@crawlee/core 4.0.0-beta.115 → 4.0.0-beta.116

package/autoscaling/cpu_load_signal.js

@@ -1,3 +1,4 @@
+import { EventType } from '../events/event_manager.js';
 import { serviceLocator } from '../service_locator.js';
 import { SnapshotStore } from './load_signal.js';
 /**
@@ -24,10 +25,10 @@ export class CpuLoadSignal {
         // Resolved here rather than in the constructor, so an instance built ahead of time (to be wrapped, or shared
         // between systems) cannot capture whichever event manager happened to be registered at that moment.
         this.#events = serviceLocator.getEventManager();
-        this.#events.on("systemInfo" /* EventType.SYSTEM_INFO */, this.handle);
+        this.#events.on(EventType.SYSTEM_INFO, this.handle);
     }
     async stop() {
-        this.#events?.off("systemInfo" /* EventType.SYSTEM_INFO */, this.handle);
+        this.#events?.off(EventType.SYSTEM_INFO, this.handle);
         this.#events = undefined;
     }
     getSample(sampleDurationMillis) {

package/autoscaling/memory_load_signal.js

@@ -1,3 +1,4 @@
+import { EventType } from '../events/event_manager.js';
 import { serviceLocator } from '../service_locator.js';
 import { getMemoryInfo } from '../system-info/memory-info.js';
 import { isContainerized } from '../system-info/runtime.js';
@@ -54,10 +55,10 @@ export class MemoryLoadSignal {
             // Fallback memory measurement in case memTotalBytes is missing from SystemInfo.
             this.#maxMemoryBytes = await this.getTotalMemoryBytes();
         }
-        this.#events.on("systemInfo" /* EventType.SYSTEM_INFO */, this.handle);
+        this.#events.on(EventType.SYSTEM_INFO, this.handle);
     }
     async stop() {
-        this.#events?.off("systemInfo" /* EventType.SYSTEM_INFO */, this.handle);
+        this.#events?.off(EventType.SYSTEM_INFO, this.handle);
         this.#events = undefined;
     }
     getSample(sampleDurationMillis) {

package/crawlers/statistics.js

@@ -1,4 +1,5 @@
 import ow from 'ow';
+import { EventType } from '../events/event_manager.js';
 import { serviceLocator } from '../service_locator.js';
 import { KeyValueStore } from '../storages/key_value_store.js';
 import { ErrorTracker } from './error_tracker.js';
@@ -238,7 +239,7 @@ export class Statistics {
         }
         if (this.#persistenceOptions.enable) {
             await this.maybeLoadStatistics();
-            this.events.on("persistState" /* EventType.PERSIST_STATE */, this.#listener);
+            this.events.on(EventType.PERSIST_STATE, this.#listener);
         }
         this.#logInterval = setInterval(() => {
             this.log.info(this.#logMessage, {
@@ -320,7 +321,7 @@ export class Statistics {
         // this can be called before a call to startCapturing happens (or in a 'finally' block)
         // Only unsubscribe if event manager was already resolved — avoid eagerly resolving it
         // (e.g. during the constructor's reset() call, which would capture the wrong context)
-        this.#events?.off("persistState" /* EventType.PERSIST_STATE */, this.#listener);
+        this.#events?.off(EventType.PERSIST_STATE, this.#listener);
         if (this.#logInterval) {
             clearInterval(this.#logInterval);
             this.#logInterval = null;

package/events/event_manager.d.ts

@@ -4,7 +4,7 @@ export interface EventManagerOptions {
     /** Interval between emitted `persistState` events in milliseconds. */
     persistStateIntervalMillis: number;
 }
-export declare const enum EventType {
+export declare enum EventType {
     PERSIST_STATE = "persistState",
     SYSTEM_INFO = "systemInfo",
     MIGRATING = "migrating",

package/events/event_manager.js

@@ -29,7 +29,7 @@ export class EventManager {
             return;
         }
         this.intervals.persistState = betterSetInterval((intervalCallback) => {
-            this.emit("persistState" /* EventType.PERSIST_STATE */, { isMigrating: false });
+            this.emit(EventType.PERSIST_STATE, { isMigrating: false });
             intervalCallback();
         }, this.#persistStateIntervalMillis);
         this.initialized = true;
@@ -45,7 +45,7 @@ export class EventManager {
         betterClearInterval(this.intervals.persistState);
         this.initialized = false;
         // Emit final PERSIST_STATE event
-        this.emit("persistState" /* EventType.PERSIST_STATE */, { isMigrating: false });
+        this.emit(EventType.PERSIST_STATE, { isMigrating: false });
         // Wait for PERSIST_STATE to process
         await this.waitForAllListenersToComplete();
     }

