@crawlee/basic 4.0.0-beta.99 → 4.0.0-rc.1

package/internals/sitemap_request_loader.js

@@ -0,0 +1,432 @@
+import { Transform } from 'node:stream';
+import { EnqueueStrategy } from '@crawlee/utils';
+import { parseArgument, parseSitemap, schemas } from '@crawlee/utils/internal';
+import { minimatch } from 'minimatch';
+import { z } from 'zod';
+import { EventType, KeyValueStore, purgeDefaultStorages, Request, RequestManagerTandem, RequestQueue, serviceLocator, } from '@crawlee/core';
+import { constructUrlPatternObjects, urlPatternSchema } from './enqueue_links/index.js';
+const sitemapRequestLoaderOptionsSchema = z.strictObject({
+    sitemapUrls: schemas.arrayOf(z.string(), 'strings'),
+    proxyUrl: z.string().optional(),
+    persistStateKey: z.string().optional(),
+    signal: z.unknown().optional(),
+    timeoutMillis: schemas.anyNumber.optional(),
+    maxBufferSize: schemas.anyNumber.default(200),
+    enqueueStrategy: z.enum(EnqueueStrategy).default(EnqueueStrategy.SameHostname),
+    parseSitemapOptions: z.looseObject({}).optional(),
+    include: schemas.arrayOf(urlPatternSchema, 'URL patterns').optional(),
+    exclude: schemas.arrayOf(urlPatternSchema, 'URL patterns').optional(),
+    persistenceOptions: z.looseObject({}).optional(),
+    httpClient: schemas.httpClient.optional(),
+});
+/** @internal */
+const STATE_PERSISTENCE_KEY = 'SITEMAP_REQUEST_LOADER_STATE';
+/**
+ * A list of URLs to crawl parsed from a sitemap.
+ *
+ * The loading of the sitemap is performed in the background so that crawling can start before the sitemap is fully loaded.
+ */
+export class SitemapRequestLoader {
+    /**
+     * Set of URLs that were returned by `fetchNextRequest()` and not marked as handled yet.
+     * @internal
+     */
+    inProgress = new Set();
+    /**
+     * Map of returned Request objects that have not been marked as handled yet.
+     *
+     * We use this to persist custom user fields on the in-progress requests.
+     */
+    #requestData = new Map();
+    /**
+     * Object for keeping track of the sitemap parsing progress.
+     */
+    #sitemapParsingProgress = {
+        /**
+         * URL of the sitemap that is currently being parsed. `null` if no sitemap is being parsed.
+         */
+        inProgressSitemapUrl: null,
+        /**
+         * Buffer for URLs from the currently parsed sitemap. Used for tracking partially loaded sitemaps across migrations.
+         */
+        inProgressEntries: new Set(),
+        /**
+         * Set of sitemap URLs that have not been parsed yet. If the set is empty and `inProgressSitemapUrl` is `null`, the sitemap loading is finished.
+         */
+        pendingSitemapUrls: new Set(),
+    };
+    /**
+     * Object stream of URLs parsed from the sitemaps.
+     * Using `highWaterMark`, this can manage the speed of the sitemap loading.
+     *
+     * Fetch the next URL to be processed using `fetchNextRequest()`.
+     */
+    #urlQueueStream;
+    /**
+     * Indicates whether the request list sitemap loading was aborted.
+     *
+     * If the loading was aborted before the sitemaps were fully loaded, the request list might be missing some URLs.
+     * The `isSitemapFullyLoaded` method can be used to check if the sitemaps were fully loaded.
+     *
+     * If the loading is aborted and all the requests are handled, `checkReadiness()` will report `finished`.
+     */
+    #abortLoading = false;
+    /** Number of URLs that were marked as handled */
+    #handledUrlCount = 0;
+    #persistStateKey;
+    #store;
+    #closed = false;
+    /**
+     * Proxy URL to be used for sitemap loading.
+     */
+    #proxyUrl;
+    /**
+     * Enqueue strategy applied to sitemap-derived URLs and stamped onto the emitted `Request` objects.
+     */
+    #enqueueStrategy;
+    /**
+     * Logger instance.
