@crawlee/puppeteer 4.0.0-beta.31 → 4.0.0-beta.32
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.
|
@@ -4,8 +4,10 @@ import type { PuppeteerController, PuppeteerPlugin } from '@crawlee/browser-pool
|
|
|
4
4
|
import type { Dictionary } from '@crawlee/types';
|
|
5
5
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
6
6
|
import type { HTTPResponse, LaunchOptions, Page } from 'puppeteer';
|
|
7
|
+
import type { EnqueueLinksByClickingElementsOptions } from './enqueue-links/click-elements.js';
|
|
7
8
|
import type { PuppeteerLaunchContext } from './puppeteer-launcher.js';
|
|
8
|
-
import type {
|
|
9
|
+
import type { InterceptHandler } from './utils/puppeteer_request_interception.js';
|
|
10
|
+
import type { BlockRequestsOptions, DirectNavigationOptions, InfiniteScrollOptions, InjectFileOptions, PuppeteerContextUtils, SaveSnapshotOptions } from './utils/puppeteer_utils.js';
|
|
9
11
|
export interface PuppeteerCrawlingContext<UserData extends Dictionary = Dictionary> extends BrowserCrawlingContext<Page, HTTPResponse, PuppeteerController, UserData>, PuppeteerContextUtils {
|
|
10
12
|
}
|
|
11
13
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
@@ -208,6 +210,24 @@ export declare class PuppeteerCrawler<ContextExtension = Dictionary<never>, Exte
|
|
|
208
210
|
* All `PuppeteerCrawler` parameters are passed via an options object.
|
|
209
211
|
*/
|
|
210
212
|
constructor(options?: PuppeteerCrawlerOptions<ContextExtension, ExtendedContext>);
|
|
213
|
+
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
214
|
+
protected buildContextPipeline(): import("@crawlee/browser").ContextPipeline<import("@crawlee/browser").CrawlingContext<Dictionary>, BrowserCrawlingContext<Page, HTTPResponse, PuppeteerController, Dictionary> & {
|
|
215
|
+
injectFile: (filePath: string, options?: InjectFileOptions) => Promise<unknown>;
|
|
216
|
+
injectJQuery: () => Promise<void>;
|
|
217
|
+
waitForSelector: (selector: string, timeoutMs?: number) => Promise<void>;
|
|
218
|
+
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
219
|
+
parseWithCheerio: (selector?: string, timeoutMs?: number) => Promise<import("@crawlee/browser").CheerioAPI>;
|
|
220
|
+
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
221
|
+
enqueueLinksByClickingElements: (options: Omit<EnqueueLinksByClickingElementsOptions, "page" | "requestQueue">) => Promise<import("@crawlee/types").BatchAddRequestsResult>;
|
|
222
|
+
blockRequests: (options?: BlockRequestsOptions) => Promise<void>;
|
|
223
|
+
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
224
|
+
compileScript: (scriptString: string, ctx?: Dictionary) => import("./utils/puppeteer_utils.js").CompiledScriptFunction;
|
|
225
|
+
