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

package/crawlers/error_tracker.js

@@ -1,308 +0,0 @@
-import { inspect } from 'node:util';
-import { ErrorSnapshotter } from './error_snapshotter.js';
-const extractPathFromStackTraceLine = (line) => {
-    const lastStartingRoundBracketIndex = line.lastIndexOf('(');
-    if (lastStartingRoundBracketIndex !== -1) {
-        const closingRoundBracketIndex = line.indexOf(')', lastStartingRoundBracketIndex);
-        if (closingRoundBracketIndex !== -1) {
-            return line.slice(lastStartingRoundBracketIndex + 1, closingRoundBracketIndex);
-        }
-    }
-    return line;
-};
-// https://v8.dev/docs/stack-trace-api#appendix%3A-stack-trace-format
-const getPathFromStackTrace = (stack) => {
-    for (const line of stack) {
-        const path = extractPathFromStackTraceLine(line);
-        if (path.startsWith('node:') || path.includes('/node_modules/') || path.includes('\\node_modules\\')) {
-            continue;
-        }
-        return path;
-    }
-    return extractPathFromStackTraceLine(stack[0]);
-};
-const getStackTraceGroup = (error, storage, showFullStack) => {
-    const stack = error.stack?.split('\n').map((line) => line.trim());
-    let sliceAt = -1;
-    if (stack) {
-        for (let i = 0; i < stack.length; i++) {
-            if (stack[i].startsWith('at ') || stack[i].startsWith('eval at ')) {
-                sliceAt = i;
-                break;
-            }
-        }
-    }
-    let normalizedStackTrace = null;
-    if (sliceAt !== -1) {
-        normalizedStackTrace = showFullStack
-            ? stack
-                .slice(sliceAt)
-                .map((x) => x.trim())
-                .join('\n')
-            : getPathFromStackTrace(stack.slice(sliceAt));
-    }
-    if (!normalizedStackTrace) {
-        normalizedStackTrace = 'missing stack trace';
-    }
-    if (!(normalizedStackTrace in storage)) {
-        storage[normalizedStackTrace] = Object.create(null);
-    }
-    return storage[normalizedStackTrace];
-};
-const getErrorCodeGroup = (error, storage) => {
-    let { code } = error;
-    if (code === undefined) {
-        code = 'missing error code';
-    }
-    if (!(code in storage)) {
-        storage[code] = Object.create(null);
-    }
-    return storage[String(code)];
-};
-const getErrorNameGroup = (error, storage) => {
-    const { name } = error;
-    if (!(name in storage)) {
-        storage[name] = Object.create(null);
-    }
-    return storage[name];
-};
-const findBiggestWordIntersection = (a, b) => {
-    let maxStreak = 0;
-    let bStreakIndex = -1;
-    let aStreakIndex = -1;
-    for (let aIndex = 0; aIndex < a.length; aIndex++) {
-        let bIndex = -1;
-        do {
-            let aWalkIndex = aIndex;
-            bIndex = b.indexOf(a[aIndex], bIndex + 1);
-            let bWalkIndex = bIndex;
-            let streak = 0;
-            while (aWalkIndex < a.length && bWalkIndex < b.length && b[bWalkIndex++] === a[aWalkIndex++]) {
-                streak++;
-            }
-            if (streak > maxStreak) {
-                maxStreak = streak;
-                aStreakIndex = aIndex;
-                bStreakIndex = bIndex;
-            }
-        } while (bIndex !== -1);
-    }
-    return {
-        maxStreak,
-        aStreakIndex,
-        bStreakIndex,
-    };
-};
-const arrayCount = (array, target) => {
-    let result = 0;
-    for (const item of array) {
-        if (item === target) {
-            result++;
-        }
-    }
-    return result;
-};
-const calculatePlaceholder = (a, b) => {
-    const { maxStreak, aStreakIndex, bStreakIndex } = findBiggestWordIntersection(a, b);
-    if (maxStreak === 0) {
-        return ['_'];
-    }
-    const leftA = a.slice(0, aStreakIndex);
-    const leftB = b.slice(0, bStreakIndex);
-    const rightA = a.slice(aStreakIndex + maxStreak);
-    const rightB = b.slice(bStreakIndex + maxStreak);
-    const output = [];
-    if (leftA.length !== 0 || leftB.length !== 0) {
-        output.push(...calculatePlaceholder(leftA, leftB));
-    }
-    output.push(...a.slice(aStreakIndex, aStreakIndex + maxStreak));
-    if (rightA.length !== 0 || rightB.length !== 0) {
-        output.push(...calculatePlaceholder(rightA, rightB));
-    }
-    return output;
-};
-const normalizedCalculatePlaceholder = (a, b) => {
-    const output = calculatePlaceholder(a, b);
-    // We can't be too general
-    if (arrayCount(output, '_') / output.length >= 0.5) {
-        return ['_'];
-    }
-    return output;
-};
-// Merge A (missing placeholders) into B (can contain placeholders but does not have to)
-const mergeMessages = (a, b, storage) => {
-    const placeholder = normalizedCalculatePlaceholder(a.split(' '), b.split(' ')).join(' ');
-    if (placeholder === '_') {
-        return undefined;
-    }
-    const count = storage[a].count + storage[b].count;
-    delete storage[a];
-    delete storage[b];
-    storage[placeholder] = Object.assign(Object.create(null), {
-        count,
-    });
-    return placeholder;
-};
-const getErrorMessageGroup = (error, storage, showFullMessage) => {
-    let { message } = error;
-    if (!message) {
-        try {
-            message =
-                typeof error === 'string'
-                    ? error
-                    : `Unknown error message. Received non-error object: ${JSON.stringify(error)}`;
-        }
-        catch {