+     */
+    #log;
+    #urlExcludePatternObjects = [];
+    #urlPatternObjects = [];
+    /** EventManager used to handle persistence */
+    #events;
+    #persistenceOptions;
+    /** @internal */
+    constructor(options) {
+        const { include, exclude, persistStateKey, persistenceOptions, proxyUrl, maxBufferSize, sitemapUrls, enqueueStrategy, } = parseArgument(options, sitemapRequestLoaderOptionsSchema, 'SitemapRequestLoaderOptions');
+        this.#log = serviceLocator.getLogger().child({ prefix: 'SitemapRequestLoader' });
+        if (exclude?.length) {
+            this.#urlExcludePatternObjects.push(...constructUrlPatternObjects(exclude));
+        }
+        if (include?.length) {
+            this.#urlPatternObjects.push(...constructUrlPatternObjects(include));
+        }
+        this.#persistStateKey = persistStateKey;
+        this.#persistenceOptions = { enable: true, ...persistenceOptions };
+        this.#proxyUrl = proxyUrl;
+        this.#enqueueStrategy = enqueueStrategy;
+        this.#urlQueueStream = this.#createNewStream(maxBufferSize);
+        this.#sitemapParsingProgress.pendingSitemapUrls = new Set(sitemapUrls);
+        this.#events = serviceLocator.getEventManager();
+        this.persistState = this.persistState.bind(this);
+    }
+    /**
+     * Creates a new object stream with the specified highWaterMark.
+     * @param highWaterMark High water mark for the stream (the maximum number of objects the stream will buffer).
+     * @returns A new object stream.
+     */
+    #createNewStream(highWaterMark) {
+        return new Transform({
+            objectMode: true,
+            highWaterMark,
+        }).pause();
+    }
+    /**
+     * Returns a function that checks whether the provided pattern matches the closure URL.
+     * @param url URL to be checked.
+     * @returns A matcher function that checks whether the pattern matches the closure URL.
+     */
+    #matchesUrl(url) {
+        return (patternObject) => {
+            const { regexp, glob } = patternObject;
+            const matchesRegex = (regexp && url.match(regexp)) || false;
+            const matchesGlob = (glob && minimatch(url, glob, { nocase: true })) || false;
+            return Boolean(matchesRegex || matchesGlob);
+        };
+    }
+    /**
+     * Checks whether the URL matches the `include` / `exclude` patterns provided in the `options`.
+     * @param url URL to be checked.
+     * @returns `true` if the URL matches the patterns, `false` otherwise.
+     */
+    #isUrlMatchingPatterns(url) {
+        return (!this.#urlExcludePatternObjects.some(this.#matchesUrl(url)) &&
+            (this.#urlPatternObjects.length === 0 || this.#urlPatternObjects.some(this.#matchesUrl(url))));
+    }
+    /**
+     * Adds a URL to the queue of parsed URLs.
+     *
+     * Blocks if the stream is full until it is drained.
+     */
+    async #pushNextUrl(url) {
+        return new Promise((resolve) => {
+            if (this.#closed || (url && !this.#isUrlMatchingPatterns(url))) {
+                resolve();
+                return;
+            }
+            if (!this.#urlQueueStream.push(url)) {
+                // This doesn't work with the 'drain' event (it's not emitted for some reason).
+                this.#urlQueueStream.once('readdata', () => {
+                    resolve();
+                });
+            }
+            else {
+                resolve();
+            }
+        });
+    }
+    /**
+     * Reads the next URL from the queue of parsed URLs.
+     *
+     * If the stream is empty, blocks until a new URL is pushed.
+     * @returns The next URL from the queue or `null` if we have read all URLs.
+     */
+    async #readNextUrl() {
+        return new Promise((resolve) => {
+            if (this.#closed) {
+                resolve(null);
+                return;
+            }
+            const result = this.#urlQueueStream.read();
+            if (!result && !this.isSitemapFullyLoaded()) {
+                this.#urlQueueStream.once('readable', () => {
+                    const nextUrl = this.#urlQueueStream.read();
+                    resolve(nextUrl);
+                });
+            }
+            else {
+                resolve(result);
+            }
+            this.#urlQueueStream.emit('readdata');
+        });
+    }
+    /**
+     * Indicates whether the background processing of sitemap contents has successfully finished.
+     *
+     * If this is `false`, the background processing is either still in progress or was aborted.
+     */
+    isSitemapFullyLoaded() {
+        return (this.#sitemapParsingProgress.inProgressSitemapUrl === null &&
+            this.#sitemapParsingProgress.pendingSitemapUrls.size === 0);
+    }
+    /**
+     * Start processing the sitemaps and loading the URLs.