package/events/local_event_manager.js

@@ -1,6 +1,6 @@
 import { betterClearInterval, betterSetInterval } from '@apify/utilities';
 import { serviceLocator } from '../service_locator.js';
-import { EventManager } from './event_manager.js';
+import { EventManager, EventType } from './event_manager.js';
 export class LocalEventManager extends EventManager {
     #systemInfoIntervalMillis;
     constructor(options) {
@@ -47,7 +47,7 @@ export class LocalEventManager extends EventManager {
         const info = await this.createSystemInfo({
             maxUsedCpuRatio: serviceLocator.getConfiguration().maxUsedCpuRatio,
         });
-        this.events.emit("systemInfo" /* EventType.SYSTEM_INFO */, info);
+        this.events.emit(EventType.SYSTEM_INFO, info);
         intervalCallback();
     }
     /**

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@crawlee/core",
-    "version": "4.0.0-beta.115",
+    "version": "4.0.0-beta.116",
     "description": "The scalable web crawling and scraping library for JavaScript/Node.js. Enables development of data extraction and web automation jobs (not only) with headless Chrome and Puppeteer.",
     "engines": {
         "node": ">=22.0.0"
@@ -52,9 +52,9 @@
         "@apify/log": "^2.5.18",
         "@apify/timeout": "^0.4.4",
         "@apify/utilities": "^2.15.5",
-        "@crawlee/fs-storage": "4.0.0-beta.115",
-        "@crawlee/types": "4.0.0-beta.115",
-        "@crawlee/utils": "4.0.0-beta.115",
+        "@crawlee/fs-storage": "4.0.0-beta.116",
+        "@crawlee/types": "4.0.0-beta.116",
+        "@crawlee/utils": "4.0.0-beta.116",
         "@sapphire/async-queue": "^1.5.5",
         "@sapphire/shapeshift": "^4.0.0",
         "@vladfrangu/async_event_emitter": "^2.4.6",
@@ -78,5 +78,5 @@
             }
         }
     },
-    "gitHead": "9bb6f54b10a88a709f0461486601a98365a90403"
+    "gitHead": "2bdc17928f47674f08fff7a83de861d881b7947a"
 }

package/recoverable_state.js

@@ -1,4 +1,4 @@
-import { KeyValueStore, serviceLocator } from '@crawlee/core';
+import { EventType, KeyValueStore, serviceLocator } from '@crawlee/core';
 /**
  * A class for managing persistent recoverable state using a plain JavaScript object.
  *
@@ -65,7 +65,7 @@ export class RecoverableState {
         await this.loadSavedState();
         // Register for persist state events
         const eventManager = serviceLocator.getEventManager();
-        eventManager.on("persistState" /* EventType.PERSIST_STATE */, this.persistState);
+        eventManager.on(EventType.PERSIST_STATE, this.persistState);
         return this.currentValue;
     }
     /**
@@ -79,7 +79,7 @@ export class RecoverableState {
             return;
         }
         const eventManager = serviceLocator.getEventManager();
-        eventManager.off("persistState" /* EventType.PERSIST_STATE */, this.persistState);
+        eventManager.off(EventType.PERSIST_STATE, this.persistState);
         await this.persistState();
     }
     /**

package/session_pool/session_pool.js

@@ -1,5 +1,6 @@
 import { AsyncQueue } from '@sapphire/async-queue';
 import ow from 'ow';
+import { EventType } from '../events/event_manager.js';
 import { serviceLocator } from '../service_locator.js';
 import { KeyValueStore } from '../storages/key_value_store.js';
 import { MAX_POOL_SIZE, PERSIST_STATE_KEY } from './consts.js';
@@ -150,7 +151,7 @@ export class SessionPool {
         // in case of migration happened and SessionPool state should be restored from the keyValueStore.
         await this.maybeLoadSessionPool();
         this.#listener = this.persistState.bind(this);
-        this.#events.on("persistState" /* EventType.PERSIST_STATE */, this.#listener);
+        this.#events.on(EventType.PERSIST_STATE, this.#listener);
     }
     /**
      * Adds a new session to the session pool. The pool automatically creates sessions up to the maximum size of the pool,
@@ -265,7 +266,7 @@ export class SessionPool {
             return;
         await this.ensureInitialized();
         if (this.#listener) {
-            this.#events.off("persistState" /* EventType.PERSIST_STATE */, this.#listener);
+            this.#events.off(EventType.PERSIST_STATE, this.#listener);
         }
         await this.persistState();
     }

