
  "keywords": [
@@ -46,20 +40,20 @@
46
40
  "scripts": {
47
41
  "build": "yarn clean && yarn compile && yarn copy",
48
42
  "clean": "rimraf ./dist",
49
- "compile": "tsc -p tsconfig.build.json && gen-esm-wrapper ./index.js ./index.mjs",
43
+ "compile": "tsc -p tsconfig.build.json",
50
44
  "copy": "tsx ../../scripts/copy.ts"
51
45
  },
52
46
  "publishConfig": {
53
47
  "access": "public"
54
48
  },
55
49
  "dependencies": {
56
- "@apify/timeout": "^0.3.0",
57
- "@apify/utilities": "^2.7.10",
58
- "@crawlee/http": "^3.13.6-beta.0",
59
- "@crawlee/types": "^3.13.6-beta.0",
60
- "linkedom": "^0.18.0",
61
- "ow": "^0.28.2",
62
- "tslib": "^2.4.0"
50
+ "@apify/timeout": "^0.3.2",
51
+ "@apify/utilities": "^2.15.5",
52
+ "@crawlee/http": "4.0.0-beta.0",
53
+ "@crawlee/types": "4.0.0-beta.0",
54
+ "linkedom": "^0.18.10",
55
+ "ow": "^2.0.0",
56
+ "tslib": "^2.8.1"
63
57
  },
64
58
  "lerna": {
65
59
  "command": {
@@ -68,5 +62,5 @@
68
62
  }
69
63
  }
70
64
  },
71
- "gitHead": "595148da5825258b9c92e49071ae6c76bc824c86"
65
+ "gitHead": "927bdafa403ff347327158b01d20b817378168a7"
72
66
  }