+     *
+     * Resolves once all the sitemaps URLs have been fully loaded (sets `isSitemapFullyLoaded` to `true`).
+     */
+    async #load({ parseSitemapOptions, }) {
+        while (!this.isSitemapFullyLoaded() && !this.#abortLoading) {
+            const sitemapUrl = this.#sitemapParsingProgress.inProgressSitemapUrl ??
+                this.#sitemapParsingProgress.pendingSitemapUrls.values().next().value;
+            try {
+                for await (const item of parseSitemap([{ type: 'url', url: sitemapUrl }], this.#proxyUrl, {
+                    ...parseSitemapOptions,
+                    maxDepth: 0,
+                    emitNestedSitemaps: true,
+                    enqueueStrategy: this.#enqueueStrategy,
+                })) {
+                    if (!item.originSitemapUrl) {
+                        // This is a nested sitemap
+                        this.#sitemapParsingProgress.pendingSitemapUrls.add(item.loc);
+                        continue;
+                    }
+                    if (!this.#sitemapParsingProgress.inProgressEntries.has(item.loc)) {
+                        await this.#pushNextUrl(item.loc);
+                        this.#sitemapParsingProgress.inProgressEntries.add(item.loc);
+                    }
+                }
+            }
+            catch (e) {
+                this.#log.error('Error loading sitemap contents:', e);
+            }
+            this.#sitemapParsingProgress.pendingSitemapUrls.delete(sitemapUrl);
+            this.#sitemapParsingProgress.inProgressEntries.clear();
+            this.#sitemapParsingProgress.inProgressSitemapUrl = null;
+        }
+        this.#urlQueueStream.end();
+    }
+    /**
+     * Open a sitemap and start processing it.
+     *
+     * Resolves to a new instance of `SitemapRequestLoader`, which **might not be fully loaded yet** - i.e. the sitemap might still be loading in the background.
+     *
+     * Track the loading progress using the `isSitemapFullyLoaded` property.
+     */
+    static async open(options) {
+        const { httpClient, ...restOptions } = options;
+        const requestList = new SitemapRequestLoader({
+            ...restOptions,
+            persistStateKey: options.persistStateKey ?? STATE_PERSISTENCE_KEY,
+        });
+        await requestList.#restoreState();
+        void requestList.#load({
+            parseSitemapOptions: { logger: serviceLocator.getLogger(), ...options.parseSitemapOptions, httpClient },
+        });
+        if (requestList.#persistenceOptions.enable) {
+            requestList.#events.on(EventType.PERSIST_STATE, requestList.persistState);
+        }
+        options?.signal?.addEventListener('abort', () => {
+            requestList.#abortLoading = true;
+        });
+        if (options.timeoutMillis) {
+            setTimeout(() => {
+                requestList.#abortLoading = true;
+            }, options.timeoutMillis);
+        }
+        return requestList;
+    }
+    /**
+     * @inheritDoc
+     */
+    async getTotalCount() {
+        // Total known so far = not-yet-fetched (still buffered in the stream) + in-progress (fetched but not
+        // yet handled) + already handled.
+        return this.#urlQueueStream.readableLength + this.inProgress.size + this.#handledUrlCount;
+    }
+    /**
+     * @inheritDoc
+     */
+    async getPendingCount() {
+        // Pending = everything not yet handled = not-yet-fetched + in-progress.
+        return this.#urlQueueStream.readableLength + this.inProgress.size;
+    }
+    /**
+     * Combines this list with a request manager (a {@link RequestQueue} by default) into a
+     * {@link RequestManagerTandem}, allowing requests to be added and reclaimed while still
+     * being read from this list first.
+     */
+    async toTandem(requestManager) {
+        return new RequestManagerTandem(this, requestManager ?? (await RequestQueue.open()));
+    }
+    /**
+     * @inheritDoc
+     */
+    async checkReadiness() {
+        if (this.#urlQueueStream.readableLength > 0) {
+            return { status: 'ready' };
+        }
+        // The parser is still running, so more URLs may yet arrive - on no schedule, hence no `readyAt`.
