@crawlee/jsdom 3.13.1-beta.9 → 3.13.2-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/README.md
CHANGED
|
@@ -28,7 +28,7 @@ Crawlee is available as the [`crawlee`](https://www.npmjs.com/package/crawlee) N
|
|
|
28
28
|
|
|
29
29
|
## Installation
|
|
30
30
|
|
|
31
|
-
We recommend visiting the [Introduction tutorial](https://crawlee.dev/docs/introduction) in Crawlee documentation for more information.
|
|
31
|
+
We recommend visiting the [Introduction tutorial](https://crawlee.dev/js/docs/introduction) in Crawlee documentation for more information.
|
|
32
32
|
|
|
33
33
|
> Crawlee requires **Node.js 16 or higher**.
|
|
34
34
|
|
|
@@ -78,7 +78,7 @@ const crawler = new PlaywrightCrawler({
|
|
|
78
78
|
await crawler.run(['https://crawlee.dev']);
|
|
79
79
|
```
|
|
80
80
|
|
|
81
|
-
By default, Crawlee stores data to `./storage` in the current working directory. You can override this directory via Crawlee configuration. For details, see [Configuration guide](https://crawlee.dev/docs/guides/configuration), [Request storage](https://crawlee.dev/docs/guides/request-storage) and [Result storage](https://crawlee.dev/docs/guides/result-storage).
|
|
81
|
+
By default, Crawlee stores data to `./storage` in the current working directory. You can override this directory via Crawlee configuration. For details, see [Configuration guide](https://crawlee.dev/js/docs/guides/configuration), [Request storage](https://crawlee.dev/js/docs/guides/request-storage) and [Result storage](https://crawlee.dev/js/docs/guides/result-storage).
|
|
82
82
|
|
|
83
83
|
### Installing pre-release versions
|
|
84
84
|
|
|
@@ -1,8 +1,8 @@
|
|
|
1
|
-
import type { IncomingMessage } from 'http';
|
|
2
|
-
import type { EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler,
|
|
1
|
+
import type { IncomingMessage } from 'node:http';
|
|
2
|
+
import type { Configuration, EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler, RequestProvider, RouterRoutes } from '@crawlee/http';
|
|
3
3
|
import { HttpCrawler } from '@crawlee/http';
|
|
4
4
|
import type { Dictionary } from '@crawlee/types';
|
|
5
|
-
import { type CheerioRoot } from '@crawlee/utils';
|
|
5
|
+
import { type CheerioRoot, type RobotsTxtFile } from '@crawlee/utils';
|
|
6
6
|
import type { DOMWindow } from 'jsdom';
|
|
7
7
|
import { VirtualConsole } from 'jsdom';
|
|
8
8
|
export type JSDOMErrorHandler<UserData extends Dictionary = any, // with default to Dictionary we cant use a typed router in untyped crawler
|
|
@@ -122,6 +122,8 @@ export declare class JSDOMCrawler extends HttpCrawler<JSDOMCrawlingContext> {
|
|
|
122
122
|
statusMessageCallback: import("ow").Predicate<Function> & import("ow").BasePredicate<Function | undefined>;
|
|
123
123
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
124
124
|
retryOnBlocked: import("ow").BooleanPredicate & import("ow").BasePredicate<boolean | undefined>;
|
|
125
|
+
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
126
|
+
respectRobotsTxtFile: import("ow").BooleanPredicate & import("ow").BasePredicate<boolean | undefined>;
|
|
125
127
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
126
128
|
httpClient: import("ow").ObjectPredicate<object> & import("ow").BasePredicate<object | undefined>;
|
|
127
129
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
@@ -173,12 +175,13 @@ interface EnqueueLinksInternalOptions {
|
|
|
173
175
|
options?: EnqueueLinksOptions;
|
|
174
176
|
window: DOMWindow | null;
|
|
175
177
|
requestQueue: RequestProvider;
|
|
178
|
+
robotsTxtFile?: RobotsTxtFile;
|
|
176
179
|
originalRequestUrl: string;
|
|
177
180
|
finalRequestUrl?: string;
|
|
178
181
|
}
|
|
179
182
|
/** @internal */
|
|
180
183
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
181
|
-
export declare function domCrawlerEnqueueLinks({ options, window, requestQueue, originalRequestUrl, finalRequestUrl, }: EnqueueLinksInternalOptions): Promise<import("@crawlee/types").BatchAddRequestsResult>;
|
|
184
|
+
export declare function domCrawlerEnqueueLinks({ options, window, requestQueue, robotsTxtFile, originalRequestUrl, finalRequestUrl, }: EnqueueLinksInternalOptions): Promise<import("@crawlee/types").BatchAddRequestsResult>;
|
|
182
185
|
/**
|
|
183
186
|
* Creates new {@apilink Router} instance that works based on request labels.
