@crawlee/basic 4.0.0-beta.10 → 4.0.0-beta.11

package/index.d.ts

@@ -1,4 +1,3 @@
 export * from '@crawlee/core';
 export * from './internals/basic-crawler.js';
-export * from './internals/constants.js';
 //# sourceMappingURL=index.d.ts.map

package/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,8BAA8B,CAAC;AAC7C,cAAc,0BAA0B,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,8BAA8B,CAAC"}

package/index.js

@@ -1,4 +1,3 @@
 export * from '@crawlee/core';
 export * from './internals/basic-crawler.js';
-export * from './internals/constants.js';
 //# sourceMappingURL=index.js.map

package/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,8BAA8B,CAAC;AAC7C,cAAc,0BAA0B,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,eAAe,CAAC;AAC9B,cAAc,8BAA8B,CAAC"}

package/internals/basic-crawler.d.ts

@@ -1,38 +1,14 @@
-import type { AddRequestsBatchedOptions, AddRequestsBatchedResult, AutoscaledPoolOptions, BaseHttpClient, CrawlingContext, DatasetExportOptions, EnqueueLinksOptions, EventManager, FinalStatistics, GetUserDataFromRequest, IRequestList, LoadedContext, ProxyInfo, Request, RequestOptions, RestrictedCrawlingContext, RouterHandler, RouterRoutes, Session, SessionPoolOptions, SkippedRequestCallback, Source, StatisticsOptions, StatisticState } from '@crawlee/core';
-import { AutoscaledPool, Configuration, Dataset, RequestProvider, SessionPool, Statistics } from '@crawlee/core';
-import type { Awaitable, BatchAddRequestsResult, Dictionary, SetStatusMessageOptions } from '@crawlee/types';
+import type { AddRequestsBatchedOptions, AddRequestsBatchedResult, AutoscaledPoolOptions, BaseHttpClient, CrawlingContext, DatasetExportOptions, EventManager, FinalStatistics, GetUserDataFromRequest, IRequestList, ProxyInfo, Request, RequestOptions, RouterHandler, RouterRoutes, Session, SessionPoolOptions, SkippedRequestCallback, Source, StatisticsOptions, StatisticState } from '@crawlee/core';
+import { AutoscaledPool, Configuration, ContextPipeline, Dataset, RequestProvider, SessionPool, Statistics } from '@crawlee/core';
+import type { Awaitable, Dictionary, SetStatusMessageOptions } from '@crawlee/types';
 import { RobotsTxtFile } from '@crawlee/utils';
-import type { SetRequired } from 'type-fest';
+import type { ReadonlyDeep } from 'type-fest';
 import type { Log } from '@apify/log';
 import { TimeoutError } from '@apify/timeout';
-export interface BasicCrawlingContext<UserData extends Dictionary = Dictionary> extends CrawlingContext<BasicCrawler, UserData> {
-    /**
-     * This function automatically finds and enqueues links from the current page, adding them to the {@link RequestQueue}
-     * currently used by the crawler.
-     *
-     * Optionally, the function allows you to filter the target links' URLs using an array of globs or regular expressions
-     * and override settings of the enqueued {@link Request} objects.
-     *
-     * Check out the [Crawl a website with relative links](https://crawlee.dev/js/docs/examples/crawl-relative-links) example
-     * for more details regarding its usage.
-     *
-     * **Example usage**
-     *
-     * ```ts
-     * async requestHandler({ enqueueLinks }) {
-     *     await enqueueLinks({
-     *       urls: [...],
-     *     });
-     * },
-     * ```
-     *
-     * @param [options] All `enqueueLinks()` parameters are passed via an options object.
-     * @returns Promise that resolves to {@link BatchAddRequestsResult} object.
-     */
-    enqueueLinks(options?: SetRequired<EnqueueLinksOptions, 'urls'>): Promise<BatchAddRequestsResult>;
+export interface BasicCrawlingContext<UserData extends Dictionary = Dictionary> extends CrawlingContext<UserData> {
 }
-export type RequestHandler<Context extends CrawlingContext = LoadedContext<BasicCrawlingContext & RestrictedCrawlingContext>> = (inputs: LoadedContext<Context>) => Awaitable<void>;
-export type ErrorHandler<Context extends CrawlingContext = LoadedContext<BasicCrawlingContext & RestrictedCrawlingContext>> = (inputs: LoadedContext<Context>, error: Error) => Awaitable<void>;
+export type RequestHandler<Context extends CrawlingContext = CrawlingContext> = (inputs: Context) => Awaitable<void>;
+export type ErrorHandler<Context extends CrawlingContext = CrawlingContext, ExtendedContext extends Context = Context> = (inputs: Context & Partial<ExtendedContext>, error: Error) => Awaitable<void>;
 export interface StatusMessageCallbackParams<Context extends CrawlingContext = BasicCrawlingContext, Crawler extends BasicCrawler<any> = BasicCrawler<Context>> {
     state: StatisticState;
     crawler: Crawler;
@@ -40,7 +16,10 @@ export interface StatusMessageCallbackParams<Context extends CrawlingContext = B
     message: string;
 }
 export type StatusMessageCallback<Context extends CrawlingContext = BasicCrawlingContext, Crawler extends BasicCrawler<any> = BasicCrawler<Context>> = (params: StatusMessageCallbackParams<Context, Crawler>) => Awaitable<void>;
-export interface BasicCrawlerOptions<Context extends CrawlingContext = BasicCrawlingContext> {
+export type RequireContextPipeline<DefaultContextType extends CrawlingContext, FinalContextType extends DefaultContextType> = DefaultContextType extends FinalContextType ? {} : {
+    contextPipelineBuilder: () => ContextPipeline<CrawlingContext, FinalContextType>;
+};
+export interface BasicCrawlerOptions<Context extends CrawlingContext = CrawlingContext, ContextExtension = {}, ExtendedContext extends Context = Context & ContextExtension> {
     /**
      * User-provided function that performs the logic of the crawler. It is called for each URL to crawl.
      *
@@ -58,7 +37,35 @@ export interface BasicCrawlerOptions<Context extends CrawlingContext = BasicCraw
      * The exceptions are logged to the request using the
      * {@link Request.pushErrorMessage|`Request.pushErrorMessage()`} function.
      */
-    requestHandler?: RequestHandler<Context>;
+    requestHandler?: RequestHandler<ExtendedContext>;
+    /**
+     * Allows the user to extend the crawling context passed to the request handler with custom functionality.
+     *
+     * **Example usage:**
+     *
+     * ```javascript
+     * import { BasicCrawler } from 'crawlee';
+     *
+     * // Create a crawler instance
+     * const crawler = new BasicCrawler({
+     *     extendContext(context) => ({
+     *         async customHelper() {
+     *             await context.pushData({ url: context.request.url })
+     *         }
+     *     }),
+     *     async requestHandler(context) {
+     *         await context.customHelper();
+     *     },
+     * });
+     * ```
+     */
+    extendContext?: (context: Context) => Awaitable<ContextExtension>;
+    /**
+     * *Intended for BasicCrawler subclasses*. Prepares a context pipeline that transforms the initial crawling context into the shape given by the `Context` type parameter.
+     *
+     * The option is not required if your crawler subclass does not extend the crawling context with custom information or helpers.
+     */
+    contextPipelineBuilder?: () => ContextPipeline<CrawlingContext, Context>;
     /**
      * Static list of URLs to be processed.