+        if (!this.isSitemapFullyLoaded() && !this.#abortLoading) {
+            return { status: 'waiting' };
+        }
+        return this.inProgress.size === 0 ? { status: 'finished' } : { status: 'waiting' };
+    }
+    /**
+     * @inheritDoc
+     */
+    async getHandledCount() {
+        return this.#handledUrlCount;
+    }
+    /**
+     * @inheritDoc
+     */
+    async persistState() {
+        if (this.#persistStateKey === undefined) {
+            return;
+        }
+        this.#store ??= await KeyValueStore.open();
+        const urlQueue = [];
+        while (this.#urlQueueStream.readableLength > 0) {
+            const url = this.#urlQueueStream.read();
+            if (url === null) {
+                break;
+            }
+            urlQueue.push(url);
+        }
+        // Create a new stream, as we have read all the URLs from the current one.
+        // Pushing the urls back to the original stream might not be possible if it has been ended.
+        const previousStream = this.#urlQueueStream;
+        const newStream = this.#createNewStream(previousStream.readableHighWaterMark);
+        for (const url of urlQueue) {
+            newStream.push(url);
+        }
+        if (previousStream.writableEnded) {
+            newStream.end();
+        }
+        this.#urlQueueStream = newStream;
+        // A `pushNextUrl()` call may be blocked on backpressure, waiting for a `readdata` event on the
+        // previous stream. That event is only ever emitted by `readNextUrl()` on the current stream, so
+        // after the swap the waiter would never be notified and the background sitemap loading would hang.
+        // Re-emit `readdata` on the previous stream to release any such pending waiter (its URL has already
+        // been transferred to the new stream above).
+        previousStream.emit('readdata');
+        await this.#store.setValue(this.#persistStateKey, {
+            sitemapParsingProgress: {
+                pendingSitemapUrls: Array.from(this.#sitemapParsingProgress.pendingSitemapUrls),
+                inProgressSitemapUrl: this.#sitemapParsingProgress.inProgressSitemapUrl,
+                inProgressEntries: Array.from(this.#sitemapParsingProgress.inProgressEntries),
+            },
+            // Re-queue in-progress requests to the front so they are retried if the state is restored.
+            urlQueue: [...this.inProgress, ...urlQueue],
+            requestData: Array.from(this.#requestData.entries()),
+            abortLoading: this.#abortLoading,
+            closed: this.#closed,
+        });
+    }
+    async #restoreState() {
+        await purgeDefaultStorages({ onlyPurgeOnce: true });
+        if (this.#persistStateKey === undefined) {
+            return;
+        }
+        this.#store ??= await KeyValueStore.open();
+        const state = await this.#store.getValue(this.#persistStateKey);
+        if (state === null) {
+            return;
+        }
+        this.#sitemapParsingProgress = {
+            pendingSitemapUrls: new Set(state.sitemapParsingProgress.pendingSitemapUrls),
+            inProgressSitemapUrl: state.sitemapParsingProgress.inProgressSitemapUrl,
+            inProgressEntries: new Set(state.sitemapParsingProgress.inProgressEntries),
+        };
+        this.#requestData = new Map(state.requestData ?? []);
+        for (const url of state.urlQueue) {
+            this.#urlQueueStream.push(url);
+        }
+        this.#abortLoading = state.abortLoading;
+        this.#closed = state.closed;
+    }
+    /**
+     * @inheritDoc
+     */
+    async fetchNextRequest() {
+        const nextUrl = await this.#readNextUrl();
+        if (!nextUrl) {
+            return null;
+        }
+        // A restored in-progress request already has its Request data; don't overwrite it.
+        if (!this.#requestData.has(nextUrl)) {
+            this.#requestData.set(nextUrl, new Request({ url: nextUrl, enqueueStrategy: this.#enqueueStrategy }));
+        }
+        this.inProgress.add(nextUrl);
+        return this.#requestData.get(nextUrl);
+    }
+    /**
+     * @inheritDoc
+     */
+    async *[Symbol.asyncIterator]() {
+        while ((await this.checkReadiness()).status !== 'finished') {
+            const request = await this.fetchNextRequest();
+            if (!request)
+                break;
+            yield request;
+        }
+    }
+    /**
+     * Aborts the internal sitemap loading, stops the processing of the sitemap contents and drops all the pending URLs.
+     *
+     * Calling `fetchNextRequest()` after this method will always return `null`.
