@crawlee/jsdom 3.17.1-beta.9 → 3.18.0

package/index.mjs

@@ -53,6 +53,7 @@ export const RequestQueue = mod.RequestQueue;
 export const RequestQueueV1 = mod.RequestQueueV1;
 export const RequestQueueV2 = mod.RequestQueueV2;
 export const RequestState = mod.RequestState;
+export const RequestValidationError = mod.RequestValidationError;
 export const RetryRequestError = mod.RetryRequestError;
 export const Router = mod.Router;
 export const STATE_PERSISTENCE_KEY = mod.STATE_PERSISTENCE_KEY;
@@ -84,6 +85,7 @@ export const createHttpRouter = mod.createHttpRouter;
 export const createJSDOMRouter = mod.createJSDOMRouter;
 export const createRequestOptions = mod.createRequestOptions;
 export const createRequests = mod.createRequests;
+export const defaultRoute = mod.defaultRoute;
 export const deserializeArray = mod.deserializeArray;
 export const domCrawlerEnqueueLinks = mod.domCrawlerEnqueueLinks;
 export const enqueueLinks = mod.enqueueLinks;
@@ -105,5 +107,6 @@ export const tryAbsoluteURL = mod.tryAbsoluteURL;
 export const updateEnqueueLinksPatternCache = mod.updateEnqueueLinksPatternCache;
 export const useState = mod.useState;
 export const validateGlobPattern = mod.validateGlobPattern;
+export const validateUserData = mod.validateUserData;
 export const validators = mod.validators;
 export const withCheckedStorageAccess = mod.withCheckedStorageAccess;

package/internals/jsdom-crawler.d.ts

@@ -1,5 +1,5 @@
 import type { IncomingMessage } from 'node:http';
-import type { BasicCrawlingContext, Configuration, EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler, RequestProvider, RouterRoutes, SkippedRequestCallback } from '@crawlee/http';
+import type { BasicCrawlingContext, Configuration, EnqueueLinksOptions, ErrorHandler, GetUserDataFromRequest, HttpCrawlerOptions, InternalHttpCrawlingContext, InternalHttpHook, RequestHandler, RequestProvider, RouterHandler, RouterRoutes, RouteSchemas, RoutesFromSchemas, SkippedRequestCallback } from '@crawlee/http';
 import { HttpCrawler } from '@crawlee/http';
 import type { Dictionary } from '@crawlee/types';
 import { type CheerioRoot, type RobotsTxtFile } from '@crawlee/utils';
@@ -218,6 +218,7 @@ export declare function domCrawlerEnqueueLinks(options: EnqueueLinksInternalOpti
  * await crawler.run();
  * ```
  */
-// @ts-ignore optional peer dependency or compatibility with es2022
-export declare function createJSDOMRouter<Context extends JSDOMCrawlingContext = JSDOMCrawlingContext, UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(routes?: RouterRoutes<Context, UserData>): import("@crawlee/http").RouterHandler<Context>;
+export declare function createJSDOMRouter<Context extends JSDOMCrawlingContext = JSDOMCrawlingContext, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<Context['request']>>>(routes?: RouterRoutes<Context, Routes>): RouterHandler<Context, Routes>;
+export declare function createJSDOMRouter<Context extends JSDOMCrawlingContext = JSDOMCrawlingContext, UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(routes?: RouterRoutes<Context, Record<string, UserData>>): RouterHandler<Context, Record<string, UserData>>;
+export declare function createJSDOMRouter<Context extends JSDOMCrawlingContext = JSDOMCrawlingContext, const Schemas extends RouteSchemas = RouteSchemas>(schemas: Schemas): RouterHandler<Context, RoutesFromSchemas<Schemas>>;
 export {};

package/internals/jsdom-crawler.js

@@ -267,8 +267,8 @@ async function domCrawlerEnqueueLinks(options) {
     if (containsEnqueueLinks(options)) {
         return options.enqueueLinks({
             urls,
-            baseUrl,
             ...enqueueLinksOptions,
+            baseUrl,
         });
     }
     return (0, http_1.enqueueLinks)({
@@ -276,8 +276,8 @@ async function domCrawlerEnqueueLinks(options) {
         robotsTxtFile: options.robotsTxtFile,
         onSkippedRequest: options.onSkippedRequest,
         urls,
-        baseUrl,
         ...enqueueLinksOptions,
+        baseUrl,
     });
 }
 /**
@@ -296,30 +296,6 @@ function extractUrlsFromWindow(window, selector, baseUrl) {
     })
         .filter((href) => href !== undefined && href !== '');
 }
-/**
- * Creates new {@link Router} instance that works based on request labels.
- * This instance can then serve as a `requestHandler` of your {@link JSDOMCrawler}.
- * Defaults to the {@link JSDOMCrawlingContext}.
- *
- * > Serves as a shortcut for using `Router.create<JSDOMCrawlingContext>()`.
- *
- * ```ts
- * import { JSDOMCrawler, createJSDOMRouter } from 'crawlee';
- *
- * const router = createJSDOMRouter();
- * router.addHandler('label-a', async (ctx) => {
- *    ctx.log.info('...');
- * });
- * router.addDefaultHandler(async (ctx) => {
- *    ctx.log.info('...');
- * });
- *
- * const crawler = new JSDOMCrawler({
- *     requestHandler: router,
- * });
- * await crawler.run();
- * ```
- */
-function createJSDOMRouter(routes) {
-    return http_1.Router.create(routes);
+function createJSDOMRouter(routesOrSchemas) {
+    return http_1.Router.create(routesOrSchemas);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@crawlee/jsdom",
-    "version": "3.17.1-beta.9",
+    "version": "3.18.0",
     "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": ">=16.0.0"
@@ -53,11 +53,11 @@
         "access": "public"
     },
     "dependencies": {
-        "@apify/timeout": "^0.3.0",
+        "@apify/timeout": "^0.4.0",
         "@apify/utilities": "^2.7.10",
-        "@crawlee/http": "3.17.1-beta.9",
-        "@crawlee/types": "3.17.1-beta.9",
-        "@crawlee/utils": "3.17.1-beta.9",
+        "@crawlee/http": "3.18.0",
+        "@crawlee/types": "3.18.0",
+        "@crawlee/utils": "3.18.0",
         "@types/jsdom": "^21.0.0",
         "cheerio": "1.0.0-rc.12",
         "jsdom": "^26.0.0",
@@ -71,5 +71,5 @@
             }
         }
     },
-    "gitHead": "c547ce4489d8617620a51516cc885187503f3136"
+    "gitHead": "49c115e1ce3b3fbf2bef61f16bffeadfdcf5bd19"
 }