      * If not provided, the crawler will open the default request queue when the {@link BasicCrawler.addRequests|`crawler.addRequests()`} function is called.
@@ -87,7 +94,7 @@ export interface BasicCrawlerOptions<Context extends CrawlingContext = BasicCraw
      * Second argument is the `Error` instance that
      * represents the last error thrown during processing of the request.
      */
-    errorHandler?: ErrorHandler<Context>;
+    errorHandler?: ErrorHandler<CrawlingContext, ExtendedContext>;
     /**
      * A function to handle requests that failed more than {@link BasicCrawlerOptions.maxRequestRetries|`maxRequestRetries`} times.
      *
@@ -96,7 +103,7 @@ export interface BasicCrawlerOptions<Context extends CrawlingContext = BasicCraw
      * Second argument is the `Error` instance that
      * represents the last error thrown during processing of the request.
      */
-    failedRequestHandler?: ErrorHandler<Context>;
+    failedRequestHandler?: ErrorHandler<CrawlingContext, ExtendedContext>;
     /**
      * Specifies the maximum number of retries allowed for a request if its processing fails.
      * This includes retries due to navigation errors or errors thrown from user-supplied functions
@@ -303,7 +310,7 @@ export interface CrawlerExperiments {
  * ```
  * @category Crawlers
  */
-export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlingContext> {
+export declare class BasicCrawler<Context extends CrawlingContext = CrawlingContext, ContextExtension = {}, ExtendedContext extends Context = Context & ContextExtension> {
     readonly config: Configuration;
     protected static readonly CRAWLEE_STATE_KEY = "CRAWLEE_STATE";
     /**
@@ -338,13 +345,16 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      * Default {@link Router} instance that will be used if we don't specify any {@link BasicCrawlerOptions.requestHandler|`requestHandler`}.
      * See {@link Router.addHandler|`router.addHandler()`} and {@link Router.addDefaultHandler|`router.addDefaultHandler()`}.
      */
-    readonly router: RouterHandler<LoadedContext<Context>>;
+    readonly router: RouterHandler<Context>;
+    private contextPipelineBuilder;
+    private _contextPipeline?;
+    get contextPipeline(): ContextPipeline<CrawlingContext, ExtendedContext>;
     running: boolean;
     hasFinishedBefore: boolean;
     readonly log: Log;
-    protected requestHandler: RequestHandler<Context>;
-    protected errorHandler?: ErrorHandler<Context>;
-    protected failedRequestHandler?: ErrorHandler<Context>;
+    protected requestHandler: RequestHandler<ExtendedContext>;
+    protected errorHandler?: ErrorHandler<CrawlingContext, ExtendedContext>;
+    protected failedRequestHandler?: ErrorHandler<CrawlingContext, ExtendedContext>;
     protected requestHandlerTimeoutMillis: number;
     protected internalTimeoutMillis: number;
     protected maxRequestRetries: number;
@@ -356,7 +366,6 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
     protected statusMessageCallback?: StatusMessageCallback;
     protected sessionPoolOptions: SessionPoolOptions;
     protected useSessionPool: boolean;
-    protected crawlingContexts: Map<string, Context>;
     protected autoscaledPoolOptions: AutoscaledPoolOptions;
     protected events: EventManager;
     protected httpClient: BaseHttpClient;
@@ -368,6 +377,10 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
     private readonly robotsTxtFileCache;
     private _experimentWarnings;
     protected static optionsShape: {
+// @ts-ignore optional peer dependency or compatibility with es2022
+        contextPipelineBuilder: import("ow").ObjectPredicate<object> & import("ow").BasePredicate<object | undefined>;
+// @ts-ignore optional peer dependency or compatibility with es2022
+        extendContext: import("ow").Predicate<Function> & import("ow").BasePredicate<Function | undefined>;
 // @ts-ignore optional peer dependency or compatibility with es2022
         requestList: import("ow").ObjectPredicate<object> & import("ow").BasePredicate<object | undefined>;
 // @ts-ignore optional peer dependency or compatibility with es2022
@@ -424,7 +437,8 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
     /**
      * All `BasicCrawler` parameters are passed via an options object.
      */
-    constructor(options?: BasicCrawlerOptions<Context>, config?: Configuration);
+    constructor(options?: BasicCrawlerOptions<Context, ContextExtension, ExtendedContext> & RequireContextPipeline<CrawlingContext, Context>, // cast because the constructor logic handles missing `contextPipelineBuilder` - the type is just for DX
+    config?: Configuration);
     /**
      * Checks if the given error is a proxy error by comparing its message to a list of known proxy error messages.
      * Used for retrying requests that failed due to proxy errors.
@@ -432,12 +446,6 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      * @param error The error to check.
      */
     protected isProxyError(error: Error): boolean;
-    /**
-     * Checks whether the given crawling context is getting blocked by anti-bot protection using several heuristics.
-     * Returns `false` if the request is not blocked, otherwise returns a string with a description of the block reason.
-     * @param _crawlingContext The crawling context to check.
-     */
-    protected isRequestBlocked(_crawlingContext: Context): Promise<string | false>;
     /**
      * This method is periodically called by the crawler, every `statusMessageLoggingInterval` seconds.
      */
@@ -473,7 +481,7 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      * @param requests The requests to add
      * @param options Options for the request queue
      */
-    addRequests(requests: (string | Source)[], options?: CrawlerAddRequestsOptions): Promise<CrawlerAddRequestsResult>;
+    addRequests(requests: ReadonlyDeep<(string | Source)[]>, options?: CrawlerAddRequestsOptions): Promise<CrawlerAddRequestsResult>;
     /**
      * Pushes data to the specified {@link Dataset}, or the default crawler {@link Dataset} by calling {@link Dataset.pushData}.
      */
@@ -492,7 +500,7 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      */
     exportData<Data>(path: string, format?: 'json' | 'csv', options?: DatasetExportOptions): Promise<Data[]>;
     protected _init(): Promise<void>;
-    protected _runRequestHandler(crawlingContext: Context): Promise<void>;
+    protected runRequestHandler(crawlingContext: CrawlingContext): Promise<void>;
     /**
      * Handles blocked request
      */
@@ -505,11 +513,6 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      * and RequestQueue is present then enqueues it to the queue first.
      */
     protected _fetchNextRequest(): Promise<Request<Dictionary> | null | undefined>;
-    /**
-     * Executed when `errorHandler` finishes or the request is successful.
-     * Can be used to clean up orphaned browser pages.
-     */
-    protected _cleanupContext(_crawlingContext: Context): Promise<void>;
     /**
      * Delays processing of the request based on the `sameDomainDelaySecs` option,
      * adding it back to the queue after the timeout passes. Returns `true` if the request
@@ -535,12 +538,17 @@ export declare class BasicCrawler<Context extends CrawlingContext = BasicCrawlin
      */
     protected _defaultIsFinishedFunction(): Promise<boolean>;
     private _rotateSession;
+    /**
+     * Unwraps errors thrown by the context pipeline to get the actual user error.
+     * RequestHandlerError and ContextPipelineInitializationError wrap the actual error.