package/storages/request_list.js

@@ -1,5 +1,6 @@
 import { downloadListOfUrls } from '@crawlee/utils';
 import ow, { ArgumentError } from 'ow';
+import { EventType } from '../events/event_manager.js';
 import { Request } from '../request.js';
 import { createDeserialize, serializeArray } from '../serialization.js';
 import { serviceLocator } from '../service_locator.js';
@@ -177,7 +178,7 @@ export class RequestList {
         if (this.#persistRequestsKey && !this.areRequestsPersisted)
             await this.persistRequests();
         if (this.#persistStateKey) {
-            serviceLocator.getEventManager().on("persistState" /* EventType.PERSIST_STATE */, this.persistState);
+            serviceLocator.getEventManager().on(EventType.PERSIST_STATE, this.persistState);
         }
         return this;
     }
@@ -268,7 +269,7 @@ export class RequestList {
      * leaking the listener (and the requests it retains) on the shared event manager.
      */
     async teardown() {
-        serviceLocator.getEventManager().off("persistState" /* EventType.PERSIST_STATE */, this.persistState);
+        serviceLocator.getEventManager().off(EventType.PERSIST_STATE, this.persistState);
         if (this.#persistStateKey) {
             await this.persistState();
         }

package/storages/request_queue.js

@@ -6,6 +6,7 @@ import { LruCache } from '@apify/datastructures';
 import { tryCancel } from '@apify/timeout';
 import { Configuration } from '../configuration.js';
 import { getObjectType } from '../debug.js';
+import { EventType } from '../events/event_manager.js';
 import { Request } from '../request.js';
 import { serviceLocator } from '../service_locator.js';
 import { activeStorageTransaction, rejectOperationInTransaction } from './transaction.js';
@@ -101,7 +102,7 @@ export class RequestQueue {
         this.requestCache = new LruCache({ maxLength: MAX_CACHED_REQUESTS });
         this.#requestSeenCache = new RequestDeduplicationCache();
         this.log = serviceLocator.getLogger().child({ prefix: `RequestQueue(${this.id}, ${this.name ?? 'no-name'})` });
-        this.#events.on("migrating" /* EventType.MIGRATING */, async () => {
+        this.#events.on(EventType.MIGRATING, async () => {
             this.#queuePausedForMigration = true;
         });
     }

package/storages/sitemap_request_loader.js

@@ -3,6 +3,7 @@ import { parseSitemap } from '@crawlee/utils';
 import { minimatch } from 'minimatch';
 import ow from 'ow';
 import { constructUrlPatternObjects } from '../enqueue_links/shared.js';
+import { EventType } from '../events/event_manager.js';
 import { Request } from '../request.js';
 import { serviceLocator } from '../service_locator.js';
 import { KeyValueStore } from './key_value_store.js';
@@ -251,7 +252,7 @@ export class SitemapRequestLoader {
             parseSitemapOptions: { logger: serviceLocator.getLogger(), ...options.parseSitemapOptions, httpClient },
         });
         if (requestList.#persistenceOptions.enable) {
-            requestList.#events.on("persistState" /* EventType.PERSIST_STATE */, requestList.persistState);
+            requestList.#events.on(EventType.PERSIST_STATE, requestList.persistState);
         }
         options?.signal?.addEventListener('abort', () => {
             requestList.#abortLoading = true;
@@ -402,7 +403,7 @@ export class SitemapRequestLoader {
     async teardown() {
         this.#closed = true;
         this.#abortLoading = true;
-        this.#events.off("persistState" /* EventType.PERSIST_STATE */, this.persistState);
+        this.#events.off(EventType.PERSIST_STATE, this.persistState);
         await this.persistState();
         this.#urlQueueStream.emit('readdata'); // unblocks the potentially waiting `pushNextUrl` call
     }