+     */
+    async teardown() {
+        this.#closed = true;
+        this.#abortLoading = true;
+        this.#events.off(EventType.PERSIST_STATE, this.persistState);
+        await this.persistState();
+        this.#urlQueueStream.emit('readdata'); // unblocks the potentially waiting `pushNextUrl` call
+    }
+    /**
+     * @inheritDoc
+     */
+    async markRequestAsHandled(request) {
+        this.#handledUrlCount += 1;
+        this.#ensureInProgress(request.url);
+        this.inProgress.delete(request.url);
+        this.#requestData.delete(request.url);
+    }
+    #ensureInProgress(url) {
+        if (!this.inProgress.has(url)) {
+            throw new Error(`The request is not being processed (url: ${url})`);
+        }
+    }
+}

package/internals/throttling_request_manager.d.ts

@@ -0,0 +1,229 @@
+import type { Dictionary } from '@crawlee/types';
+import type { AddRequestsBatchedOptions, AddRequestsBatchedResult, Configuration, IRequestManager, PacingSignal, RequestQueueOperationInfo, RequestQueueOperationOptions, RequestsLike, Source, StorageIdentifier, RequestSourceStatus, StorageOpenOptions } from '@crawlee/core';
+import { Request } from '@crawlee/core';
+/**
+ * Opens a request manager, matching the shape of storage `open` methods such as
+ * {@link RequestQueue.open|`RequestQueue.open`}.
+ *
+ * {@link ThrottlingRequestManager} calls this once per configured domain, so every per-domain queue shares the
+ * concrete type and storage backend of the manager being wrapped.
+ *
+ * Not exported: the only option that takes one is `@internal`.
+ */
+type RequestManagerOpener<T extends IRequestManager = IRequestManager> = (identifier?: string | StorageIdentifier | null, options?: StorageOpenOptions) => Promise<T>;
+/** Options for {@link ThrottlingRequestManager}. */
+export interface ThrottlingRequestManagerOptions<T extends IRequestManager = IRequestManager> {
+    /**
+     * The request manager to wrap, usually a {@link RequestQueue}. Requests for domains that are not throttled
+     * are stored here. May be a factory, so that the throttler can be constructed synchronously and the manager
+     * under it opened only on first use.
+     *
+     * Omitted, the default request queue is opened on first use through
+     * {@link ThrottlingRequestManagerOptions.requestManagerOpener|`requestManagerOpener`}.
+     */
+    inner?: T | (() => T | Promise<T>);
+    /**
+     * Which domains to throttle: a list of hostnames, or `'all'` for every domain the crawl encounters.
+     *
+     * Matching a listed hostname is case-insensitive and exact - wildcards such as `*.example.com` are not
+     * supported, so list each subdomain you care about (or set
+     * {@link ThrottlingRequestManagerOptions.throttleBy|`throttleBy: 'registrableDomain'`}). An
+     * internationalized domain may be given in either its unicode or its punycode form, and an IPv6 address has
+     * to be bracketed (`[::1]`). Requests for any other domain bypass throttling entirely.
+     *
+     * `'all'` gives each domain a queue of its own the first time it is seen, so that it can be held back
+     * without its requests being repeatedly popped and re-enqueued. One request queue per domain is not free,
+     * which is what {@link ThrottlingRequestManagerOptions.maxThrottledDomains|`maxThrottledDomains`} is
+     * there to bound.
+     */
+    domains: string[] | 'all';
+    /**
+     * A floor under the crawl delay of every throttled domain, in seconds - the proactive clock described on
+     * {@link ThrottlingRequestManager}. A domain whose robots.txt asks for a longer `Crawl-delay` gets the
+     * longer one; this is a minimum, not an override. A `minIntervalEverywhere` {@link PacingSignal} raises
+     * this floor at runtime, and never lowers it.
+     * @default 0
+     */
+    minCrawlDelaySecs?: number;
+    /**
+     * What counts as "the same domain": the exact hostname, or the registrable domain it belongs to
+     * (`example.com` for `www.example.com`, `a.example.co.uk` and so on). Hosts with no registrable domain -
+     * IP addresses, `localhost` - are always throttled per hostname.
+     *
+     * Grouping by registrable domain gives subdomains a single pair of clocks and a single queue, which is what
+     * you want when the pacing is there to be polite to one server rather than to satisfy a specific host's
+     * rate limit.
+     * @default 'hostname'
+     */
+    throttleBy?: 'hostname' | 'registrableDomain';
+    /**
+     * The most domains a run may throttle at once. Exceeding it throws, rather than silently letting the
+     * throttling lapse - one request queue per domain is not free, and a crawl that discovers domains without
+     * bound would drown the storage backend in them.