+     */
+    private unwrapError;
     /**
      * Handles errors thrown by user provided requestHandler()
      */
-    protected _requestFunctionErrorHandler(error: Error, crawlingContext: Context, source: IRequestList | RequestProvider): Promise<void>;
+    protected _requestFunctionErrorHandler(error: Error, crawlingContext: CrawlingContext, source: IRequestList | RequestProvider): Promise<void>;
     protected _tagUserHandlerError<T>(cb: () => unknown): Promise<T>;
-    protected _handleFailedRequestHandler(crawlingContext: Context, error: Error): Promise<void>;
+    protected _handleFailedRequestHandler(crawlingContext: CrawlingContext, error: Error): Promise<void>;
     /**
      * Resolves the most verbose error message from a thrown error
      * @param error The error received

package/internals/basic-crawler.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"basic-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/basic-crawler.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EACR,yBAAyB,EACzB,wBAAwB,EACxB,qBAAqB,EACrB,cAAc,EACd,eAAe,EACf,oBAAoB,EACpB,mBAAmB,EACnB,YAAY,EACZ,eAAe,EACf,sBAAsB,EACtB,YAAY,EACZ,aAAa,EACb,SAAS,EACT,OAAO,EACP,cAAc,EACd,yBAAyB,EACzB,aAAa,EACb,YAAY,EACZ,OAAO,EACP,kBAAkB,EAClB,sBAAsB,EACtB,MAAM,EACN,iBAAiB,EACjB,cAAc,EACjB,MAAM,eAAe,CAAC;AACvB,OAAO,EACH,cAAc,EACd,aAAa,EAEb,OAAO,EASP,eAAe,EAOf,WAAW,EACX,UAAU,EAEb,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EAAE,SAAS,EAAE,sBAAsB,EAAE,UAAU,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AAC7G,OAAO,EAAE,aAAa,EAAuB,MAAM,gBAAgB,CAAC;AAKpE,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAG7C,OAAO,KAAK,EAAE,GAAG,EAAE,MAAM,YAAY,CAAC;AAEtC,OAAO,EAAuB,YAAY,EAAa,MAAM,gBAAgB,CAAC;AAK9E,MAAM,WAAW,oBAAoB,CAAC,QAAQ,SAAS,UAAU,GAAG,UAAU,CAC1E,SAAQ,eAAe,CAAC,YAAY,EAAE,QAAQ,CAAC;IAC/C;;;;;;;;;;;;;;;;;;;;;;OAsBG;IACH,YAAY,CAAC,OAAO,CAAC,EAAE,WAAW,CAAC,mBAAmB,EAAE,MAAM,CAAC,GAAG,OAAO,CAAC,sBAAsB,CAAC,CAAC;CACrG;AAaD,MAAM,MAAM,cAAc,CACtB,OAAO,SAAS,eAAe,GAAG,aAAa,CAAC,oBAAoB,GAAG,yBAAyB,CAAC,IACjG,CAAC,MAAM,EAAE,aAAa,CAAC,OAAO,CAAC,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAExD,MAAM,MAAM,YAAY,CACpB,OAAO,SAAS,eAAe,GAAG,aAAa,CAAC,oBAAoB,GAAG,yBAAyB,CAAC,IACjG,CAAC,MAAM,EAAE,aAAa,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,KAAK,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAEtE,MAAM,WAAW,2BAA2B,CACxC,OAAO,SAAS,eAAe,GAAG,oBAAoB,EACtD,OAAO,SAAS,YAAY,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC;IAEzD,KAAK,EAAE,cAAc,CAAC;IACtB,OAAO,EAAE,OAAO,CAAC;IACjB,aAAa,EAAE,cAAc,CAAC;IAC9B,OAAO,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,qBAAqB,CAC7B,OAAO,SAAS,eAAe,GAAG,oBAAoB,EACtD,OAAO,SAAS,YAAY,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC,IACzD,CAAC,MAAM,EAAE,2BAA2B,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAE/E,MAAM,WAAW,mBAAmB,CAAC,OAAO,SAAS,eAAe,GAAG,oBAAoB;IACvF;;;;;;;;;;;;;;;;OAgBG;IACH,cAAc,CAAC,EAAE,cAAc,CAAC,OAAO,CAAC,CAAC;IAEzC;;;;;OAKG;IACH,WAAW,CAAC,EAAE,YAAY,CAAC;IAE3B;;;;;OAKG;IACH,YAAY,CAAC,EAAE,eAAe,CAAC;IAE/B;;;OAGG;IACH,yBAAyB,CAAC,EAAE,MAAM,CAAC;IAEnC;;;;;;;;OAQG;IACH,YAAY,CAAC,EAAE,YAAY,CAAC,OAAO,CAAC,CAAC;IAErC;;;;;;;OAOG;IACH,oBAAoB,CAAC,EAAE,YAAY,CAAC,OAAO,CAAC,CAAC;IAE7C;;;;;;;;OAQG;IACH,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAE3B;;;OAGG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;;;OAMG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;OAIG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;;;OAMG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IAE9C;;;;;OAKG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IAExB;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IAExB;;;;OAIG;IACH,oBAAoB,CAAC,EAAE,MAAM,CAAC;IAE9B;;;;OAIG;IACH,SAAS,CAAC,EAAE,OAAO,CAAC;IAEpB;;;OAGG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IAEzB;;OAEG;IACH,kBAAkB,CAAC,EAAE,kBAAkB,CAAC;IAExC;;OAEG;IACH,4BAA4B,CAAC,EAAE,MAAM,CAAC;IAEtC;;;;;;;;;;;;;;;OAeG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IAE9C;;;;;;OAMG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IAEzB;;;OAGG;IACH,oBAAoB,CAAC,EAAE,OAAO,CAAC;IAE/B;;;OAGG;IACH,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IAE1C,gBAAgB;IAChB,GAAG,CAAC,EAAE,GAAG,CAAC;IAEV;;;OAGG;IACH,WAAW,CAAC,EAAE,kBAAkB,CAAC;IAEjC;;;OAGG;IACH,iBAAiB,CAAC,EAAE,iBAAiB,CAAC;IAEtC;;;OAGG;IACH,UAAU,CAAC,EAAE,cAAc,CAAC;CAC/B;AAED;;;;;GAKG;AACH,MAAM,WAAW,kBAAkB;IAC/B;;;;;OAKG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;CAC5B;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+DG;AACH,qBAAa,YAAY,CAAC,OAAO,SAAS,eAAe,GAAG,oBAAoB;IAqHxE,QAAQ,CAAC,MAAM;IApHnB,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,mBAAmB;IAE9D;;OAEG;IACH,QAAQ,CAAC,KAAK,EAAE,UAAU,CAAC;IAE3B;;;OAGG;IACH,WAAW,CAAC,EAAE,YAAY,CAAC;IAE3B;;;;OAIG;IACH,YAAY,CAAC,EAAE,eAAe,CAAC;IAE/B;;;OAGG;IACH,WAAW,CAAC,EAAE,WAAW,CAAC;IAE1B;;;;;;OAMG;IACH,cAAc,CAAC,EAAE,cAAc,CAAC;IAEhC;;;OAGG;IACH,QAAQ,CAAC,MAAM,EAAE,aAAa,CAAC,aAAa,CAAC,OAAO,CAAC,CAAC,CAA2C;IAEjG,OAAO,UAAS;IAChB,iBAAiB,UAAS;IAE1B,QAAQ,CAAC,GAAG,EAAE,GAAG,CAAC;IAClB,SAAS,CAAC,cAAc,EAAG,cAAc,CAAC,OAAO,CAAC,CAAC;IACnD,SAAS,CAAC,YAAY,CAAC,EAAE,YAAY,CAAC,OAAO,CAAC,CAAC;IAC/C,SAAS,CAAC,oBAAoB,CAAC,EAAE,YAAY,CAAC,OAAO,CAAC,CAAC;IACvD,SAAS,CAAC,2BAA2B,EAAG,MAAM,CAAC;IAC/C,SAAS,CAAC,qBAAqB,EAAE,MAAM,CAAC;IACxC,SAAS,CAAC,iBAAiB,EAAE,MAAM,CAAC;IACpC,SAAS,CAAC,qBAAqB,EAAE,MAAM,CAAC;IACxC,SAAS,CAAC,kBAAkB,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAClD,SAAS,CAAC,mBAAmB,EAAE,MAAM,CAAC;IACtC,SAAS,CAAC,oBAAoB,EAAE,MAAM,CAAC;IACvC,SAAS,CAAC,4BAA4B,EAAE,MAAM,CAAC;IAC/C,SAAS,CAAC,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IACxD,SAAS,CAAC,kBAAkB,EAAE,kBAAkB,CAAC;IACjD,SAAS,CAAC,cAAc,EAAE,OAAO,CAAC;IAClC,SAAS,CAAC,gBAAgB,uBAA8B;IACxD,SAAS,CAAC,qBAAqB,EAAE,qBAAqB,CAAC;IACvD,SAAS,CAAC,MAAM,EAAE,YAAY,CAAC;IAC/B,SAAS,CAAC,UAAU,EAAE,cAAc,CAAC;IACrC,SAAS,CAAC,cAAc,EAAE,OAAO,CAAC;IAClC,SAAS,CAAC,oBAAoB,EAAE,OAAO,CAAC;IACxC,SAAS,CAAC,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IACpD,OAAO,CAAC,YAAY,CAAC,CAAU;IAE/B,OAAO,CAAC,WAAW,CAAqB;IACxC,OAAO,CAAC,QAAQ,CAAC,kBAAkB,CAA0B;IAC7D,OAAO,CAAC,mBAAmB,CAA0D;IAErF,SAAS,CAAC,MAAM,CAAC,YAAY;;;;;;;;;;;;;;;;;;;;;;;;;;;MAqC3B;IAEF;;OAEG;gBAEC,OAAO,GAAE,mBAAmB,CAAC,OAAO,CAAM,EACjC,MAAM,gBAAkC;IA4KrD;;;;;OAKG;IACH,SAAS,CAAC,YAAY,CAAC,KAAK,EAAE,KAAK,GAAG,OAAO;IAI7C;;;;OAIG;cACa,gBAAgB,CAAC,gBAAgB,EAAE,OAAO,GAAG,OAAO,CAAC,MAAM,GAAG,KAAK,CAAC;IAIpF;;OAEG;IACG,gBAAgB,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,GAAE,uBAA4B;IAmB7E,OAAO,CAAC,iBAAiB;IAgDzB;;;;;;;;;OASG;IACG,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,GAAG,OAAO,GAAG,cAAc,CAAC,EAAE,EAAE,OAAO,CAAC,EAAE,iBAAiB,GAAG,OAAO,CAAC,eAAe,CAAC;IA8GlH;;;;OAIG;IACH,IAAI,CAAC,OAAO,SAA6C,GAAG,IAAI;IAY1D,eAAe;IAYf,QAAQ,CAAC,KAAK,SAAS,UAAU,GAAG,UAAU,EAAE,YAAY,GAAS,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;IAKjG;;;;;;;;;;OAUG;IACG,WAAW,CACb,QAAQ,EAAE,CAAC,MAAM,GAAG,MAAM,CAAC,EAAE,EAC7B,OAAO,GAAE,yBAA8B,GACxC,OAAO,CAAC,wBAAwB,CAAC;IAsCpC;;OAEG;IACG,QAAQ,CAAC,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,eAAe,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAKjG;;OAEG;IACG,UAAU,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;IAIrD;;OAEG;IACG,OAAO,CAAC,GAAG,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IAKtF;;;OAGG;IACG,UAAU,CAAC,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,GAAG,KAAK,EAAE,OAAO,CAAC,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;cAoC9F,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;cAoBtB,kBAAkB,CAAC,eAAe,EAAE,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC;IAI3E;;OAEG;IACH,SAAS,CAAC,sBAAsB,CAAC,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM;YAQvD,6BAA6B;cAS3B,sBAAsB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,aAAa,GAAG,SAAS,CAAC;cAuBvE,iBAAiB;IAuCjC;;;OAGG;cACa,iBAAiB;IAyBjC;;;OAGG;cACa,eAAe,CAAC,gBAAgB,EAAE,OAAO;IAEzD;;;;OAIG;IACH,SAAS,CAAC,YAAY,CAAC,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,YAAY,GAAG,eAAe;IAsC/E;;;OAGG;cACa,gBAAgB;IA0JhC;;;OAGG;cACa,gBAAgB,CAC5B,OAAO,EAAE,MAAM,OAAO,CAAC,OAAO,CAAC,EAC/B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,KAAK,GAAG,MAAM,EACrB,UAAU,SAAI,EACd,OAAO,SAAI,GACZ,OAAO,CAAC,IAAI,CAAC;IAehB;;OAEG;cACa,oBAAoB;IASpC;;OAEG;cACa,0BAA0B;YAS5B,cAAc;IAQ5B;;OAEG;cACa,4BAA4B,CACxC,KAAK,EAAE,KAAK,EACZ,eAAe,EAAE,OAAO,EACxB,MAAM,EAAE,YAAY,GAAG,eAAe,GACvC,OAAO,CAAC,IAAI,CAAC;cAyDA,oBAAoB,CAAC,CAAC,EAAE,EAAE,EAAE,MAAM,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC;cAStD,2BAA2B,CAAC,eAAe,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC;IAYlG;;;;OAIG;IACH,SAAS,CAAC,oBAAoB,CAAC,KAAK,EAAE,KAAK,EAAE,UAAU,UAAQ;IAmB/D,SAAS,CAAC,oBAAoB,CAAC,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK;IAoB7D;;;;;;;OAOG;cACa,wBAAwB,IAAI,OAAO,CAAC,IAAI,CAAC;cAQzC,aAAa,CAAC,QAAQ,SAAS,CAAC,GAAG,IAAI,EAAE,GAAG,EAAE,KAAK,SAAS,CAAC,IAAI,CAAC,EAC9E,KAAK,EAAE,QAAQ,EAAE,EACjB,GAAG,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC;IASjC;;;OAGG;IACG,QAAQ,IAAI,OAAO,CAAC,IAAI,CAAC;IAc/B,SAAS,CAAC,2BAA2B,CAAC,OAAO,EAAE,OAAO;YAWxC,gBAAgB;IAc9B,SAAS,CAAC,6BAA6B,CAAC,OAAO,EAAE,OAAO;CA6C3D;AAED,MAAM,WAAW,oBAAoB;IACjC,OAAO,EAAE,OAAO,CAAC;IACjB,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,SAAS,CAAC,EAAE,SAAS,CAAC;CACzB;AAED,MAAM,WAAW,yBAA0B,SAAQ,yBAAyB;CAAG;AAE/E,MAAM,WAAW,wBAAyB,SAAQ,wBAAwB;CAAG;AAE7E,MAAM,WAAW,iBAAkB,SAAQ,yBAAyB;IAChE;;;;OAIG;IACH,iBAAiB,CAAC,EAAE,OAAO,CAAC;CAC/B;AAED;;;;;;;;;;;;;;;;;;;;;;;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,0BAEzC"}