addInterceptRequestHandler: (handler: InterceptHandler) => Promise<void>;
|
|
226
|
+
removeInterceptRequestHandler: (handler: InterceptHandler) => Promise<void>;
|
|
227
|
+
infiniteScroll: (options?: InfiniteScrollOptions) => Promise<void>;
|
|
228
|
+
saveSnapshot: (options?: SaveSnapshotOptions) => Promise<void>;
|
|
229
|
+
closeCookieModals: () => Promise<void>;
|
|
230
|
+
}>;
|
|
211
231
|
private enhanceContext;
|
|
212
232
|
protected _navigationHandler(crawlingContext: PuppeteerCrawlingContext, gotoOptions: DirectNavigationOptions): Promise<HTTPResponse | null>;
|
|
213
233
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"puppeteer-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/puppeteer-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACR,qBAAqB,EACrB,sBAAsB,EACtB,WAAW,EACX,sBAAsB,EACtB,YAAY,EACf,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,cAAc,EAAwB,MAAM,kBAAkB,CAAC;AACxE,OAAO,KAAK,EAAsB,mBAAmB,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAEtG,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAEjD,OAAO,KAAK,EAAE,YAAY,EAAE,aAAa,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;
|
|
1
|
+
{"version":3,"file":"puppeteer-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/puppeteer-crawler.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACR,qBAAqB,EACrB,sBAAsB,EACtB,WAAW,EACX,sBAAsB,EACtB,YAAY,EACf,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,cAAc,EAAwB,MAAM,kBAAkB,CAAC;AACxE,OAAO,KAAK,EAAsB,mBAAmB,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAEtG,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,gBAAgB,CAAC;AAEjD,OAAO,KAAK,EAAE,YAAY,EAAE,aAAa,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAEnE,OAAO,KAAK,EAAE,qCAAqC,EAAE,MAAM,mCAAmC,CAAC;AAC/F,OAAO,KAAK,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAC;AAEtE,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,2CAA2C,CAAC;AAClF,OAAO,KAAK,EACR,oBAAoB,EACpB,uBAAuB,EACvB,qBAAqB,EACrB,iBAAiB,EACjB,qBAAqB,EACrB,mBAAmB,EACtB,MAAM,4BAA4B,CAAC;AAGpC,MAAM,WAAW,wBAAwB,CAAC,QAAQ,SAAS,UAAU,GAAG,UAAU,CAC9E,SAAQ,sBAAsB,CAAC,IAAI,EAAE,YAAY,EAAE,mBAAmB,EAAE,QAAQ,CAAC,EAC7E,qBAAqB;CAAG;AAChC,MAAM,WAAW,aAAc,SAAQ,WAAW,CAAC,wBAAwB,EAAE,oBAAoB,CAAC;CAAG;AACrG,MAAM,MAAM,oBAAoB,GAAG,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAE/D,MAAM,WAAW,uBAAuB,CACpC,gBAAgB,GAAG,UAAU,CAAC,KAAK,CAAC,EACpC,eAAe,SAAS,wBAAwB,GAAG,wBAAwB,GAAG,gBAAgB,CAChG,SAAQ,qBAAqB,CACvB,IAAI,EACJ,YAAY,EACZ,mBAAmB,EACnB,wBAAwB,EACxB,gBAAgB,EAChB,eAAe,EACf;IAAE,cAAc,EAAE,CAAC,eAAe,CAAC,CAAA;CAAE,CACxC;IACD;;OAEG;IACH,aAAa,CAAC,EAAE,sBAAsB,CAAC;IAEvC;;;;;;;;;;;;;;;;OAgBG;IACH,kBAAkB,CAAC,EAAE,aAAa,EAAE,CAAC;IAErC;;;;;;;;;;;;;;OAcG;IACH,mBAAmB,CAAC,EAAE,aAAa,EAAE,CAAC;CACzC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