@crawlee/playwright 4.0.0-beta.223 → 4.0.0-beta.225

package/internals/playwright-browser-pool.d.ts

@@ -1,8 +1,16 @@
-import type { Configuration } from '@crawlee/browser';
-import type { BrowserPool, BrowserPoolHooks, BrowserPoolOptions, PlaywrightPlugin, RemoteBrowserPool, RemoteBrowserPoolOptions } from '@crawlee/browser-pool';
+import type { Configuration, CrawlerRemoteBrowserOptions } from '@crawlee/browser';
+import type { BrowserPool, BrowserPoolHooks, BrowserPoolOptions, PlaywrightPlugin } from '@crawlee/browser-pool';
 // @ts-ignore optional peer dependency or compatibility with es2022
 import type { Page } from 'playwright';
 import type { PlaywrightLaunchContext } from './playwright-launcher.js';
+/** {@link BrowserCrawlerOptions.remoteBrowser|`remoteBrowser`} for Playwright, with the protocol to connect over. */
+export interface PlaywrightRemoteBrowserOptions extends CrawlerRemoteBrowserOptions {
+    /**
+     * `'cdp'` uses `connectOverCDP()` (the default), `'playwright'` uses `connect()` (Playwright's own WebSocket
+     * protocol, for endpoints started with `launchServer()`).
+     */
+    protocol?: 'cdp' | 'playwright';
+}
 /** A {@link BrowserPool} of Playwright browsers, as built by {@link playwrightBrowserPool}. */
 export type PlaywrightBrowserPool = BrowserPool<{
     browserPlugins: [PlaywrightPlugin];
@@ -17,8 +25,8 @@ export interface PlaywrightBrowserPoolOptions extends Omit<BrowserPoolOptions, '
     headless?: boolean;
     /** Configuration to read the browser defaults from. Defaults to the global configuration. */
     configuration?: Configuration;
-}
-export interface RemotePlaywrightBrowserPoolOptions extends Pick<PlaywrightBrowserPoolOptions, 'launchContext' | 'headless' | 'configuration'>, Omit<RemoteBrowserPoolOptions, 'browserPlugins'> {
+    /** Connect to a remote browser service instead of launching locally; see {@link BrowserCrawlerOptions.remoteBrowser}. */
+    remoteBrowser?: PlaywrightRemoteBrowserOptions;
 }
 /**
  * Builds a {@link BrowserPool} of Playwright browsers to pass to a {@link PlaywrightCrawler} as its
@@ -45,27 +53,3 @@ export interface RemotePlaywrightBrowserPoolOptions extends Pick<PlaywrightBrows
  * @category Browser management
  */
 export declare function playwrightBrowserPool(options?: PlaywrightBrowserPoolOptions): PlaywrightBrowserPool;
-/**
- * The {@link RemoteBrowserPool} counterpart of {@link playwrightBrowserPool}: connects to a remote browser
- * service (Browserbase, Browserless, Steel, ...) with a Playwright plugin derived from `launchContext`.
- *
- * A {@link PlaywrightCrawler} accepts the same connection details directly via
- * {@link BrowserCrawlerOptions.remoteBrowser|`remoteBrowser`}; reach for this factory when you also need to
- * tune the wrapping pool, or to share one remote pool between crawlers.
- *
- * **Example usage:**
- *
- * ```javascript
- * const crawler = new PlaywrightCrawler({
- *     browserPool: remotePlaywrightBrowserPool({
- *         endpoint: 'wss://production-sfo.browserless.io?token=xxx',
- *         maxOpenBrowsers: 2,
- *         browserPoolOptions: { useFingerprints: false },
- *     }),
- *     requestHandler: async ({ page }) => { ... },
- * });
- * ```
- *
- * @category Browser management
- */
-export declare function remotePlaywrightBrowserPool(options: RemotePlaywrightBrowserPoolOptions): RemoteBrowserPool<Page>;

package/internals/playwright-browser-pool.js

@@ -27,33 +27,6 @@ export function playwrightBrowserPool(options = {}) {
     const { launchContext, headless, configuration, ...poolOptions } = options;
     return playwrightLauncher(launchContext, headless, configuration).createBrowserPool(poolOptions);
 }
-/**
- * The {@link RemoteBrowserPool} counterpart of {@link playwrightBrowserPool}: connects to a remote browser
- * service (Browserbase, Browserless, Steel, ...) with a Playwright plugin derived from `launchContext`.
- *
- * A {@link PlaywrightCrawler} accepts the same connection details directly via
- * {@link BrowserCrawlerOptions.remoteBrowser|`remoteBrowser`}; reach for this factory when you also need to
- * tune the wrapping pool, or to share one remote pool between crawlers.
- *
- * **Example usage:**
- *
- * ```javascript
- * const crawler = new PlaywrightCrawler({
- *     browserPool: remotePlaywrightBrowserPool({
- *         endpoint: 'wss://production-sfo.browserless.io?token=xxx',
- *         maxOpenBrowsers: 2,
- *         browserPoolOptions: { useFingerprints: false },
- *     }),
- *     requestHandler: async ({ page }) => { ... },
- * });
- * ```
- *
- * @category Browser management
- */
-export function remotePlaywrightBrowserPool(options) {
-    const { launchContext, headless, configuration, ...remoteOptions } = options;
-    return playwrightLauncher(launchContext, headless, configuration).createRemoteBrowserPool(remoteOptions);
-}
 function playwrightLauncher(launchContext = {}, headless, configuration) {
     return new PlaywrightLauncher(headless == null
         ? launchContext

