@midscene/web 0.30.10 → 1.0.0

package/dist/types/bridge-mode/page-browser-side.d.ts

@@ -3,12 +3,13 @@ import type { ChromePageDestroyOptions } from '../web-page';
 import { type BridgeConnectTabOptions } from './common';
 import { BridgeClient } from './io-client';
 export declare class ExtensionBridgePageBrowserSide extends ChromeExtensionProxyPage {
+    serverEndpoint?: string | undefined;
     onDisconnect: () => void;
     onLogMessage: (message: string, type: 'log' | 'status') => void;
     bridgeClient: BridgeClient | null;
     private destroyOptions?;
     private newlyCreatedTabIds;
-    constructor(onDisconnect?: () => void, onLogMessage?: (message: string, type: 'log' | 'status') => void, forceSameTabNavigation?: boolean);
+    constructor(serverEndpoint?: string | undefined, onDisconnect?: () => void, onLogMessage?: (message: string, type: 'log' | 'status') => void, forceSameTabNavigation?: boolean);
     private setupBridgeClient;
     connect(): Promise<void>;
     connectNewTabWithUrl(url: string, options?: BridgeConnectTabOptions): Promise<void>;

package/dist/types/chrome-extension/page.d.ts

@@ -7,8 +7,6 @@ export default class ChromeExtensionProxyPage implements AbstractInterface {
     forceSameTabNavigation: boolean;
     private viewportSize?;
     private activeTabId;
-    private tabIdOfDebuggerAttached;
-    private attachingDebugger;
     private destroyed;
     private isMobileEmulation;
     _continueWhenFailedToAttachDebugger: boolean;
@@ -27,26 +25,41 @@ export default class ChromeExtensionProxyPage implements AbstractInterface {
         currentActiveTab: boolean;
     }[]>;
     getTabIdOrConnectToCurrentTab(): Promise<number>;
-    private attachDebugger;
+    /**
+     * Ensure debugger is attached to the current tab.
+     * Uses lazy attach pattern - only attaches when needed.
+     */
+    private ensureDebuggerAttached;
     private showMousePointer;
     private hideMousePointer;
-    private detachDebugger;
+    /**
+     * Public method to detach debugger without destroying the page instance.
+     * Useful for error recovery scenarios where we want to remove the debugger banner
+     * without completely destroying the page.
+     */
+    detachDebugger(tabId?: number): Promise<void>;
     private enableWaterFlowAnimation;
     private disableWaterFlowAnimation;
+    /**
+     * Send a command to the debugger with automatic attach and retry on detachment.
+     * Uses lazy attach pattern - will automatically attach if not already attached.
+     */
     private sendCommandToDebugger;
     private getPageContentByCDP;
     evaluateJavaScript(script: string): Promise<any>;
     beforeInvokeAction(): Promise<void>;
     private waitUntilNetworkIdle;
     getElementsInfo(): Promise<ElementInfo[]>;
-    getXpathsById(id: string): Promise<any>;
-    getXpathsByPoint(point: Point, isOrderSensitive: boolean): Promise<any>;
+    getXpathsByPoint(point: Point, isOrderSensitive?: boolean): Promise<any>;
     getElementInfoByXpath(xpath: string): Promise<any>;
     getElementsNodeTree(): Promise<ElementTreeNode<ElementInfo>>;
     getContext(): Promise<UIContext>;
     size(): Promise<Size>;
     screenshotBase64(): Promise<string>;
     url(): Promise<string>;
+    navigate(url: string): Promise<void>;
+    reload(): Promise<void>;
+    goBack(): Promise<void>;
     scrollUntilTop(startingPoint?: Point): Promise<void>;
     scrollUntilBottom(startingPoint?: Point): Promise<void>;
     scrollUntilLeft(startingPoint?: Point): Promise<void>;