8DG;AACH,qBAAa,gBAAgB,CACzB,gBAAgB,GAAG,UAAU,CAAC,KAAK,CAAC,EACpC,eAAe,SAAS,wBAAwB,GAAG,wBAAwB,GAAG,gBAAgB,CAChG,SAAQ,cAAc,CACpB,IAAI,EACJ,YAAY,EACZ,mBAAmB,EACnB;IAAE,cAAc,EAAE,CAAC,eAAe,CAAC,CAAA;CAAE,EACrC,aAAa,EACb,wBAAwB,EACxB,gBAAgB,EAChB,eAAe,CAClB;IACG,iBAA0B,YAAY;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAGpC;IAEF;;OAEG;gBACS,OAAO,GAAE,uBAAuB,CAAC,gBAAgB,EAAE,eAAe,CAAM;cAqDjE,oBAAoB;+BAUF,MAAM,YAAY,iBAAiB;;oCAL3B,MAAM;sCAkBP,MAAM;kDAQ7B,IAAI,CAAC,qCAAqC,EAAE,MAAM,GAAG,cAAc,CAAC;kCAOjD,oBAAoB;sCAEtB,MAAM,QAAQ,UAAU;8CACV,gBAAgB;iDAEb,gBAAgB;mCAE9B,qBAAqB;iCAEvB,mBAAmB;;;YA3C5C,cAAc;cAiDH,kBAAkB,CACvC,eAAe,EAAE,wBAAwB,EACzC,WAAW,EAAE,uBAAuB;CAI3C;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAgB,qBAAqB,CACjC,OAAO,SAAS,wBAAwB,GAAG,wBAAwB,EACnE,QAAQ,SAAS,UAAU,GAAG,sBAAsB,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,EAC1E,MAAM,CAAC,EAAE,YAAY,CAAC,OAAO,EAAE,QAAQ,CAAC,qDAEzC"}
|
|
@@ -76,7 +76,7 @@ export class PuppeteerCrawler extends BrowserCrawler {
|
|
|
76
76
|
*/
|
|
77
77
|
constructor(options = {}) {
|
|
78
78
|
ow(options, 'PuppeteerCrawlerOptions', ow.object.exactShape(PuppeteerCrawler.optionsShape));
|
|
79
|
-
const { launchContext = {}, headless, proxyConfiguration, ...browserCrawlerOptions } = options;
|
|
79
|
+
const { launchContext = {}, headless, proxyConfiguration, contextPipelineBuilder, ...browserCrawlerOptions } = options;
|
|
80
80
|
const browserPoolOptions = {
|
|
81
81
|
...options.browserPoolOptions,
|
|
82
82
|
};
|
|
@@ -100,9 +100,12 @@ export class PuppeteerCrawler extends BrowserCrawler {
|
|
|
100
100
|
launchContext,
|
|
101
101
|
proxyConfiguration,
|
|
102
102
|
browserPoolOptions,
|
|
103
|
-
contextPipelineBuilder: () => this.buildContextPipeline()
|
|
103
|
+
contextPipelineBuilder: contextPipelineBuilder ?? (() => this.buildContextPipeline()),
|
|
104
104
|
});
|
|
105
105
|
}
|
|
106
|
+
buildContextPipeline() {
|
|
107
|
+
return super.buildContextPipeline().compose({ action: this.enhanceContext.bind(this) });
|
|
108
|
+
}
|
|
106
109
|
async enhanceContext(context) {
|
|
107
110
|
const waitForSelector = async (selector, timeoutMs = 5_000) => {
|
|
108
111
|
await context.page.waitForSelector(selector, { timeout: timeoutMs });
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"puppeteer-crawler.js","sourceRoot":"","sources":["../../src/internals/puppeteer-crawler.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAExE,OAAO,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAE/C,OAAO,EAAE,MAAM,IAAI,CAAC;AAKpB,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAU5D,OAAO,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AA8D1E;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8DG;AACH,MAAM,OAAO,gBAGX,SAAQ,cAST;IACa,MAAM,CAAU,YAAY,GAAG;QACrC,GAAG,cAAc,CAAC,YAAY;QAC9B,kBAAkB,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM;KACzC,CAAC;IAEF;;OAEG;IACH,YAAY,UAAsE,EAAE;QAChF,EAAE,CAAC,OAAO,EAAE,yBAAyB,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,gBAAgB,CAAC,YAAY,CAAC,CAAC,CAAC;QAE5F,MAAM,