-            message = `Unknown error message. Received non-error object, and could not stringify it: ${inspect(error, {
-                depth: 0,
-            })}`;
-        }
-    }
-    if (!showFullMessage) {
-        const newLineIndex = message.indexOf('\n');
-        message = message.slice(0, newLineIndex === -1 ? undefined : newLineIndex);
-    }
-    if (!(message in storage)) {
-        storage[message] = Object.assign(Object.create(null), {
-            count: 0,
-        });
-        // This actually safe, since we Object.create(null) so no prototype pollution can happen.
-        // eslint-disable-next-line no-restricted-syntax, guard-for-in
-        for (const existingMessage in storage) {
-            const newMessage = mergeMessages(message, existingMessage, storage);
-            if (newMessage) {
-                message = newMessage;
-                break;
-            }
-        }
-    }
-    return storage[message];
-};
-const increaseCount = (group) => {
-    if (!('count' in group)) {
-        // In case users don't want to display error message
-        group.count = 0;
-    }
-    group.count++;
-};
-/**
- * This class tracks errors and computes a summary of information like:
- * - where the errors happened
- * - what the error names are
- * - what the error codes are
- * - what is the general error message
- *
- * This is extremely useful when there are dynamic error messages, such as argument validation.
- *
- * Since the structure of the `tracker.result` object differs when using different options,
- * it's typed as `Record<string, unknown>`. The most deep object has a `count` property, which is a number.
- *
- * It's possible to get the total amount of errors via the `tracker.total` property.
- */
-export class ErrorTracker {
-    #options;
-    result;
-    total;
-    errorSnapshotter;
-    constructor(options = {}) {
-        this.#options = {
-            showErrorCode: true,
-            showErrorName: true,
-            showStackTrace: true,
-            showFullStack: false,
-            showErrorMessage: true,
-            showFullMessage: false,
-            saveErrorSnapshots: false,
-            ...options,
-        };
-        if (this.#options.saveErrorSnapshots) {
-            this.errorSnapshotter = new ErrorSnapshotter();
-        }
-        this.result = Object.create(null);
-        this.total = 0;
-    }
-    updateGroup(error) {
-        let group = this.result;
-        if (this.#options.showStackTrace) {
-            group = getStackTraceGroup(error, group, this.#options.showFullStack);
-        }
-        if (this.#options.showErrorCode) {
-            group = getErrorCodeGroup(error, group);
-        }
-        if (this.#options.showErrorName) {
-            group = getErrorNameGroup(error, group);
-        }
-        if (this.#options.showErrorMessage) {
-            group = getErrorMessageGroup(error, group, this.#options.showFullMessage);
-        }
-        increaseCount(group);
-        return group;
-    }
-    add(error) {
-        this.total++;
-        this.updateGroup(error);
-        if (typeof error.cause === 'object' && error.cause !== null) {
-            this.add(error.cause);
-        }
-    }
-    /**
-     * This method is async, because it captures a snapshot of the error context.
-     * We added this new method to avoid breaking changes.
-     */
-    async addAsync(error, context) {
-        this.total++;
-        const group = this.updateGroup(error);
-        // Capture a snapshot (screenshot and HTML) on the first occurrence of an error
-        if (group.count === 1 && context) {
-            await this.captureSnapshot(group, error, context).catch(() => { });
-        }
-        if (typeof error.cause === 'object' && error.cause !== null) {
-            await this.addAsync(error.cause);
-        }
-    }
-    getUniqueErrorCount() {
-        let count = 0;
-        const goDeeper = (group) => {
-            if ('count' in group) {
-                count++;
-                return;
-            }
-            // eslint-disable-next-line guard-for-in, no-restricted-syntax
-            for (const key in group) {
-                goDeeper(group[key]);
-            }
-        };
-        goDeeper(this.result);
-        return count;
-    }
-    getMostPopularErrors(count) {
-        const result = [];
-        const goDeeper = (group, path) => {
-            if ('count' in group) {
-                result.push([group.count, path]);
-                return;
-            }
-            // eslint-disable-next-line guard-for-in, no-restricted-syntax
-            for (const key in group) {
-                goDeeper(group[key], [...path, key]);
-            }
-        };
-        goDeeper(this.result, []);
-        return result.sort((a, b) => b[0] - a[0]).slice(0, count);
-    }
-    async captureSnapshot(storage, error, context) {
-        if (!this.errorSnapshotter) {
-            return;
-        }
-        const { screenshotFileUrl, htmlFileUrl } = await this.errorSnapshotter.captureSnapshot(error, context);
-        storage.firstErrorScreenshotUrl = screenshotFileUrl;
-        storage.firstErrorHtmlUrl = htmlFileUrl;
-    }
-    reset() {
-        // This actually safe, since we Object.create(null) so no prototype pollution can happen.
-        // eslint-disable-next-line no-restricted-syntax, guard-for-in
-        for (const key in this.result) {
-            delete this.result[key];
-        }
-    }
-}