|
|
184
187
|
* This instance can then serve as a `requestHandler` of your {@apilink JSDOMCrawler}.
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"jsdom-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/jsdom-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,
|
|
1
|
+
{"version":3,"file":"jsdom-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/jsdom-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAEjD,OAAO,KAAK,EACR,aAAa,EACb,mBAAmB,EACnB,YAAY,EACZ,sBAAsB,EACtB,kBAAkB,EAClB,2BAA2B,EAC3B,gBAAgB,EAChB,cAAc,EACd,eAAe,EACf,YAAY,EACf,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;AAE7E,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,OAAO,CAAC;AACvC,OAAO,EAAyB,cAAc,EAAE,MAAM,OAAO,CAAC;AAM9D,MAAM,MAAM,iBAAiB,CACzB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,YAAY,CAAC,oBAAoB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE3D,MAAM,WAAW,mBAAmB,CAChC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,kBAAkB,CAAC,oBAAoB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;IAClE;;OAEG;IACH,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB;;OAEG;IACH,mBAAmB,CAAC,EAAE,OAAO,CAAC;CACjC;AAED,MAAM,MAAM,SAAS,CACjB,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,gBAAgB,CAAC,oBAAoB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAE/D,MAAM,WAAW,oBAAoB,CACjC,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,CACnC,SAAQ,2BAA2B,CAAC,QAAQ,EAAE,QAAQ,EAAE,YAAY,CAAC;IACnE,MAAM,EAAE,SAAS,CAAC;IAClB,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,mBAAmB,CAC3B,QAAQ,SAAS,UAAU,GAAG,GAAG,EAAE,2EAA2E;AAC9G,QAAQ,SAAS,UAAU,GAAG,GAAG,IACjC,cAAc,CAAC,oBAAoB,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;AAkF7D,qBAAa,YAAa,SAAQ,WAAW,CAAC,oBAAoB,CAAC;IAC/D,iBAA0B,YAAY;;;mCAyIlC,IAAI,+BAA+B,IAAI;sCACN,IAAI,2BAExB,IAAI;gCAEL,IAAL,4BACK,IAAV;oCAA2E,IAAG;wCAIvE,IAAI,2BAA2B,IAAI;sCAGnB,IAAG,2BAA2B,IAAI;0EAGvC,IAAI;yCACgB,IAAI,4BACxC,IACZ;+CAKI,IAAC;2CAIoB,IAAI;mCAEJ,IAAI,mCACA,IAAI;oCAKxB,IAAI,mCACkB,IAAI;4BAItB,IAAI,mCAER,IAAI;6BAGL,IAAC,mCAAmC,IAAI;+BAGxB,IAAI,+BAA+B,IACzD;sCAEQ,IAAI,+BACG,IAAI;0CAIb,IAAE,2BACW,IAAI;yCAErB,IAAE,2BAGE,IAAL;6BAGO,IAAI,+BAGd,IAEC;qCAIY,IAAI,+BAA+B,IAAI;4CACd,IAAI,+BAClC,IAAG;kCACwC,IAAI,2BAClC,IAAI;oCAET,IAAG,2BACP,IACV;oCAEY,IAAI,2BAA2B,IAAI;oCAIP,IAAG,2BAA2B,IAAI;sCACH,IAAI,mCACpD,IAAI;mCAEiB,IAAI,mCAE