package/dist/types/playwright/ai-fixture.d.ts

@@ -21,6 +21,12 @@ export declare const PlaywrightAiFixture: (options?: {
     ai: ({ page }: {
         page: OriginPlaywrightPage;
     }, use: any, testInfo: TestInfo) => Promise<void>;
+    aiAct: ({ page }: {
+        page: OriginPlaywrightPage;
+    }, use: any, testInfo: TestInfo) => Promise<void>;
+    /**
+     * @deprecated Use {@link PlaywrightAiFixture.aiAct} instead.
+     */
     aiAction: ({ page }: {
         page: OriginPlaywrightPage;
     }, use: any, testInfo: TestInfo) => Promise<void>;
@@ -78,6 +84,9 @@ export declare const PlaywrightAiFixture: (options?: {
     evaluateJavaScript: ({ page }: {
         page: OriginPlaywrightPage;
     }, use: any, testInfo: TestInfo) => Promise<void>;
+    recordToReport: ({ page }: {
+        page: OriginPlaywrightPage;
+    }, use: any, testInfo: TestInfo) => Promise<void>;
     logScreenshot: ({ page }: {
         page: OriginPlaywrightPage;
     }, use: any, testInfo: TestInfo) => Promise<void>;
@@ -90,8 +99,12 @@ export declare const PlaywrightAiFixture: (options?: {
 };
 export type PlayWrightAiFixtureType = {
     agentForPage: (page?: any, opts?: any) => Promise<PageAgent<PlaywrightWebPage>>;
-    ai: <T = any>(prompt: string) => Promise<T>;
-    aiAction: (taskPrompt: string) => ReturnType<PageAgent['aiAction']>;
+    ai: <T = any>(...args: Parameters<PageAgent['ai']>) => Promise<T>;
+    aiAct: (...args: Parameters<PageAgent['aiAct']>) => ReturnType<PageAgent['aiAct']>;
+    /**
+     * @deprecated Use {@link PlayWrightAiFixtureType.aiAct} instead.
+     */
+    aiAction: (...args: Parameters<PageAgent['aiAction']>) => ReturnType<PageAgent['aiAction']>;
     aiTap: (...args: Parameters<PageAgent['aiTap']>) => ReturnType<PageAgent['aiTap']>;
     aiRightClick: (...args: Parameters<PageAgent['aiRightClick']>) => ReturnType<PageAgent['aiRightClick']>;
     aiDoubleClick: (...args: Parameters<PageAgent['aiDoubleClick']>) => ReturnType<PageAgent['aiDoubleClick']>;
@@ -110,6 +123,7 @@ export type PlayWrightAiFixtureType = {
     runYaml: (...args: Parameters<PageAgent['runYaml']>) => ReturnType<PageAgent['runYaml']>;
     setAIActionContext: (...args: Parameters<PageAgent['setAIActionContext']>) => ReturnType<PageAgent['setAIActionContext']>;
     evaluateJavaScript: (...args: Parameters<PageAgent['evaluateJavaScript']>) => ReturnType<PageAgent['evaluateJavaScript']>;
+    recordToReport: (...args: Parameters<PageAgent['recordToReport']>) => ReturnType<PageAgent['recordToReport']>;
     logScreenshot: (...args: Parameters<PageAgent['logScreenshot']>) => ReturnType<PageAgent['logScreenshot']>;
     freezePageContext: (...args: Parameters<PageAgent['freezePageContext']>) => ReturnType<PageAgent['freezePageContext']>;
     unfreezePageContext: (...args: Parameters<PageAgent['unfreezePageContext']>) => ReturnType<PageAgent['unfreezePageContext']>;

package/dist/types/playwright/index.d.ts

@@ -5,6 +5,7 @@ export type { PlayWrightAiFixtureType } from './ai-fixture';
 export { PlaywrightAiFixture } from './ai-fixture';
 export { overrideAIConfig } from '@midscene/shared/env';
 export { WebPage as PlaywrightWebPage } from './page';
+export type { WebPageAgentOpt } from '../web-element';
 import type { WebPageAgentOpt } from '../web-element';
 export declare class PlaywrightAgent extends PageAgent<PlaywrightWebPage> {
     constructor(page: PlaywrightPage, opts?: WebPageAgentOpt);