package/crawlers/index.d.ts

@@ -1,5 +0,0 @@
-export * from './context_pipeline.js';
-export * from './crawler_commons.js';
-export * from './statistics.js';
-export * from './error_tracker.js';
-export * from './error_snapshotter.js';

package/crawlers/index.js

@@ -1,5 +0,0 @@
-export * from './context_pipeline.js';
-export * from './crawler_commons.js';
-export * from './statistics.js';
-export * from './error_tracker.js';
-export * from './error_snapshotter.js';

package/crawlers/internals/types.d.ts

@@ -1,7 +0,0 @@
-export interface BrowserPage {
-    content: () => Promise<string>;
-}
-export interface SnapshottableProperties {
-    body?: unknown;
-    page?: BrowserPage;
-}

package/crawlers/statistics.d.ts

@@ -1,209 +0,0 @@
-import type { CrawleeLogger } from '../log.js';
-import { KeyValueStore } from '../storages/key_value_store.js';
-import { ErrorTracker } from './error_tracker.js';
-/**
- * Persistence-related options to control how and when crawler's data gets persisted.
- */
-export interface PersistenceOptions {
-    /**
-     * Use this flag to disable or enable periodic persistence to key value store.
-     * @default true
-     */
-    enable?: boolean;
-}
-/**
- * The statistics class provides an interface to collecting and logging run
- * statistics for requests.
- *
- * All statistic information is saved on key value store
- * under the key `CRAWLEE_CRAWLER_STATISTICS_*`, persists between
- * migrations and abort/resurrect
- *
- * @category Crawlers
- */
-export declare class Statistics {
-    private static id;
-    /**
-     * An error tracker for final retry errors.
-     */
-    readonly errorTracker: ErrorTracker;
-    /**
-     * An error tracker for retry errors prior to the final retry.
-     */
-    readonly errorTrackerRetry: ErrorTracker;
-    /**
-     * Statistic instance id.
-     */
-    readonly id: string;
-    /**
-     * Current statistic state used for doing calculations on {@link Statistics.calculate} calls
-     */
-    state: StatisticState;
-    /**
-     * Contains the current retries histogram. Index 0 means 0 retries, index 2, 2 retries, and so on
-     */
-    readonly requestRetryHistogram: number[];
-    protected keyValueStore?: KeyValueStore;
-    protected readonly persistStateKey: string;
-    private logIntervalMillis;
-    private logMessage;
-    private listener;
-    private requestsInProgress;
-    private readonly log;
-    private instanceStart;
-    private logInterval;
-    private _events?;
-    private persistenceOptions;
-    private get events();
-    /**
-     * @internal
-     */
-    constructor(options?: StatisticsOptions);
-    /**
-     * Set the current statistic instance to pristine values
-     */
-    reset(): void;
-    /**
-     * @param options - Override the persistence options provided in the constructor
-     */
-    resetStore(options?: PersistenceOptions): Promise<void>;
-    /**
-     * Increments the status code counter.
-     */
-    registerStatusCode(code: number): void;
-    /**
-     * Starts a job
-     * @ignore
-     */
-    startJob(id: number | string): void;
-    /**
-     * Mark job as finished and sets the state
-     * @ignore
-     */
-    finishJob(id: number | string, retryCount: number): void;
-    /**
-     * Mark job as failed and sets the state
-     * @ignore
-     */
-    failJob(id: number | string, retryCount: number): void;
-    /**
-     * Discards a started job without affecting the finished/failed counters, e.g. when a request
-     * turns out to be skipped (robots.txt, enqueue strategy) after `startJob` was already called for it.
-     * @ignore
-     */
-    discardJob(id: number | string): void;
-    /**
-     * Calculate the current statistics
-     */
-    calculate(): {
-        requestAvgFailedDurationMillis: number;