7C,IACL;+BAEG,IAAI,4BAA4B,IACnC;6CAGC,IAAG,2BAA2B,IAAI;sCAIP,IAAI,+BACP,IAAI;+BAKH,IAAI,4BACP,IAAI;qCACU,IAAI,4BAA4B,IAAI;2BAE3D,IAAI,mCAEV,IAAI;+BAAoE,IAAI,2BAA2B,IAAI;+BAAoE,IAAI,2BAA2B,IAAI;qCAA0E,IAAI,2BAA2B,IAAI;0BAA+D,IAAI,4BAA4B,IAAI;oBAA0D,IAAI,mCAAmC,IAAI;4BAAiE,IAAI,mCAAmC,IAAI;kCAAuE,IAAI,mCAAmC,IAAI;MA3PvuB;IAEF,SAAS,CAAC,UAAU,EAAE,OAAO,CAAC;IAC9B,SAAS,CAAC,mBAAmB,EAAE,OAAO,CAAC;IACvC,SAAS,CAAC,cAAc,EAAE,cAAc,GAAG,IAAI,CAAQ;gBAE3C,OAAO,GAAE,mBAAwB,EAAE,MAAM,CAAC,EAAE,aAAa;IASrE;;;;;;;;;;;;;OAaG;IACH,iBAAiB;IAgBjB,OAAO,CAAC,QAAQ,CAAC,iBAAiB,CAAuE;cAEhF,eAAe,CAAC,OAAO,EAAE,oBAAoB;cAK7C,UAAU,CAC/B,QAAQ,EAAE,eAAe,EACzB,KAAK,EAAE,OAAO,EACd,eAAe,EAAE,oBAAoB;;;;wCAgEK,mBAAmB;;IAalD,kBAAkB,CAAC,OAAO,EAAE,oBAAoB;CA0BlE;AAED,UAAU,2BAA2B;IACjC,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,MAAM,EAAE,SAAS,GAAG,IAAI,CAAC;IACzB,YAAY,EAAE,eAAe,CAAC;IAC9B,aAAa,CAAC,EAAE,aAAa,CAAC;IAC9B,kBAAkB,EAAE,MAAM,CAAC;IAC3B,eAAe,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,gBAAgB;AAChB,wBAAsB,sBAAsB,CAAC,EACzC,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,kBAAkB,EAClB,eAAe,GAClB,EAAE,2BAA2B,4DAyB7B;AAmBD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,iBAAiB,CAC7B,OAAO,SAAS,oBAAoB,GAAG,oBAAoB,EAC3D,QAAQ,SAAS,UAAU,GAAG,sBAAsB,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,EAC1E,MAAM,CAAC,EAAE,YAAY,CAAC,OAAO,EAAE,QAAQ,CAAC,kDAEzC"}
|
|
@@ -4,13 +4,13 @@ exports.JSDOMCrawler = void 0;
|
|
|
4
4
|
exports.domCrawlerEnqueueLinks = domCrawlerEnqueueLinks;
|
|
5
5
|
exports.createJSDOMRouter = createJSDOMRouter;
|
|
6
6
|
const tslib_1 = require("tslib");
|
|
7
|
-
const timeout_1 = require("@apify/timeout");
|
|
8
|
-
const utilities_1 = require("@apify/utilities");
|
|
9
7
|
const http_1 = require("@crawlee/http");
|
|
10
8
|
const utils_1 = require("@crawlee/utils");
|
|
11
9
|
const cheerio = tslib_1.__importStar(require("cheerio"));
|
|
12
10
|
const jsdom_1 = require("jsdom");
|
|
13
11
|
const ow_1 = tslib_1.__importDefault(require("ow"));
|
|
12
|
+
const timeout_1 = require("@apify/timeout");
|
|
13
|
+
const utilities_1 = require("@apify/utilities");
|
|
14
14
|
/**
|
|
15
15
|
* Provides a framework for the parallel crawling of web pages using plain HTTP requests and
|
|
16
16
|
* [jsdom](https://www.npmjs.com/package/jsdom) JSDOM implementation.