package/internals/playwright-crawler.d.ts

@@ -4,6 +4,7 @@ import type { Dictionary } from '@crawlee/types';
 // @ts-ignore optional peer dependency or compatibility with es2022
 import type { Page, Response } from 'playwright';
 import { z } from 'zod';
+import { type PlaywrightRemoteBrowserOptions } from './playwright-browser-pool.js';
 import type { PlaywrightLaunchContext } from './playwright-launcher.js';
 import type { DirectNavigationOptions, HandleCloudflareChallengeOptions, PlaywrightContextUtils } from './utils/playwright-utils.js';
 export type PlaywrightGotoOptions = NonNullable<Parameters<Page['goto']>[1]>;
@@ -20,6 +21,8 @@ export interface PlaywrightCrawlerOptions<ContextExtension = Dictionary<never>,
      * Can be also set via {@link Configuration}.
      */
     headless?: boolean;
+    /** See {@link BrowserCrawlerOptions.remoteBrowser}; Playwright additionally lets you pick the `protocol`. */
+    remoteBrowser?: PlaywrightRemoteBrowserOptions;
     /**
      * Async functions that are sequentially evaluated before the navigation. Good for setting additional cookies
      * or browser properties before navigation. The function receives the `crawlingContext`; the options object

package/internals/playwright-crawler.js

@@ -1,7 +1,7 @@
 import { BrowserCrawler, RequestState, Router, serviceLocator } from '@crawlee/browser';
 import { assertBrowserPoolNotConfigured, parseArgument, schemas } from '@crawlee/utils/internal';
 import { z } from 'zod';
-import { playwrightBrowserPool, remotePlaywrightBrowserPool } from './playwright-browser-pool.js';
+import { playwrightBrowserPool } from './playwright-browser-pool.js';
 import { blockRequests, compileScript, enqueueLinksByClickingElements, gotoExtended, handleCloudflareChallenge, infiniteScroll, injectFile, injectJQuery, parseWithCheerio, saveSnapshot, } from './utils/playwright-utils.js';
 /**
  * Provides a simple framework for parallel crawling of web pages
@@ -99,9 +99,7 @@ export class PlaywrightCrawler extends BrowserCrawler {
         super({
             ...browserCrawlerOptions,
             configuration,
-            browserPoolBuilder: (remoteBrowser) => remoteBrowser
-                ? remotePlaywrightBrowserPool({ ...remoteBrowser, launchContext, headless, configuration })
-                : playwrightBrowserPool({ launchContext, headless, configuration }),
+            browserPoolBuilder: (remoteBrowser) => playwrightBrowserPool({ launchContext, headless, configuration, remoteBrowser }),
             contextPipelineBuilder: () => this.#buildContextPipeline(),
         });
     }

package/internals/playwright-launcher.js

@@ -1,5 +1,5 @@
 import { BrowserLauncher, Configuration } from '@crawlee/browser';
-import { PlaywrightPlugin } from '@crawlee/browser-pool';
+import { PlaywrightPlugin, RemotePlaywrightPlugin } from '@crawlee/browser-pool';
 import { parseArgument, schemas } from '@crawlee/utils/internal';
 import { z } from 'zod';
 /**
@@ -35,6 +35,7 @@ export class PlaywrightLauncher extends BrowserLauncher {
         }, configuration);
         this.configuration = configuration;
         this.Plugin = PlaywrightPlugin;
+        this.RemotePlugin = RemotePlaywrightPlugin;
     }
 }
 /**

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@crawlee/playwright",
-    "version": "4.0.0-beta.223",
+    "version": "4.0.0-beta.225",
     "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.",
     "engines": {
         "node": ">=22.13.0"
@@ -49,13 +49,13 @@
     "dependencies": {
         "@apify/datastructures": "^3.0.1",
         "@apify/timeout": "^1.0.1",
-        "@crawlee/basic": "4.0.0-beta.223",
-        "@crawlee/browser": "4.0.0-beta.223",
-        "@crawlee/browser-pool": "4.0.0-beta.223",
-        "@crawlee/cheerio": "4.0.0-beta.223",
-        "@crawlee/core": "4.0.0-beta.223",
-        "@crawlee/types": "4.0.0-beta.223",
-        "@crawlee/utils": "4.0.0-beta.223",
+        "@crawlee/basic": "4.0.0-beta.225",
+        "@crawlee/browser": "4.0.0-beta.225",
+        "@crawlee/browser-pool": "4.0.0-beta.225",
+        "@crawlee/cheerio": "4.0.0-beta.225",
+        "@crawlee/core": "4.0.0-beta.225",
+        "@crawlee/types": "4.0.0-beta.225",
+        "@crawlee/utils": "4.0.0-beta.225",
         "cheerio": "^1.0.0",
         "jquery": "^3.7.1",
         "ml-logistic-regression": "^2.0.0",
@@ -79,5 +79,5 @@
             }
         }
     },
-    "gitHead": "cfcbcbebeabd9ca2dfd7a17afe559db16307250d"
+    "gitHead": "5d71fa0f785d7dc5736d95bdf5f4ec907e5a1c42"
 }