+{"version":3,"file":"basic-crawler.d.ts","sourceRoot":"","sources":["../../src/internals/basic-crawler.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EACR,yBAAyB,EACzB,wBAAwB,EACxB,qBAAqB,EACrB,cAAc,EACd,eAAe,EACf,oBAAoB,EACpB,YAAY,EACZ,eAAe,EACf,sBAAsB,EACtB,YAAY,EACZ,SAAS,EACT,OAAO,EACP,cAAc,EACd,aAAa,EACb,YAAY,EACZ,OAAO,EACP,kBAAkB,EAClB,sBAAsB,EACtB,MAAM,EACN,iBAAiB,EACjB,cAAc,EACjB,MAAM,eAAe,CAAC;AACvB,OAAO,EACH,cAAc,EACd,aAAa,EACb,eAAe,EAKf,OAAO,EAUP,eAAe,EAOf,WAAW,EACX,UAAU,EAEb,MAAM,eAAe,CAAC;AACvB,OAAO,KAAK,EAAE,SAAS,EAAE,UAAU,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AACrF,OAAO,EAAE,aAAa,EAAuB,MAAM,gBAAgB,CAAC;AAKpE,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AAG9C,OAAO,KAAK,EAAE,GAAG,EAAE,MAAM,YAAY,CAAC;AAEtC,OAAO,EAAuB,YAAY,EAAa,MAAM,gBAAgB,CAAC;AAK9E,MAAM,WAAW,oBAAoB,CAAC,QAAQ,SAAS,UAAU,GAAG,UAAU,CAAE,SAAQ,eAAe,CAAC,QAAQ,CAAC;CAAG;AAapH,MAAM,MAAM,cAAc,CAAC,OAAO,SAAS,eAAe,GAAG,eAAe,IAAI,CAAC,MAAM,EAAE,OAAO,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAErH,MAAM,MAAM,YAAY,CACpB,OAAO,SAAS,eAAe,GAAG,eAAe,EACjD,eAAe,SAAS,OAAO,GAAG,OAAO,IACzC,CAAC,MAAM,EAAE,OAAO,GAAG,OAAO,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,KAAK,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAElF,MAAM,WAAW,2BAA2B,CACxC,OAAO,SAAS,eAAe,GAAG,oBAAoB,EACtD,OAAO,SAAS,YAAY,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC;IAEzD,KAAK,EAAE,cAAc,CAAC;IACtB,OAAO,EAAE,OAAO,CAAC;IACjB,aAAa,EAAE,cAAc,CAAC;IAC9B,OAAO,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,qBAAqB,CAC7B,OAAO,SAAS,eAAe,GAAG,oBAAoB,EACtD,OAAO,SAAS,YAAY,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC,IACzD,CAAC,MAAM,EAAE,2BAA2B,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,SAAS,CAAC,IAAI,CAAC,CAAC;AAE/E,MAAM,MAAM,sBAAsB,CAC9B,kBAAkB,SAAS,eAAe,EAC1C,gBAAgB,SAAS,kBAAkB,IAC3C,kBAAkB,SAAS,gBAAgB,GACzC,EAAE,GACF;IAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC,eAAe,EAAE,gBAAgB,CAAC,CAAA;CAAE,CAAC;AAE3F,MAAM,WAAW,mBAAmB,CAChC,OAAO,SAAS,eAAe,GAAG,eAAe,EACjD,gBAAgB,GAAG,EAAE,EACrB,eAAe,SAAS,OAAO,GAAG,OAAO,GAAG,gBAAgB;IAE5D;;;;;;;;;;;;;;;;OAgBG;IACH,cAAc,CAAC,EAAE,cAAc,CAAC,eAAe,CAAC,CAAC;IAEjD;;;;;;;;;;;;;;;;;;;;OAoBG;IACH,aAAa,CAAC,EAAE,CAAC,OAAO,EAAE,OAAO,KAAK,SAAS,CAAC,gBAAgB,CAAC,CAAC;IAElE;;;;OAIG;IACH,sBAAsB,CAAC,EAAE,MAAM,eAAe,CAAC,eAAe,EAAE,OAAO,CAAC,CAAC;IAEzE;;;;;OAKG;IACH,WAAW,CAAC,EAAE,YAAY,CAAC;IAE3B;;;;;OAKG;IACH,YAAY,CAAC,EAAE,eAAe,CAAC;IAE/B;;;OAGG;IACH,yBAAyB,CAAC,EAAE,MAAM,CAAC;IAEnC;;;;;;;;OAQG;IACH,YAAY,CAAC,EAAE,YAAY,CAAC,eAAe,EAAE,eAAe,CAAC,CAAC;IAE9D;;;;;;;OAOG;IACH,oBAAoB,CAAC,EAAE,YAAY,CAAC,eAAe,EAAE,eAAe,CAAC,CAAC;IAEtE;;;;;;;;OAQG;IACH,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAE3B;;;OAGG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;;;OAMG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;OAIG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAE7B;;;;;;OAMG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IAE9C;;;;;OAKG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IAExB;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IAExB;;;;OAIG;IACH,oBAAoB,CAAC,EAAE,MAAM,CAAC;IAE9B;;;;OAIG;IACH,SAAS,CAAC,EAAE,OAAO,CAAC;IAEpB;;;OAGG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IAEzB;;OAEG;IACH,kBAAkB,CAAC,EAAE,kBAAkB,CAAC;IAExC;;OAEG;IACH,4BAA4B,CAAC,EAAE,MAAM,CAAC;IAEtC;;;;;;;;;;;;;;;OAeG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IAE9C;;;;;;OAMG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IAEzB;;;OAGG;IACH,oBAAoB,CAAC,EAAE,OAAO,CAAC;IAE/B;;;OAGG;IACH,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IAE1C,gBAAgB;IAChB,GAAG,CAAC,EAAE,GAAG,CAAC;IAEV;;;OAGG;IACH,WAAW,CAAC,EAAE,kBAAkB,CAAC;IAEjC;;;OAGG;IACH,iBAAiB,CAAC,EAAE,iBAAiB,CAAC;IAEtC;;;OAGG;IACH,UAAU,CAAC,EAAE,cAAc,CAAC;CAC/B;AAED;;;;;GAKG;AACH,MAAM,WAAW,kBAAkB;IAC/B;;;;;OAKG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;CAC5B;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+DG;AACH,qBAAa,YAAY,CACrB,OAAO,SAAS,eAAe,GAAG,eAAe,EACjD,gBAAgB,GAAG,EAAE,EACrB,eAAe,SAAS,OAAO,GAAG,OAAO,GAAG,gBAAgB;IAoIxD,QAAQ,CAAC,MAAM;IAlInB,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,mBAAmB;IAE9D;;OAEG;IACH,QAAQ,CAAC,KAAK,EAAE,UAAU,CAAC;IAE3B;;;OAGG;IACH,WAAW,CAAC,EAAE,YAAY,CAAC;IAE3B;;;;OAIG;IACH,YAAY,CAAC,EAAE,eAAe,CAAC;IAE/B;;;OAGG;IACH,WAAW,CAAC,EAAE,WAAW,CAAC;IAE1B;;;;;;OAMG;IACH,cAAc,CAAC,EAAE,cAAc,CAAC;IAEhC;;;OAGG;IACH,QAAQ,CAAC,MAAM,EAAE,aAAa,CAAC,OAAO,CAAC,CAA4B;IAEnE,OAAO,CAAC,sBAAsB,CAA0D;IACxF,OAAO,CAAC,gBAAgB,CAAC,CAAoD;IAE7E,IAAI,eAAe,IAAI,eAAe,CAAC,eAAe,EAAE,eAAe,CAAC,CAMvE;IAED,OAAO,UAAS;IAChB,iBAAiB,UAAS;IAE1B,QAAQ,CAAC,GAAG,EAAE,GAAG,CAAC;IAClB,SAAS,CAAC,cAAc,EAAG,cAAc,CAAC,eAAe,CAAC,CAAC;IAC3D,SAAS,CAAC,YAAY,CAAC,EAAE,YAAY,CAAC,eAAe,EAAE,eAAe,CAAC,CAAC;IACxE,SAAS,CAAC,oBAAoB,CAAC,EAAE,YAAY,CAAC,eAAe,EAAE,eAAe,CAAC,CAAC;IAChF,SAAS,CAAC,2BAA2B,EAAG,MAAM,CAAC;IAC/C,SAAS,CAAC,qBAAqB,EAAE,MAAM,CAAC;IACxC,SAAS,CAAC,iBAAiB,EAAE,MAAM,CAAC;IACpC,SAAS,CAAC,qBAAqB,EAAE,MAAM,CAAC;IACxC,SAAS,CAAC,kBAAkB,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAClD,SAAS,CAAC,mBAAmB,EAAE,MAAM,CAAC;IACtC,SAAS,CAAC,oBAAoB,EAAE,MAAM,CAAC;IACvC,SAAS,CAAC,4BAA4B,EAAE,MAAM,CAAC;IAC/C,SAAS,CAAC,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IACxD,SAAS,CAAC,kBAAkB,EAAE,kBAAkB,CAAC;IACjD,SAAS,CAAC,cAAc,EAAE,OAAO,CAAC;IAClC,SAAS,CAAC,qBAAqB,EAAE,qBAAqB,CAAC;IACvD,SAAS,CAAC,MAAM,EAAE,YAAY,CAAC;IAC/B,SAAS,CAAC,UAAU,EAAE,cAAc,CAAC;IACrC,SAAS,CAAC,cAAc,EAAE,OAAO,CAAC;IAClC,SAAS,CAAC,oBAAoB,EAAE,OAAO,CAAC;IACxC,SAAS,CAAC,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IACpD,OAAO,CAAC,YAAY,CAAC,CAAU;IAE/B,OAAO,CAAC,WAAW,CAAqB;IACxC,OAAO,CAAC,QAAQ,CAAC,kBAAkB,CAA0B;IAC7D,OAAO,CAAC,mBAAmB,CAA0D;IAErF,SAAS,CAAC,MAAM,CAAC,YAAY;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAwC3B;IAEF;;OAEG;gBAEC,OAAO,GAAE,mBAAmB,CAAC,OAAO,EAAE,gBAAgB,EAAE,eAAe,CAAC,GACpE,sBAAsB,CAAC,eAAe,EAAE,OAAO,CAAa,EAAE,wGAAwG;IACjK,MAAM,gBAAkC;IA2MrD;;;;;OAKG;IACH,SAAS,CAAC,YAAY,CAAC,KAAK,EAAE,KAAK,GAAG,OAAO;IAI7C;;OAEG;IACG,gBAAgB,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,GAAE,uBAA4B;IAmB7E,OAAO,CAAC,iBAAiB;IAgDzB;;;;;;;;;OASG;IACG,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,GAAG,OAAO,GAAG,cAAc,CAAC,EAAE,EAAE,OAAO,CAAC,EAAE,iBAAiB,GAAG,OAAO,CAAC,eAAe,CAAC;IA8GlH;;;;OAIG;IACH,IAAI,CAAC,OAAO,SAA6C,GAAG,IAAI;IAY1D,eAAe;IAYf,QAAQ,CAAC,KAAK,SAAS,UAAU,GAAG,UAAU,EAAE,YAAY,GAAS,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;IAKjG;;;;;;;;;;OAUG;IACG,WAAW,CACb,QAAQ,EAAE,YAAY,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC,EAAE,CAAC,EAC3C,OAAO,GAAE,yBAA8B,GACxC,OAAO,CAAC,wBAAwB,CAAC;IAsCpC;;OAEG;IACG,QAAQ,CAAC,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,eAAe,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAKjG;;OAEG;IACG,UAAU,CAAC,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;IAIrD;;OAEG;IACG,OAAO,CAAC,GAAG,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IAKtF;;;OAGG;IACG,UAAU,CAAC,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,GAAG,KAAK,EAAE,OAAO,CAAC,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;cAoC9F,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;cAoBtB,iBAAiB,CAAC,eAAe,EAAE,eAAe,GAAG,OAAO,CAAC,IAAI,CAAC;IAUlF;;OAEG;IACH,SAAS,CAAC,sBAAsB,CAAC,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM;YAQvD,6BAA6B;cAS3B,sBAAsB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,aAAa,GAAG,SAAS,CAAC;cAuBvE,iBAAiB;IAuCjC;;;OAGG;cACa,iBAAiB;IAyBjC;;;;OAIG;IACH,SAAS,CAAC,YAAY,CAAC,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,YAAY,GAAG,eAAe;IAsC/E;;;OAGG;cACa,gBAAgB;IAyJhC;;;OAGG;cACa,gBAAgB,CAC5B,OAAO,EAAE,MAAM,OAAO,CAAC,OAAO,CAAC,EAC/B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,KAAK,GAAG,MAAM,EACrB,UAAU,SAAI,EACd,OAAO,SAAI,GACZ,OAAO,CAAC,IAAI,CAAC;IAehB;;OAEG;cACa,oBAAoB;IASpC;;OAEG;cACa,0BAA0B;YAS5B,cAAc;IAQ5B;;;OAGG;IACH,OAAO,CAAC,WAAW;IAWnB;;OAEG;cACa,4BAA4B,CACxC,KAAK,EAAE,KAAK,EACZ,eAAe,EAAE,eAAe,EAChC,MAAM,EAAE,YAAY,GAAG,eAAe,GACvC,OAAO,CAAC,IAAI,CAAC;cA4DA,oBAAoB,CAAC,CAAC,EAAE,EAAE,EAAE,MAAM,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC;cAStD,2BAA2B,CAAC,eAAe,EAAE,eAAe,EAAE,KAAK,EAAE,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC;IAe1G;;;;OAIG;IACH,SAAS,CAAC,oBAAoB,CAAC,KAAK,EAAE,KAAK,EAAE,UAAU,UAAQ;IAmB/D,SAAS,CAAC,oBAAoB,CAAC,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK;IAoB7D;;;;;;;OAOG;cACa,wBAAwB,IAAI,OAAO,CAAC,IAAI,CAAC;cAQzC,aAAa,CAAC,QAAQ,SAAS,CAAC,GAAG,IAAI,EAAE,GAAG,EAAE,KAAK,SAAS,CAAC,IAAI,CAAC,EAC9E,KAAK,EAAE,QAAQ,EAAE,EACjB,GAAG,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC;IASjC;;;OAGG;IACG,QAAQ,IAAI,OAAO,CAAC,IAAI,CAAC;IAY/B,SAAS,CAAC,2BAA2B,CAAC,OAAO,EAAE,OAAO;YAWxC,gBAAgB;IAc9B,SAAS,CAAC,6BAA6B,CAAC,OAAO,EAAE,OAAO;CA6C3D;AAED,MAAM,WAAW,oBAAoB;IACjC,OAAO,EAAE,OAAO,CAAC;IACjB,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,SAAS,CAAC,EAAE,SAAS,CAAC;CACzB;AAED,MAAM,WAAW,yBAA0B,SAAQ,yBAAyB;CAAG;AAE/E,MAAM,WAAW,wBAAyB,SAAQ,wBAAwB;CAAG;AAE7E,MAAM,WAAW,iBAAkB,SAAQ,yBAAyB;IAChE;;;;OAIG;IACH,iBAAiB,CAAC,EAAE,OAAO,CAAC;CAC/B;AAED;;;;;;;;;;;;;;;;;;;;;;;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,0BAEzC"}

package/internals/basic-crawler.js

@@ -1,6 +1,6 @@
 import { writeFile } from 'node:fs/promises';
 import { dirname } from 'node:path';
-import { AutoscaledPool, Configuration, CriticalError, Dataset, enqueueLinks, EnqueueStrategy, GotScrapingHttpClient, KeyValueStore, mergeCookies, NonRetryableError, purgeDefaultStorages, RequestProvider, RequestQueue, RequestQueueV1, RequestState, RetryRequestError, Router, SessionError, SessionPool, Statistics, validators, } from '@crawlee/core';
+import { AutoscaledPool, Configuration, ContextPipeline, ContextPipelineCleanupError, ContextPipelineInitializationError, ContextPipelineInterruptedError, CriticalError, Dataset, enqueueLinks, EnqueueStrategy, GotScrapingHttpClient, KeyValueStore, mergeCookies, NonRetryableError, purgeDefaultStorages, RequestHandlerError, RequestProvider, RequestQueue, RequestQueueV1, RequestState, RetryRequestError, Router, SessionError, SessionPool, Statistics, validators, } from '@crawlee/core';
 import { RobotsTxtFile, ROTATE_PROXY_ERRORS } from '@crawlee/utils';
 import { stringify } from 'csv-stringify/sync';
 import { ensureDir, writeJSON } from 'fs-extra/esm';
@@ -121,6 +121,14 @@ export class BasicCrawler {
      * See {@link Router.addHandler|`router.addHandler()`} and {@link Router.addDefaultHandler|`router.addDefaultHandler()`}.
      */
     router = Router.create();
+    contextPipelineBuilder;
+    _contextPipeline;
+    get contextPipeline() {
+        if (this._contextPipeline === undefined) {