|
|
@@ -206,6 +206,7 @@ class JSDOMCrawler extends http_1.HttpCrawler {
|
|
|
206
206
|
options: enqueueOptions,
|
|
207
207
|
window,
|
|
208
208
|
requestQueue: await this.getRequestQueue(),
|
|
209
|
+
robotsTxtFile: await this.getRobotsTxtFileForUrl(crawlingContext.request.url),
|
|
209
210
|
originalRequestUrl: crawlingContext.request.url,
|
|
210
211
|
finalRequestUrl: crawlingContext.request.loadedUrl,
|
|
211
212
|
});
|
|
@@ -218,7 +219,8 @@ class JSDOMCrawler extends http_1.HttpCrawler {
|
|
|
218
219
|
if ($(selector).get().length === 0) {
|
|
219
220
|
if (timeoutMs) {
|
|
220
221
|
await (0, utils_1.sleep)(50);
|
|
221
|
-
|
|
222
|
+
await context.waitForSelector(selector, Math.max(timeoutMs - 50, 0));
|
|
223
|
+
return;
|
|
222
224
|
}
|
|
223
225
|
throw new Error(`Selector '${selector}' not found.`);
|
|
224
226
|
}
|
|
@@ -245,7 +247,7 @@ Object.defineProperty(JSDOMCrawler, "optionsShape", {
|
|
|
245
247
|
}
|
|
246
248
|
});
|
|
247
249
|
/** @internal */
|
|
248
|
-
async function domCrawlerEnqueueLinks({ options, window, requestQueue, originalRequestUrl, finalRequestUrl, }) {
|
|
250
|
+
async function domCrawlerEnqueueLinks({ options, window, requestQueue, robotsTxtFile, originalRequestUrl, finalRequestUrl, }) {
|
|
249
251
|
if (!window) {
|
|
250
252
|
throw new Error('Cannot enqueue links because the JSDOM is not available.');
|
|
251
253
|
}
|
|
@@ -258,6 +260,7 @@ async function domCrawlerEnqueueLinks({ options, window, requestQueue, originalR
|
|
|
258
260
|
const urls = extractUrlsFromWindow(window, options?.selector ?? 'a', options?.baseUrl ?? finalRequestUrl ?? originalRequestUrl);
|
|
259
261
|
return (0, http_1.enqueueLinks)({
|
|
260
262
|
requestQueue,
|
|
263
|
+
robotsTxtFile,
|
|
261
264
|
urls,
|
|
262
265
|
baseUrl,
|
|
263
266
|
...options,
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"jsdom-crawler.js","sourceRoot":"","sources":["../../src/internals/jsdom-crawler.ts"],"names":[],"mappings":";;;
|
|
1
|
+
{"version":3,"file":"jsdom-crawler.js","sourceRoot":"","sources":["../../src/internals/jsdom-crawler.ts"],"names":[],"mappings":";;;AAgWA,wDAgCC;AA2CD,8CAKC;;AAlaD,wCAMuB;AAEvB,0CAA6E;AAC7E,yDAAmC;AAEnC,iCAA8D;AAC9D,oDAAoB;AAEpB,4CAAqD;AACrD,gDAAwD;AAoExD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAwEG;AACH,MAAM,SAAS,GAAG,IAAI,sBAAc,CAAC;IACjC,6EAA6E;IAC7E,kDAAkD;IAClD,SAAS,EACL,uHAAuH