package/dist/types/playwright/reporter/index.d.ts

@@ -6,6 +6,7 @@ declare class MidsceneReporter implements Reporter {
     private mergedFilename?;
     private testTitleToFilename;
     mode?: 'merged' | 'separate';
+    private tempFiles;
     constructor(options?: MidsceneReporterOptions);
     private static getMode;
     private getSeparatedFilename;
@@ -14,5 +15,6 @@ declare class MidsceneReporter implements Reporter {
     onBegin(config: FullConfig, suite: Suite): Promise<void>;
     onTestBegin(_test: TestCase, _result: TestResult): void;
     onTestEnd(test: TestCase, result: TestResult): void;
+    onEnd(): void;
 }
 export default MidsceneReporter;

package/dist/types/puppeteer/agent-launcher.d.ts

@@ -1,11 +1,12 @@
 import { PuppeteerAgent } from './index';
-import type { Cache, MidsceneYamlScriptWebEnv } from '@midscene/core';
+import type { AgentOpt, MidsceneYamlScriptWebEnv } from '@midscene/core';
 import { type Browser } from 'puppeteer';
 export declare const defaultUA = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36";
 export declare const defaultViewportWidth = 1440;
 export declare const defaultViewportHeight = 768;
 export declare const defaultViewportScale: number;
 export declare const defaultWaitForNetworkIdleTimeout = 2000;
+export declare function resolveAiActionContext(target: MidsceneYamlScriptWebEnv, preference?: Partial<Pick<AgentOpt, 'aiActionContext' | 'aiActContext'>>): AgentOpt['aiActionContext'] | undefined;
 interface FreeFn {
     name: string;
     fn: () => void;
@@ -20,9 +21,7 @@ export declare function launchPuppeteerPage(target: MidsceneYamlScriptWebEnv, pr
 export declare function puppeteerAgentForTarget(target: MidsceneYamlScriptWebEnv, preference?: {
     headed?: boolean;
     keepWindow?: boolean;
-    testId?: string;
-    cache?: Cache;
-}, browser?: Browser): Promise<{
+} & Partial<Pick<AgentOpt, 'testId' | 'groupName' | 'groupDescription' | 'generateReport' | 'autoPrintReportMsg' | 'reportFileName' | 'replanningCycleLimit' | 'cache' | 'aiActionContext'>>, browser?: Browser): Promise<{
     agent: PuppeteerAgent;
     freeFn: FreeFn[];
 }>;