+            this._contextPipeline = this.contextPipelineBuilder();
+        }
+        return this._contextPipeline;
+    }
     running = false;
     hasFinishedBefore = false;
     log;
@@ -138,7 +146,6 @@ export class BasicCrawler {
     statusMessageCallback;
     sessionPoolOptions;
     useSessionPool;
-    crawlingContexts = new Map();
     autoscaledPoolOptions;
     events;
     httpClient;
@@ -150,6 +157,8 @@ export class BasicCrawler {
     robotsTxtFileCache;
     _experimentWarnings = {};
     static optionsShape = {
+        contextPipelineBuilder: ow.optional.object,
+        extendContext: ow.optional.function,
         requestList: ow.optional.object.validate(validators.requestList),
         requestQueue: ow.optional.object.validate(validators.requestQueue),
         // Subclasses override this function instead of passing it
@@ -185,7 +194,8 @@ export class BasicCrawler {
     /**
      * All `BasicCrawler` parameters are passed via an options object.
      */
-    constructor(options = {}, config = Configuration.getGlobalConfig()) {
+    constructor(options = {}, // cast because the constructor logic handles missing `contextPipelineBuilder` - the type is just for DX
+    config = Configuration.getGlobalConfig()) {
         this.config = config;
         ow(options, 'BasicCrawlerOptions', ow.object.exactShape(BasicCrawler.optionsShape));
         const { requestList, requestQueue, maxRequestRetries = 3, sameDomainDelaySecs = 0, maxSessionRotations = 10, maxRequestsPerCrawl, autoscaledPoolOptions = {}, keepAlive, sessionPoolOptions = {}, useSessionPool = true, 
@@ -193,6 +203,32 @@ export class BasicCrawler {
         minConcurrency, maxConcurrency, maxRequestsPerMinute, retryOnBlocked = false, respectRobotsTxtFile = false, onSkippedRequest, requestHandler, requestHandlerTimeoutSecs, errorHandler, failedRequestHandler, statusMessageLoggingInterval = 10, statusMessageCallback, statisticsOptions, httpClient, 
         // internal
         log = defaultLog.child({ prefix: this.constructor.name }), experiments = {}, } = options;
+        // Store the builder so that it can be run when the contextPipeline is needed.
+        // Invoking it immediately would cause problems with parent constructor call order.
+        this.contextPipelineBuilder = () => {
+            let contextPipeline = (options.contextPipelineBuilder?.() ??
+                ContextPipeline.create()); // Thanks to the RequireContextPipeline, contextPipeline will only be undefined if InitialContextType is CrawlingContext
+            if (options.extendContext !== undefined) {
+                contextPipeline = contextPipeline.compose({
+                    action: async (context) => await options.extendContext(context),
+                });
+            }
+            contextPipeline = contextPipeline.compose({
+                action: async (context) => {
+                    const { request } = context;
+                    if (!this.requestMatchesEnqueueStrategy(request)) {
+                        // eslint-disable-next-line dot-notation
+                        const message = `Skipping request ${request.id} (starting url: ${request.url} -> loaded url: ${request.loadedUrl}) because it does not match the enqueue strategy (${request['enqueueStrategy']}).`;
+                        this.log.debug(message);
+                        request.noRetry = true;
+                        request.state = RequestState.SKIPPED;
+                        throw new ContextPipelineInterruptedError(message);
+                    }
+                    return context;
+                },
+            });
+            return contextPipeline;
+        };
         this.requestList = requestList;
         this.requestQueue = requestQueue;
         this.httpClient = httpClient ?? new GotScrapingHttpClient();
@@ -247,7 +283,6 @@ export class BasicCrawler {
             }
         }
         this.useSessionPool = useSessionPool;
-        this.crawlingContexts = new Map();
         const maxSignedInteger = 2 ** 31 - 1;
         if (this.requestHandlerTimeoutMillis > maxSignedInteger) {
             log.warning(`requestHandlerTimeoutMillis ${this.requestHandlerTimeoutMillis}` +
@@ -310,14 +345,6 @@ export class BasicCrawler {
     isProxyError(error) {
         return ROTATE_PROXY_ERRORS.some((x) => this._getMessageFromError(error)?.includes(x));
     }
-    /**
-     * Checks whether the given crawling context is getting blocked by anti-bot protection using several heuristics.
-     * Returns `false` if the request is not blocked, otherwise returns a string with a description of the block reason.
-     * @param _crawlingContext The crawling context to check.
-     */
-    async isRequestBlocked(_crawlingContext) {
-        throw new Error('the "isRequestBlocked" method is not implemented in this crawler.');
-    }
     /**
      * This method is periodically called by the crawler, every `statusMessageLoggingInterval` seconds.