;CAC9H,CAAC,CAAC;AAEH,MAAa,YAAa,SAAQ,kBAAiC;IAW/D,YAAY,UAA+B,EAAE,EAAE,MAAsB;QACjE,MAAM,EAAE,UAAU,GAAG,KAAK,EAAE,mBAAmB,GAAG,KAAK,EAAE,GAAG,WAAW,EAAE,GAAG,OAAO,CAAC;QAEpF,KAAK,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;QAPrB;;;;;WAAoB;QACpB;;;;;WAA6B;QAC7B;;;;mBAAwC,IAAI;WAAC;QAyCtC;;;;mBAAoB,CAAC,KAAY,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,0BAA0B,EAAE,KAAK,CAAC;WAAC;QAlCrG,IAAI,CAAC,UAAU,GAAG,UAAU,CAAC;QAC7B,IAAI,CAAC,mBAAmB,GAAG,mBAAmB,CAAC;IACnD,CAAC;IAED;;;;;;;;;;;;;OAaG;IACH,iBAAiB;QACb,IAAI,IAAI,CAAC,cAAc,EAAE,CAAC;YACtB,OAAO,IAAI,CAAC,cAAc,CAAC;QAC/B,CAAC;QAED,IAAI,CAAC,cAAc,GAAG,IAAI,sBAAc,EAAE,CAAC;QAE3C,IAAI,CAAC,IAAI,CAAC,mBAAmB,EAAE,CAAC;YAC5B,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC,OAAO,EAAE,EAAE,eAAe,EAAE,IAAI,EAAE,CAAC,CAAC;QACnE,CAAC;QAED,IAAI,CAAC,cAAc,CAAC,EAAE,CAAC,YAAY,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC;QAE7D,OAAO,IAAI,CAAC,cAAc,CAAC;IAC/B,CAAC;IAIkB,KAAK,CAAC,eAAe,CAAC,OAA6B;QAClE,IAAI,CAAC,iBAAiB,EAAE,CAAC,GAAG,CAAC,YAAY,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC;QACnE,OAAO,CAAC,MAAM,EAAE,KAAK,EAAE,CAAC;IAC5B,CAAC;IAEkB,KAAK,CAAC,UAAU,CAC/B,QAAyB,EACzB,KAAc,EACd,eAAqC;QAErC,MAAM,IAAI,GAAG,MAAM,IAAA,gCAAoB,EAAC,QAAQ,CAAC,CAAC;QAElD,MAAM,EAAE,MAAM,EAAE,GAAG,IAAI,aAAK,CAAC,IAAI,EAAE;YAC/B,GAAG,EAAE,QAAQ,CAAC,GAAG;YACjB,WAAW,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,WAAW;YAC7C,UAAU,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,SAAS;YACvD,SAAS;YACT,cAAc,EAAE,IAAI,CAAC,iBAAiB,EAAE;YACxC,iBAAiB,EAAE,IAAI;SAC1B,CAAC,CAAC;QAEH,kEAAkE;QAClE,MAAM,CAAC,cAAc,CAAC,MAAM,EAAE,YAAY,EAAE;YACxC,QAAQ,EAAE,IAAI;YACd,KAAK,EAAE,CAAC,KAAc,EAAO,EAAE,CAAC,CAAC;gBAC7B,OAAO,EAAE,KAAK;gBACd,KAAK,EAAE,KAAK;gBACZ,QAAQ,EAAE,IAAI;gBACd,WAAW,EAAE,GAAG,EAAE,GAAE,CAAC;gBACrB,cAAc,EAAE,GAAG,EAAE,GAAE,CAAC;gBACxB,gBAAgB,EAAE,GAAG,EAAE,GAAE,CAAC;gBAC1B,mBAAmB,EAAE,GAAG,EAAE,GAAE,CAAC;gBAC7B,aAAa,EAAE,GAAG,EAAE,GAAE,CAAC;aAC1B,CAAC;SACL,CAAC,CAAC;QACH,MAAM,CAAC,QAAQ,CAAC,WAAW,GAAG,GAAG,EAAE;YAC/B,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;YACjC,KAAK,CAAC,qBAAqB,GAAG,GAAG,EAAE,CAAC,CAAC,EAAE,CAAQ,CAAC;YAChD,KAAK,CAAC,cAAc,GAAG,GAAG,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC,IAAW,EAAE,MAAM,EAAE,CAAC,EAAE,CAAQ,CAAC;YAC7E,OAAO,KAAK,CAAC;QACjB,CAAC,CAAC;QAEF,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YAClB,IAAI,CAAC;gBACD,MAAM,IAAA,6BAAmB,EACrB,KAAK,IAAI,EAAE;oBACP,OAAO,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE;wBACjC,MAAM,CAAC,gBAAgB,CACnB,MAAM,EACN,GAAG,EAAE;4BACD,OAAO,EAAE,CAAC;wBACd,CAAC,EACD,KAAK,CACR,CAAC;oBACN,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC;gBACf,CAAC,EACD,KAAM,EACN,+CAA+C,CAClD,CAAC,KAAK,EAAE,CAAC;YACd,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACT,IAAI,CAAC,GAAG,CAAC,KAAK,CAAE,CAAW,CAAC,OAAO,CAAC,CAAC;YACzC,CAAC;QACL,CAAC;QAED,OAAO;YACH,MAAM;YACN,IAAI,IAAI;gBACJ,OAAO,MAAM,CAAC,QAAQ,CAAC,eAAe,CAAC,SAAS,CAAC;YACrD,CAAC;YACD,IAAI,QAAQ;gBACR,OAAO,MAAM,CAAC,QAAQ,CAAC;YAC3B,CAAC;YACD,YAAY,EAAE,KAAK,EAAE,cAAoC,EAAE,EAAE;gBACzD,OAAO,sBAAsB,CAAC;oBAC1B,OAAO,EAAE,cAAc;oBACvB,MAAM;oBACN,YAAY,EAAE,MAAM,IAAI,CAAC,eAAe,EAAE;oBAC1C,aAAa,EAAE,MAAM,IAAI,CAAC,sBAAsB,CAAC,eAAe,CAAC,OAAO,CAAC,GAAG,CAAC;oBAC7E,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,OAA6B;QAC3D,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;;AAlKL,oCAmKC;AAlK6B;;;;WAAe;QACrC,GAAG,kBAAW,CAAC,YAAY;QAC3B,UAAU,EAAE,YAAE,CAAC,QAAQ,CAAC,OAAO;QAC/B,mBAAmB,EAAE,YAAE,CAAC,QAAQ,CAAC,OAAO;KAC3C;EAJqC,CAIpC;AAyKN,gBAAgB;AACT,KAAK,UAAU,sBAAsB,CAAC,EACzC,OAAO,EACP,MAAM,EACN,YAAY,EACZ,aAAa,EACb,kBAAkB,EAClB,eAAe,GACW;IAC1B,IAAI,CAAC,MAAM,EAAE,CAAC;QACV,MAAM,IAAI,KAAK,CAAC,0DAA0D,CAAC,CAAC;IAChF,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,IAAI;QACJ,OAAO;QACP,GAAG,OAAO;KACb,CAAC,CAAC;AACP,CAAC;AAED;;;GAGG;AACH,SAAS,qBAAqB,CAAC,MAAiB,EAAE,QAAgB,EAAE,OAAe;IAC/E,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,iBAAiB,CAG/B,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/jsdom",
|
|
3
|
-
"version": "3.13.
|
|
3
|
+
"version": "3.13.2-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
6
|
"node": ">=16.0.0"
|
|
@@ -55,9 +55,9 @@
|
|
|
55
55
|
"dependencies": {
|
|
56
56
|
"@apify/timeout": "^0.3.0",
|
|
57
57
|
"@apify/utilities": "^2.7.10",
|
|
58
|
-
"@crawlee/http": "3.13.
|
|
59
|
-
"@crawlee/types": "3.13.
|
|
60
|
-
"@crawlee/utils": "3.13.
|
|
58
|
+
"@crawlee/http": "^3.13.2-beta.0",
|
|
59
|
+
"@crawlee/types": "^3.13.2-beta.0",
|
|
60
|
+
"@crawlee/utils": "^3.13.2-beta.0",
|
|
61
61
|
"@types/jsdom": "^21.0.0",
|
|
62
62
|
"cheerio": "1.0.0-rc.12",
|
|
63
63
|
"jsdom": "^26.0.0",
|
|
@@ -71,5 +71,5 @@
|
|
|
71
71
|
}
|
|
72
72
|
}
|
|
73
73
|
},
|
|
74
|
-
"gitHead": "
|
|
74
|
+
"gitHead": "5decdceeb4642c5fff872f9ac27e78bb9485735a"
|
|
75
75
|
}
|