package/dist/types/puppeteer/base-page.d.ts

@@ -1,6 +1,7 @@
 import { type WebPageAgentOpt } from '../web-element';
 import type { DeviceAction, ElementCacheFeature, ElementTreeNode, Point, Rect, Size, UIContext } from '@midscene/core';
 import type { AbstractInterface } from '@midscene/core/device';
+import type { IModelConfig } from '@midscene/shared/env';
 import type { ElementInfo } from '@midscene/shared/extractor';
 import { type DebugFunction } from '@midscene/shared/logger';
 import type { Page as PlaywrightPage } from 'playwright';
@@ -15,6 +16,7 @@ export declare class Page<AgentType extends 'puppeteer' | 'playwright', Interfac
     private onBeforeInvokeAction?;
     private onAfterInvokeAction?;
     private customActions?;
+    private enableTouchEventsInActionSpace;
     interfaceType: AgentType;
     actionSpace(): DeviceAction[];
     private evaluate;
@@ -23,11 +25,11 @@ export declare class Page<AgentType extends 'puppeteer' | 'playwright', Interfac
     waitForNavigation(): Promise<void>;
     waitForNetworkIdle(): Promise<void>;
     getElementsInfo(): Promise<ElementInfo[]>;
-    getXpathsById(id: string): Promise<any>;
-    getXpathsByPoint(point: Point, isOrderSensitive: boolean): Promise<any>;
-    getElementInfoByXpath(xpath: string): Promise<any>;
-    cacheFeatureForRect(rect: Rect, opt?: {
-        _orderSensitive: boolean;
+    private getXpathsByPoint;
+    private getElementInfoByXpath;
+    cacheFeatureForRect(rect: Rect, options?: {
+        targetDescription?: string;
+        modelConfig?: IModelConfig;
     }): Promise<ElementCacheFeature>;
     rectMatchesCacheFeature(feature: ElementCacheFeature): Promise<Rect>;
     getElementsNodeTree(): Promise<ElementTreeNode<ElementInfo>>;
@@ -74,6 +76,8 @@ export declare class Page<AgentType extends 'puppeteer' | 'playwright', Interfac
     scrollLeft(distance?: number, startingPoint?: Point): Promise<void>;
     scrollRight(distance?: number, startingPoint?: Point): Promise<void>;
     navigate(url: string): Promise<void>;
+    reload(): Promise<void>;
+    goBack(): Promise<void>;
     beforeInvokeAction(name: string, param: any): Promise<void>;
     afterInvokeAction(name: string, param: any): Promise<void>;
     destroy(): Promise<void>;
@@ -88,3 +92,12 @@ export declare class Page<AgentType extends 'puppeteer' | 'playwright', Interfac
     longPress(x: number, y: number, duration?: number): Promise<void>;
 }
 export declare function forceClosePopup(page: PuppeteerPage | PlaywrightPage, debugProfile: DebugFunction): void;
+/**
+ * Force Chrome to render select elements using base-select appearance instead of OS-native rendering.
+ * This makes select elements visible in screenshots captured by Playwright/Puppeteer.
+ *
+ * Reference: https://developer.chrome.com/blog/a-customizable-select
+ *
+ * Adds a style tag with CSS rules to make all select elements use base-select appearance.
+ */
+export declare function forceChromeSelectRendering(page: PuppeteerPage | PlaywrightPage): void;

package/dist/types/puppeteer/index.d.ts

@@ -3,6 +3,7 @@ import { Agent as PageAgent } from '@midscene/core/agent';
 import type { Page as PuppeteerPage } from 'puppeteer';
 import { PuppeteerWebPage } from './page';
 export { PuppeteerWebPage } from './page';
+export type { WebPageAgentOpt } from '../web-element';
 export declare class PuppeteerAgent extends PageAgent<PuppeteerWebPage> {
     constructor(page: PuppeteerPage, opts?: WebPageAgentOpt);
 }

package/dist/types/static/static-page.d.ts

@@ -16,7 +16,6 @@ export default class StaticPage implements AbstractInterface {
     evaluateJavaScript<T = unknown>(script: string): Promise<T>;
     getElementsInfo(): Promise<never>;
     getElementsNodeTree(): Promise<never>;
-    getXpathsById(id: string): Promise<never>;
     getXpathsByPoint(point: Point): Promise<never>;
     getElementInfoByXpath(xpath: string): Promise<never>;
     size(): Promise<{

package/dist/types/utils.d.ts

@@ -0,0 +1,6 @@
+/**
+ * Get a require function that won't be processed by webpack.
+ * Returns __non_webpack_require__ if available (in webpack environment),
+ * otherwise falls back to the standard require.
+ */
+export declare function getWebpackRequire(): typeof require;

package/dist/types/web-element.d.ts

@@ -11,6 +11,16 @@ export type WebPageOpt = {
     waitForNavigationTimeout?: number;
     waitForNetworkIdleTimeout?: number;
     forceSameTabNavigation?: boolean;
+    enableTouchEventsInActionSpace?: boolean;
+    /**
+     * Force Chrome to render select elements using base-select appearance instead of OS-native rendering.
+     * This makes select elements visible in screenshots captured by Playwright/Puppeteer.
+     *
+     * Reference: https://developer.chrome.com/blog/a-customizable-select
+     *
+     * When enabled, adds a style tag with `select { appearance: base-select !important; }` to the page.
+     */
+    forceChromeSelectRendering?: boolean;
     beforeInvokeAction?: () => Promise<void>;
     afterInvokeAction?: () => Promise<void>;
     customActions?: DeviceAction<any>[];