-        requestAvgFinishedDurationMillis: number;
-        requestsFinishedPerMinute: number;
-        requestsFailedPerMinute: number;
-        requestTotalDurationMillis: number;
-        requestsTotal: number;
-        crawlerRuntimeMillis: number;
-    };
-    /**
-     * Initializes the key value store for persisting the statistics,
-     * displaying the current state in predefined intervals
-     */
-    startCapturing(): Promise<void>;
-    /**
-     * Stops logging and remove event listeners, then persist
-     */
-    stopCapturing(): Promise<void>;
-    private saveRetryCountForJob;
-    /**
-     * Persist internal state to the key value store
-     * @param options - Override the persistence options provided in the constructor
-     */
-    persistState(options?: PersistenceOptions): Promise<void>;
-    /**
-     * Loads the current statistic from the key value store if any
-     */
-    protected maybeLoadStatistics(): Promise<void>;
-    private teardown;
-    /**
-     * Make this class serializable when called with `JSON.stringify(statsInstance)` directly
-     * or through `keyValueStore.setValue('KEY', statsInstance)`
-     */
-    toJSON(): StatisticPersistedState;
-}
-/**
- * Configuration for the {@link Statistics} instance used by the crawler
- */
-export interface StatisticsOptions {
-    /**
-     * Interval in seconds to log the current statistics
-     * @default 60
-     */
-    logIntervalSecs?: number;
-    /**
-     * Message to log with the current statistics
-     * @default 'Statistics'
-     */
-    logMessage?: string;
-    /**
-     * Parent logger instance, the statistics will create a child logger from this.
-     * @default crawler.log
-     */
-    log?: CrawleeLogger;
-    /**
-     * Key value store instance to persist the statistics.
-     * If not provided, the default one will be used when capturing starts
-     */
-    keyValueStore?: KeyValueStore;
-    /**
-     * Control how and when to persist the statistics.
-     */
-    persistenceOptions?: PersistenceOptions;
-    /**
-     * Save HTML snapshot (and a screenshot if possible) when an error occurs.
-     * @default false
-     */
-    saveErrorSnapshots?: boolean;
-    /**
-     * A unique identifier for this statistics instance. This ID is used for persistence
-     * to the key value store, ensuring the same statistics can be loaded after script restarts.
-     *
-     * If not provided, an auto-incremented ID will be used for backward compatibility.
-     * This means statistics may not persist correctly across script restarts
-     * if crawler creation order changes.
-     */
-    id?: string;
-}
-/**
- * Format of the persisted stats
- */
-export interface StatisticPersistedState extends Omit<StatisticState, 'statsPersistedAt'> {
-    requestRetryHistogram: number[];
-    statsId: string;
-    requestAvgFailedDurationMillis: number;
-    requestAvgFinishedDurationMillis: number;
-    requestTotalDurationMillis: number;
-    requestsTotal: number;
-    crawlerLastStartTimestamp: number;
-    statsPersistedAt: string;
-}
-/**
- * Contains the statistics state
- */
-export interface StatisticState {
-    requestsFinished: number;
-    requestsFailed: number;
-    requestsRetries: number;
-    requestsFailedPerMinute: number;
-    requestsFinishedPerMinute: number;
-    requestMinDurationMillis: number;
-    requestMaxDurationMillis: number;
-    requestTotalFailedDurationMillis: number;
-    requestTotalFinishedDurationMillis: number;
-    crawlerStartedAt: Date | string | null;
-    crawlerFinishedAt: Date | string | null;
-    crawlerRuntimeMillis: number;
-    statsPersistedAt: Date | string | null;
-    errors: Record<string, unknown>;
-    retryErrors: Record<string, unknown>;
-    requestsWithStatusCode: Record<string, number>;
-}