+     *
+     * Only domains discovered under `domains: 'all'` count against this; an explicit list is taken at face value.
+     * @default 100
+     */
+    maxThrottledDomains?: number;
+    /**
+     * The key under which the discovered domain list is kept in the default key-value store, so that a restart
+     * with `purgeOnStart` disabled reopens their queues instead of stranding whatever they still hold. Only
+     * written under `domains: 'all'`.
+     *
+     * Give each manager its own key when running several of them against the same storage.
+     * @default 'CRAWLEE_THROTTLED_DOMAINS'
+     */
+    persistStateKey?: string;
+    /**
+     * Opens the per-domain queues, one per throttled domain, each under the alias `throttled-<domain>`.
+     * @default RequestQueue.open
+     * @internal
+     */
+    requestManagerOpener?: RequestManagerOpener<T>;
+    /**
+     * The delay applied after a domain's first HTTP 429, doubled on each subsequent one.
+     * @default 2
+     */
+    baseDelaySecs?: number;
+    /**
+     * Upper bound on the delay between requests to a rate-limited domain, applied to both the exponential
+     * backoff and a `Retry-After` value.
+     * @default 60
+     */
+    maxDelaySecs?: number;
+    /**
+     * How long a domain may rate-limit us without a single request getting through before the crawl is
+     * abandoned with a {@link PersistentRateLimitError}.
+     *
+     * A domain that keeps answering 429 for this long is not going to be crawled by waiting longer - the
+     * concurrency is too high for it, or it has blocked us outright. Its requests are deliberately left in
+     * their queue, so re-running the crawl with `purgeOnStart` disabled picks them up once the domain recovers.
+     *
+     * A crawler running with `keepAlive` is exempt - outliving a domain that will not let us through is the
+     * whole point there.
+     * @default 900
+     */
+    maxDomainStallSecs?: number;
+}
+/**
+ * A request manager that wraps another one and paces requests per domain.
+ *
+ * Requests for a throttled domain are routed into their own queue when they are added, so each request lives in
+ * exactly one place and deduplication keeps working. Everything else goes to the wrapped manager untouched.
+ *
+ * {@link ThrottlingRequestManager.fetchNextRequest|`fetchNextRequest()`} serves the domain that has been waiting
+ * longest and skips any that are backing off, falling back to the wrapped manager. It never blocks: while every
+ * remaining request belongs to a throttled domain it returns `null` and
+ * {@link ThrottlingRequestManager.checkReadiness|`checkReadiness()`} reports `waiting` with the moment the
+ * earliest of them comes due, so the crawler idles instead of holding a concurrency slot open.
+ *
+ * Each throttled domain runs two independent clocks, and may be dispatched to once **both** have run out:
+ * - **Backoff**, set by HTTP 429 responses - honouring `Retry-After`, and otherwise doubling from `baseDelaySecs`.
+ *   Reactive and temporary: it decays once the domain stops turning us away. The crawlers report the 429s
+ *   themselves; a request held back this way is retried later without counting against `maxRequestRetries` and
+ *   without penalising its session.
+ * - **Crawl delay**, the minimum interval between two dispatches to the domain, armed after each one. Proactive
+ *   and constant: whatever the domain's robots.txt asks for, floored by
+ *   {@link ThrottlingRequestManagerOptions.minCrawlDelaySecs|`minCrawlDelaySecs`}. Either may be absent, in
+ *   which case the other one is the delay.
+ *
+ * Which domains get those clocks is {@link ThrottlingRequestManagerOptions.domains|`domains`} - a list, or
+ * `'all'` for every domain the crawl encounters.
+ *
+ * Pass one as a crawler's `requestManager`; the `sameDomainDelaySecs` shorthand builds one with `domains: 'all'`
+ * and `throttleBy: 'registrableDomain'`. Construct it yourself to name the domains or tune the delays - one
+ * covering every domain also makes `sameDomainDelaySecs` land on it as a floor rather than adding a second pacer.
+ *
+ * Signals - 429s, robots.txt `Crawl-delay`, that floor - arrive through
+ * {@link IRequestManager.recordPacingSignal|`recordPacingSignal`}, which wrapping managers forward, so this
+ * works wherever it sits in a composition, including inside a {@link RequestManagerTandem}.