      */
@@ -590,8 +617,10 @@ export class BasicCrawler {
         }
         await this._loadHandledRequestCount();
     }
-    async _runRequestHandler(crawlingContext) {
-        await this.requestHandler(crawlingContext);
+    async runRequestHandler(crawlingContext) {
+        await this.contextPipeline.call(crawlingContext, async (finalContext) => {
+            await addTimeoutToPromise(async () => this.requestHandler(finalContext), this.requestHandlerTimeoutMillis, `requestHandler timed out after ${this.requestHandlerTimeoutMillis / 1000} seconds (${finalContext.request.id}).`);
+        });
     }
     /**
      * Handles blocked request
@@ -686,11 +715,6 @@ export class BasicCrawler {
         await this.requestList.markRequestHandled(request);
         return this.requestQueue.fetchNextRequest();
     }
-    /**
-     * Executed when `errorHandler` finishes or the request is successful.
-     * Can be used to clean up orphaned browser pages.
-     */
-    async _cleanupContext(_crawlingContext) { }
     /**
      * Delays processing of the request based on the `sameDomainDelaySecs` option,
      * adding it back to the queue after the timeout passes. Returns `true` if the request
@@ -759,18 +783,14 @@ export class BasicCrawler {
         request.loadedUrl = undefined;
         const statisticsId = request.id || request.uniqueKey;
         this.stats.startJob(statisticsId);
-        // Shared crawling context
-        // @ts-expect-error
-        // All missing properties (that extend CrawlingContext) are set dynamically,
-        // but TS does not know that, so otherwise it would throw when compiling.
+        const deferredCleanup = [];
         const crawlingContext = {
             id: cryptoRandomObjectId(10),
-            crawler: this,
             log: this.log,
             request,
             session,
             enqueueLinks: async (options) => {
-                return enqueueLinks({
+                return await enqueueLinks({
                     // specify the RQ first to allow overriding it
                     requestQueue: await this.getRequestQueue(),
                     robotsTxtFile: await this.getRobotsTxtFileForUrl(request.url),
@@ -778,17 +798,21 @@ export class BasicCrawler {
                     ...options,
                 });
             },
-            addRequests: this.addRequests.bind(this),
+            addRequests: async (requests, options) => {
+                await this.addRequests(requests, options);
+            },
             pushData: this.pushData.bind(this),
             useState: this.useState.bind(this),
             sendRequest: createSendRequest(this.httpClient, request, session, () => crawlingContext.proxyInfo?.url),
             getKeyValueStore: async (idOrName) => KeyValueStore.open(idOrName, { config: this.config }),
+            registerDeferredCleanup: (cleanup) => {
+                deferredCleanup.push(cleanup);
+            },
         };
-        this.crawlingContexts.set(crawlingContext.id, crawlingContext);
         let isRequestLocked = true;
         try {
             request.state = RequestState.REQUEST_HANDLER;
-            await addTimeoutToPromise(async () => this._runRequestHandler(crawlingContext), this.requestHandlerTimeoutMillis, `requestHandler timed out after ${this.requestHandlerTimeoutMillis / 1000} seconds (${request.id}).`);
+            await this.runRequestHandler(crawlingContext);
             await this._timeoutAndRetry(async () => source.markRequestHandled(request), this.internalTimeoutMillis, `Marking request ${request.url} (${request.id}) as handled timed out after ${this.internalTimeoutMillis / 1e3} seconds.`);
             isRequestLocked = false; // markRequestHandled succeeded and unlocked the request
             this.stats.finishJob(statisticsId, request.retryCount);
@@ -797,7 +821,8 @@ export class BasicCrawler {
             request.state = RequestState.DONE;
             crawlingContext.session?.markGood();
         }
-        catch (err) {
+        catch (rawError) {
+            const err = this.unwrapError(rawError);
             try {
                 request.state = RequestState.ERROR_HANDLER;
                 await addTimeoutToPromise(async () => this._requestFunctionErrorHandler(err, crawlingContext, source), this.internalTimeoutMillis, `Handling request failure of ${request.url} (${request.id}) timed out after ${this.internalTimeoutMillis / 1e3} seconds.`);
@@ -807,24 +832,24 @@ export class BasicCrawler {
                 request.state = RequestState.DONE;
             }
             catch (secondaryError) {
-                if (!secondaryError.triggeredFromUserHandler &&
+                const unwrappedSecondaryError = this.unwrapError(secondaryError);
+                if (!unwrappedSecondaryError.triggeredFromUserHandler &&
                     // avoid reprinting the same critical error multiple times, as it will be printed by Nodejs at the end anyway
-                    !(secondaryError instanceof CriticalError)) {
+                    !(unwrappedSecondaryError instanceof CriticalError)) {
                     const apifySpecific = process.env.APIFY_IS_AT_HOME
                         ? `This may have happened due to an internal error of Apify's API or due to a misconfigured crawler.`
                         : '';
-                    this.log.exception(secondaryError, 'An exception occurred during handling of failed request. ' +
+                    this.log.exception(unwrappedSecondaryError, 'An exception occurred during handling of failed request. ' +
                         `This places the crawler and its underlying storages into an unknown state and crawling will be terminated. ${apifySpecific}`);
                 }
                 request.state = RequestState.ERROR;
-                throw secondaryError;
+                throw unwrappedSecondaryError;
             }
             // decrease the session score if the request fails (but the error handler did not throw)
             crawlingContext.session?.markBad();
         }
         finally {
-            await this._cleanupContext(crawlingContext);
-            this.crawlingContexts.delete(crawlingContext.id);
+            await Promise.all(deferredCleanup.map((cleanup) => cleanup()));
             // Safety net - release the lock if nobody managed to do it before
             if (isRequestLocked && source instanceof RequestProvider) {
                 try {
@@ -883,6 +908,18 @@ export class BasicCrawler {
         request.sessionRotationCount++;
         crawlingContext.session?.retire();
     }
+    /**
+     * Unwraps errors thrown by the context pipeline to get the actual user error.
+     * RequestHandlerError and ContextPipelineInitializationError wrap the actual error.
+     */
+    unwrapError(error) {
+        if (error instanceof RequestHandlerError ||
+            error instanceof ContextPipelineInitializationError ||
+            error instanceof ContextPipelineCleanupError) {
+            return this.unwrapError(error.cause);
+        }
+        return error;
+    }
     /**
      * Handles errors thrown by user provided requestHandler()
      */
@@ -895,7 +932,8 @@ export class BasicCrawler {
         const shouldRetryRequest = this._canRequestBeRetried(request, error);
         if (shouldRetryRequest) {
             await this.stats.errorTrackerRetry.addAsync(error, crawlingContext);
-            await this.errorHandler?.(crawlingContext, error);
+            await this.errorHandler?.(crawlingContext, // valid cast - ExtendedContext transitively extends CrawlingContext
+            error);
             if (error instanceof SessionError) {
                 await this._rotateSession(crawlingContext);
             }
@@ -947,7 +985,8 @@ export class BasicCrawler {
         const message = this._getMessageFromError(error, true);
         this.log.error(`Request failed and reached maximum retries. ${message}`, { id, url, method, uniqueKey });
         if (this.failedRequestHandler) {
-            await this.failedRequestHandler?.(crawlingContext, error);
+            await this.failedRequestHandler?.(crawlingContext, // valid cast - ExtendedContext transitively extends CrawlingContext
+            error);
         }
     }
     /**
@@ -1013,9 +1052,7 @@ export class BasicCrawler {
      */
     async teardown() {
         this.events.emit("persistState" /* EventType.PERSIST_STATE */, { isMigrating: false });
-        if (this.useSessionPool) {
-            await this.sessionPool.teardown();
-        }
+        await this.sessionPool?.teardown();
         if (this._closeEvents) {
             await this.events.close();
         }