|
|
1
|
+
{"version":3,"file":"puppeteer-crawler.js","sourceRoot":"","sources":["../../src/internals/puppeteer-crawler.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAExE,OAAO,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAE/C,OAAO,EAAE,MAAM,IAAI,CAAC;AAKpB,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAU5D,OAAO,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AA8D1E;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8DG;AACH,MAAM,OAAO,gBAGX,SAAQ,cAST;IACa,MAAM,CAAU,YAAY,GAAG;QACrC,GAAG,cAAc,CAAC,YAAY;QAC9B,kBAAkB,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM;KACzC,CAAC;IAEF;;OAEG;IACH,YAAY,UAAsE,EAAE;QAChF,EAAE,CAAC,OAAO,EAAE,yBAAyB,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,gBAAgB,CAAC,YAAY,CAAC,CAAC,CAAC;QAE5F,MAAM,EACF,aAAa,GAAG,EAAE,EAClB,QAAQ,EACR,kBAAkB,EAClB,sBAAsB,EACtB,GAAG,qBAAqB,EAC3B,GAAG,OAAO,CAAC;QAEZ,MAAM,kBAAkB,GAAG;YACvB,GAAG,OAAO,CAAC,kBAAkB;SACV,CAAC;QAExB,IAAI,aAAa,CAAC,QAAQ,EAAE,CAAC;YACzB,MAAM,IAAI,KAAK,CACX,oFAAoF;gBAChF,gDAAgD,CACvD,CAAC;QACN,CAAC;QAED,2EAA2E;QAC3E,uFAAuF;QACvF,IAAI,kBAAkB,CAAC,cAAc,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,sFAAsF,CAAC,CAAC;QAC5G,CAAC;QAED,IAAI,QAAQ,IAAI,IAAI,EAAE,CAAC;YACnB,aAAa,CAAC,aAAa,KAAK,EAAmB,CAAC;YACpD,aAAa,CAAC,aAAa,CAAC,QAAQ,GAAG,QAAmB,CAAC;QAC/D,CAAC;QAED,MAAM,iBAAiB,GAAG,IAAI,iBAAiB,CAAC,aAAa,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC;QAEtF,kBAAkB,CAAC,cAAc,GAAG,CAAC,iBAAiB,CAAC,mBAAmB,EAAE,CAAC,CAAC;QAE9E,KAAK,CAAC;YACF,GAAI,qBAOF;YACF,aAAa;YACb,kBAAkB;YAClB,kBAAkB;YAClB,sBAAsB,EAAE,sBAAsB,IAAI,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,oBAAoB,EAAE,CAAC;SACxF,CAAC,CAAC;IACP,CAAC;IAEkB,oBAAoB;QACnC,OAAO,KAAK,CAAC,oBAAoB,EAAE,CAAC,OAAO,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC5F,CAAC;IAEO,KAAK,CAAC,cAAc,CAAC,OAAwE;QACjG,MAAM,eAAe,GAAG,KAAK,EAAE,QAAgB,EAAE,SAAS,GAAG,KAAK,EAAE,EAAE;YAClE,MAAM,OAAO,CAAC,IAAI,CAAC,eAAe,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC,CAAC;QACzE,CAAC,CAAC;QAEF,OAAO;YACH,UAAU,EAAE,KAAK,EAAE,QAAgB,EAAE,OAA2B,EAAE,EAAE,CAChE,cAAc,CAAC,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,QAAQ,EAAE,OAAO,CAAC;YAC9D,YAAY,EAAE,KAAK,IAAI,EAAE;gBACrB,IAAI,OAAO,CAAC,OAAO,CAAC,KAAK,KAAK,YAAY,CAAC,UAAU,EAAE,CAAC;oBACpD,OAAO,CAAC,GAAG,CAAC,OAAO,CACf,mIAAmI,CACtI,CAAC;oBACF,MAAM,cAAc,CAAC,YAAY,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;oBAChD,OAAO;gBACX,CAAC;gBACD