@crawlee/stagehand 3.17.1-beta.53 → 3.17.1-beta.55

package/index.mjs

@@ -50,6 +50,7 @@ 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;
@@ -81,6 +82,7 @@ export const createEventLoopLoadSignal = mod.createEventLoopLoadSignal;
 export const createRequestOptions = mod.createRequestOptions;
 export const createRequests = mod.createRequests;
 export const createStagehandRouter = mod.createStagehandRouter;
+export const defaultRoute = mod.defaultRoute;
 export const deserializeArray = mod.deserializeArray;
 export const enqueueLinks = mod.enqueueLinks;
 export const evaluateLoadSignalSample = mod.evaluateLoadSignalSample;
@@ -103,5 +105,6 @@ 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/stagehand-crawler.d.ts

@@ -1,5 +1,5 @@
 import type { Action, ActOptions, ActResult, AgentConfig, ExtractOptions, LLMClient, ModelConfiguration, NonStreamingAgentInstance, ObserveOptions, Stagehand, StreamingAgentInstance } from '@browserbasehq/stagehand';
-import type { BrowserCrawlerOptions, BrowserCrawlingContext, BrowserHook, BrowserRequestHandler, GetUserDataFromRequest, LoadedContext, RouterHandler, RouterRoutes } from '@crawlee/browser';
+import type { BrowserCrawlerOptions, BrowserCrawlingContext, BrowserHook, BrowserRequestHandler, GetUserDataFromRequest, LoadedContext, RouterHandler, RouterRoutes, RouteSchemas, RoutesFromSchemas } from '@crawlee/browser';
 import { BrowserCrawler, Configuration } from '@crawlee/browser';
 import type { Dictionary } from '@crawlee/types';
 // @ts-ignore optional peer dependency or compatibility with es2022
@@ -408,6 +408,9 @@ export declare class StagehandCrawler extends BrowserCrawler<{
 // @ts-ignore optional peer dependency or compatibility with es2022
         statisticsOptions: import("ow").ObjectPredicate<object> & import("ow").BasePredicate<object | undefined>;
     };
+    /** Set while `_runRequestHandler` has swapped `userProvidedRequestHandler` for its page-enhancing wrapper. */
+    private unwrappedRequestHandler?;
+    protected get userRequestHandler(): BrowserRequestHandler<StagehandCrawlingContext>;
     /**
      * Creates a new instance of StagehandCrawler.
      *
@@ -460,3 +463,4 @@ export declare class StagehandCrawler extends BrowserCrawler<{
  */
 export declare function createStagehandRouter<Context extends StagehandCrawlingContext = StagehandCrawlingContext, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<Context['request']>>>(routes?: RouterRoutes<Context, Routes>): RouterHandler<Context, Routes>;
 export declare function createStagehandRouter<Context extends StagehandCrawlingContext = StagehandCrawlingContext, UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(routes?: RouterRoutes<Context, Record<string, UserData>>): RouterHandler<Context, Record<string, UserData>>;
+export declare function createStagehandRouter<Context extends StagehandCrawlingContext = StagehandCrawlingContext, const Schemas extends RouteSchemas = RouteSchemas>(schemas: Schemas): RouterHandler<Context, RoutesFromSchemas<Schemas>>;

package/internals/stagehand-crawler.js

@@ -57,6 +57,9 @@ const stagehand_utils_1 = require("./utils/stagehand-utils");
  * ```
  */
 class StagehandCrawler extends browser_1.BrowserCrawler {
+    get userRequestHandler() {
+        return this.unwrappedRequestHandler ?? super.userRequestHandler;
+    }
     /**
      * Creates a new instance of StagehandCrawler.
      *
@@ -88,6 +91,13 @@ class StagehandCrawler extends browser_1.BrowserCrawler {
             writable: true,
             value: config
         });
+        /** Set while `_runRequestHandler` has swapped `userProvidedRequestHandler` for its page-enhancing wrapper. */
+        Object.defineProperty(this, "unwrappedRequestHandler", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
     }
     /**
      * Overrides the request handler to enhance the page with Stagehand AI methods.
@@ -105,6 +115,9 @@ class StagehandCrawler extends browser_1.BrowserCrawler {
     async _runRequestHandler(crawlingContext) {
         // Store the original handler (could be this.requestHandler or this.router)
         const originalHandler = this.userProvidedRequestHandler;
+        // Keep `userRequestHandler` pointing at the unwrapped handler: the swap below is live while the user's
+        // handler runs, which is exactly when `context.addRequests`/`enqueueLinks` resolve router metadata.
+        this.unwrappedRequestHandler = originalHandler;
         // Replace with a wrapper that enhances the page before calling the user's handler
         this.userProvidedRequestHandler = async (ctx) => {
             // Get Stagehand instance from controller
@@ -122,6 +135,7 @@ class StagehandCrawler extends browser_1.BrowserCrawler {
         finally {
             // Restore original handler
             this.userProvidedRequestHandler = originalHandler;
+            this.unwrappedRequestHandler = undefined;
         }
     }
     /**
@@ -144,6 +158,6 @@ Object.defineProperty(StagehandCrawler, "optionsShape", {
         browserPoolOptions: ow_1.default.optional.object,
     }
 });
-function createStagehandRouter(routes) {
-    return browser_1.Router.create(routes);
+function createStagehandRouter(routesOrSchemas) {
+    return browser_1.Router.create(routesOrSchemas);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@crawlee/stagehand",
-    "version": "3.17.1-beta.53",
+    "version": "3.17.1-beta.55",
     "description": "AI-powered web crawling with Stagehand integration for Crawlee - enables natural language browser automation with act(), extract(), and observe() methods.",
     "engines": {
         "node": ">=16.0.0"
@@ -58,11 +58,11 @@
     "dependencies": {
         "@apify/log": "^2.4.0",
         "@apify/timeout": "^0.3.1",
-        "@crawlee/browser": "3.17.1-beta.53",
-        "@crawlee/browser-pool": "3.17.1-beta.53",
-        "@crawlee/core": "3.17.1-beta.53",
-        "@crawlee/types": "3.17.1-beta.53",
-        "@crawlee/utils": "3.17.1-beta.53",
+        "@crawlee/browser": "3.17.1-beta.55",
+        "@crawlee/browser-pool": "3.17.1-beta.55",
+        "@crawlee/core": "3.17.1-beta.55",
+        "@crawlee/types": "3.17.1-beta.55",
+        "@crawlee/utils": "3.17.1-beta.55",
         "ow": "^0.28.1",
         "tslib": "^2.4.0"
     },
@@ -94,5 +94,5 @@
             }
         }
     },
-    "gitHead": "59a0d19d96e7402cd8d0edd8b90ad4f5abc8f7db"
+    "gitHead": "93d29e305f01ca6ca9bb79264a5ca553b19766ef"
 }