package/dist/types/web-page.d.ts

@@ -36,6 +36,9 @@ export interface ChromePageDestroyOptions {
     closeTab?: boolean;
 }
 export declare abstract class AbstractWebPage extends AbstractInterface {
+    navigate?(url: string): Promise<void>;
+    reload?(): Promise<void>;
+    goBack?(): Promise<void>;
     get mouse(): MouseAction;
     get keyboard(): KeyboardAction;
     clearInput(element: ElementInfo): Promise<void>;
@@ -56,4 +59,4 @@ export declare abstract class AbstractWebPage extends AbstractInterface {
         y: number;
     }, duration?: number): Promise<void>;
 }
-export declare const commonWebActionsForWebPage: <T extends AbstractWebPage>(page: T) => DeviceAction<any>[];
+export declare const commonWebActionsForWebPage: <T extends AbstractWebPage>(page: T, includeTouchEvents?: boolean) => DeviceAction<any>[];

package/package.json

@@ -8,7 +8,7 @@
     "Browser use",
     "Android use"
   ],
-  "version": "0.30.10",
+  "version": "1.0.0",
   "repository": "https://github.com/web-infra-dev/midscene",
   "homepage": "https://midscenejs.com/",
   "main": "./dist/lib/index.js",
@@ -86,15 +86,6 @@
       "import": "./dist/es/static/index.mjs"
     }
   },
-  "watch": {
-    "build": {
-      "patterns": [
-        "src"
-      ],
-      "extensions": "tsx,less,scss,css,js,jsx,ts",
-      "quiet": false
-    }
-  },
   "bin": {
     "midscene-playground": "./bin/midscene-playground"
   },
@@ -112,22 +103,24 @@
     "http-server": "14.1.1",
     "socket.io": "^4.8.1",
     "socket.io-client": "4.8.1",
-    "@midscene/playground": "0.30.10",
-    "@midscene/core": "0.30.10",
-    "@midscene/shared": "0.30.10"
+    "@midscene/core": "1.0.0",
+    "@midscene/playground": "1.0.0",
+    "@midscene/shared": "1.0.0"
   },
   "devDependencies": {
-    "@types/js-yaml": "4.0.9",
-    "devtools-protocol": "0.0.1380148",
     "@playwright/test": "^1.44.1",
-    "@rslib/core": "^0.11.2",
+    "@rslib/core": "^0.18.3",
     "@types/chrome": "0.0.279",
     "@types/cors": "^2.8.17",
     "@types/http-server": "^0.12.4",
+    "@types/js-yaml": "4.0.9",
     "@types/node": "^18.0.0",
-    "playwright": "1.44.1",
-    "puppeteer": "24.2.0",
+    "@types/semver": "7.7.0",
+    "devtools-protocol": "0.0.1380148",
     "js-yaml": "4.1.0",
+    "langsmith": "0.3.74",
+    "playwright": "1.44.1",
+    "puppeteer": "24.6.0",
     "tsx": "^4.19.2",
     "typescript": "^5.8.3",
     "vitest": "3.0.5"
@@ -154,10 +147,10 @@
   },
   "license": "MIT",
   "scripts": {
-    "dev": "npm run build && npx npm-watch",
+    "dev": "npm run build:watch",
     "dev:server": "npm run build && ./bin/midscene-playground",
     "build": "rslib build",
-    "build:watch": "rslib build --watch",
+    "build:watch": "rslib build --watch --no-clean",
     "playground": "DEBUG=midscene:* tsx demo/playground.ts",
     "test": "vitest --run",
     "test:u": "vitest --run -u",