,MAAM,cAAc,CAAC,YAAY,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,kBAAkB,EAAE,KAAK,EAAE,CAAC,CAAC;YACnF,CAAC;YACD,eAAe;YACf,gBAAgB,EAAE,KAAK,EAAE,QAAiB,EAAE,SAAS,GAAG,KAAK,EAAE,EAAE;gBAC7D,IAAI,QAAQ,EAAE,CAAC;oBACX,MAAM,eAAe,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;gBAC/C,CAAC;gBAED,OAAO,cAAc,CAAC,gBAAgB,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,iBAAiB,EAAE,IAAI,CAAC,aAAa,CAAC,CAAC;YACrG,CAAC;YACD,8BAA8B,EAAE,KAAK,EACjC,OAA6E,EAC/E,EAAE,CACA,cAAc,CAAC,8BAA8B,CAAC;gBAC1C,IAAI,EAAE,OAAO,CAAC,IAAI;gBAClB,YAAY,EAAE,IAAI,CAAC,YAAa;gBAChC,GAAG,OAAO;aACb,CAAC;YACN,aAAa,EAAE,KAAK,EAAE,OAA8B,EAAE,EAAE,CACpD,cAAc,CAAC,aAAa,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC;YACvD,aAAa,EAAE,CAAC,YAAoB,EAAE,GAAgB,EAAE,EAAE,CAAC,cAAc,CAAC,aAAa,CAAC,YAAY,EAAE,GAAG,CAAC;YAC1G,0BAA0B,EAAE,KAAK,EAAE,OAAyB,EAAE,EAAE,CAC5D,cAAc,CAAC,0BAA0B,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC;YACpE,6BAA6B,EAAE,KAAK,EAAE,OAAyB,EAAE,EAAE,CAC/D,cAAc,CAAC,6BAA6B,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC;YACvE,cAAc,EAAE,KAAK,EAAE,OAA+B,EAAE,EAAE,CACtD,cAAc,CAAC,cAAc,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC;YACxD,YAAY,EAAE,KAAK,EAAE,OAA6B,EAAE,EAAE,CAClD,cAAc,CAAC,YAAY,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,GAAG,OAAO,EAAE,MAAM,EAAE,cAAc,CAAC,gBAAgB,EAAE,EAAE,CAAC;YACxG,iBAAiB,EAAE,KAAK,IAAI,EAAE,CAAC,cAAc,CAAC,iBAAiB,CAAC,OAAO,CAAC,IAAI,CAAC;SAChF,CAAC;IACN,CAAC;IAEkB,KAAK,CAAC,kBAAkB,CACvC,eAAyC,EACzC,WAAoC;QAEpC,OAAO,YAAY,CAAC,eAAe,CAAC,IAAI,EAAE,eAAe,CAAC,OAAO,EAAE,WAAW,CAAC,CAAC;IACpF,CAAC;;AAGL;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,MAAM,UAAU,qBAAqB,CAGnC,MAAwC;IACtC,OAAO,MAAM,CAAC,MAAM,CAAU,MAAM,CAAC,CAAC;AAC1C,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@crawlee/puppeteer",
|
|
3
|
-
"version": "4.0.0-beta.
|
|
3
|
+
"version": "4.0.0-beta.32",
|
|
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": ">=22.0.0"
|
|
@@ -49,10 +49,10 @@
|
|
|
49
49
|
"dependencies": {
|
|
50
50
|
"@apify/datastructures": "^2.0.3",
|
|
51
51
|
"@apify/log": "^2.5.18",
|
|
52
|
-
"@crawlee/browser": "4.0.0-beta.
|
|
53
|
-
"@crawlee/browser-pool": "4.0.0-beta.
|
|
54
|
-
"@crawlee/types": "4.0.0-beta.
|
|
55
|
-
"@crawlee/utils": "4.0.0-beta.
|
|
52
|
+
"@crawlee/browser": "4.0.0-beta.32",
|
|
53
|
+
"@crawlee/browser-pool": "4.0.0-beta.32",
|
|
54
|
+
"@crawlee/types": "4.0.0-beta.32",
|
|
55
|
+
"@crawlee/utils": "4.0.0-beta.32",
|
|
56
56
|
"cheerio": "^1.0.0",
|
|
57
57
|
"devtools-protocol": "*",
|
|
58
58
|
"idcac-playwright": "^0.1.3",
|
|
@@ -79,5 +79,5 @@
|
|
|
79
79
|
}
|
|
80
80
|
}
|
|
81
81
|
},
|
|
82
|
-
"gitHead": "
|
|
82
|
+
"gitHead": "86c266f094f0ac2bcf8c376512e16b677279a614"
|
|
83
83
|
}
|