+ *
+ * **Example usage:**
+ *
+ * ```ts
+ * const crawler = new CheerioCrawler({
+ *     requestManager: new ThrottlingRequestManager({
+ *         domains: ['api.example.com', 'slow-site.org'],
+ *     }),
+ *     requestHandler: async ({ request }) => { ... },
+ * });
+ * ```
+ *
+ * @category Sources
+ */
+export declare class ThrottlingRequestManager<T extends IRequestManager = IRequestManager> implements IRequestManager {
+    #private;
+    constructor(options: ThrottlingRequestManagerOptions<T>, config?: Configuration);
+    /**
+     * The wrapped manager, holding every request whose domain is not throttled. `undefined` until an `inner`
+     * passed as a factory is resolved - reading this never forces it, because a getter should not open a queue
+     * behind a caller's back.
+     *
+     * @internal
+     */
+    get innerManager(): T | undefined;
+    /**
+     * Records a pacing signal: a refusal puts the URL's domain into backoff, a declared interval becomes its
+     * crawl delay, and a crawl-wide floor raises {@link recordEverywhereFloor|the floor under all of them}.
+     *
+     * @returns `false` if the domain the signal covers is not throttled, in which case this is a no-op.
+     * @throws If the signal's scope is one this manager cannot honour - see {@link assertScopeHonourable}.
+     * @inheritdoc
+     */
+    recordPacingSignal(signal: PacingSignal): boolean;
+    addRequest(requestLike: Source, options?: RequestQueueOperationOptions): Promise<RequestQueueOperationInfo>;
+    /**
+     * Adds requests in batches, routing each one to the manager that owns its domain.
+     *
+     * Batching, validation, deduplication and `Retry-After`-free bookkeeping are all delegated to the target
+     * managers - this only decides where each request goes, one batch at a time, so a lazy or unbounded input
+     * iterable is never fully materialized.
+     */
+    addRequestsBatched(requests: RequestsLike, options?: AddRequestsBatchedOptions): Promise<AddRequestsBatchedResult>;
+    reclaimRequest(request: Request, options?: RequestQueueOperationOptions): Promise<RequestQueueOperationInfo | null>;
+    markRequestAsHandled(request: Request): Promise<RequestQueueOperationInfo | void | null>;
+    getTotalCount(): Promise<number>;
+    getPendingCount(): Promise<number>;
+    getHandledCount(): Promise<number>;
+    /**
+     * Reports whether anything can be dispatched right now, and if not, when — or why never.
+     *
+     * One traversal of the domain clocks answers all of it: only domains whose delays have run out are probed,
+     * the rest merely contribute the moment they come due. Throttled requests count as outstanding work, so a
+     * crawler gated on this idles for the backoff instead of concluding it is done.
+     *
+     * `ready` from anywhere else outranks a stalling domain and is returned without looking at the stall clocks,
+     * so one hopeless domain never ends a crawl making progress elsewhere. It cannot outrank itself, though -
+     * see the traversal.
+     */
+    checkReadiness(): Promise<RequestSourceStatus>;
+    /**
+     * Empties every manager and clears the accumulated backoff. A robots.txt `Crawl-delay` is a property of the
+     * site rather than of the run, so it survives.
+     */
+    purge(): Promise<void>;
+    setExpectedRequestProcessingTimeSecs(secs: number): Promise<void>;
+    /**
+     * @inheritdoc
+     * Extending a lock must not consume the in-flight marker that later routes
+     * {@link ThrottlingRequestManager.markRequestAsHandled} / {@link ThrottlingRequestManager.reclaimRequest}.
+     */
+    extendRequestProcessingTimeSecs(request: Request, secs: number): Promise<boolean>;
+    /**
+     * Returns the next request from a domain that is not backing off, or from the inner manager.
+     *
+     * Returns `null` while every remaining request belongs to a throttled domain - it never waits the backoff
+     * out, because a consumer parked in here holds a concurrency slot, which the autoscaler reads as spare
+     * capacity and answers by scaling up. Callers poll instead, and
+     * {@link ThrottlingRequestManager.checkReadiness|`checkReadiness()`} reports `waiting` meanwhile so the
+     * crawler's task loop idles rather than spins.
+     */
+    fetchNextRequest<R extends Dictionary = Dictionary>(): Promise<Request<R> | null>;
+    [Symbol.asyncIterator](): AsyncGenerator<Request<Dictionary>, void, unknown>;
+    /**
+     * Drops every managed storage - the wrapped manager and each per-domain queue.
+     * @internal
+     */
+    drop(): Promise<void>;
+}
+export {};