@nahisaho/katashiro-collector 0.2.2 → 0.2.4

package/dist/browser/BrowserAutomation.js

@@ -0,0 +1,463 @@
+/**
+ * BrowserAutomation - ブラウザ自動化メインクラス
+ *
+ * Puppeteerベースのブラウザ自動化機能を提供する。
+ * JavaScriptで動的に生成されるページや認証が必要なサイトから情報を収集する。
+ *
+ * @requirement REQ-COLLECT-009
+ * @design DES-COLLECT-009-BrowserAutomation
+ */
+import { ok, err } from '@nahisaho/katashiro-core';
+import { DEFAULT_BROWSER_CONFIG } from './types.js';
+import { ActionExecutor } from './ActionExecutor.js';
+import { ContentExtractor } from './ContentExtractor.js';
+import { SessionManager } from './SessionManager.js';
+/**
+ * BrowserAutomation エラー
+ */
+export class BrowserAutomationError extends Error {
+    code;
+    cause;
+    constructor(code, message, cause) {
+        super(message);
+        this.code = code;
+        this.cause = cause;
+        this.name = 'BrowserAutomationError';
+    }
+}
+/**
+ * Browser Automation
+ *
+ * @example
+ * ```typescript
+ * const browser = new BrowserAutomation();
+ *
+ * // シンプルなスクレイピング
+ * const result = await browser.scrape('https://example.com');
+ * if (isOk(result)) {
+ *   console.log(result.value.content);
+ * }
+ *
+ * // スクリーンショット
+ * const screenshot = await browser.screenshot('https://example.com');
+ *
+ * await browser.close();
+ * ```
+ */
+export class BrowserAutomation {
+    config;
+    browser = null;
+    executor;
+    extractor;
+    sessionManager;
+    puppeteer = null;
+    constructor(config = {}) {
+        this.config = { ...DEFAULT_BROWSER_CONFIG, ...config };
+        this.executor = new ActionExecutor();
+        this.extractor = new ContentExtractor();
+        this.sessionManager = new SessionManager();
+    }
+    /**
+     * Puppeteerランチャーを設定（DI用）
+     */
+    setPuppeteer(puppeteer) {
+        this.puppeteer = puppeteer;
+    }
+    /**
+     * ブラウザを初期化
+     */
+    async initialize() {
+        if (this.browser)
+            return;
+        if (!this.puppeteer) {
+            // 動的インポートを試みる
+            try {
+                // @ts-ignore - puppeteer is optional peer dependency
+                const puppeteer = await import(/* webpackIgnore: true */ 'puppeteer');
+                this.puppeteer = puppeteer.default || puppeteer;
+            }
+            catch {
+                throw new BrowserAutomationError('BROWSER_NOT_INITIALIZED', 'Puppeteer is not available. Please install puppeteer: npm install puppeteer');
+            }
+        }
+        this.browser = await this.puppeteer.launch({
+            headless: this.config.headless ? 'new' : false,
+            args: [
+                '--no-sandbox',
+                '--disable-setuid-sandbox',
+                '--disable-dev-shm-usage',
+                ...(this.config.args || []),
+            ],
+        });
+    }
+    /**
+     * ブラウザを終了
+     */
+    async close() {
+        if (this.browser) {
+            await this.browser.close();
+            this.browser = null;
+        }
+    }
+    /**
+     * ブラウザが初期化されているか
+     */
+    isInitialized() {
+        return this.browser !== null;
+    }
+    /**
+     * ページをスクレイピング
+     */
+    async scrape(url, options) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                // ページに移動
+                await page.goto(url, {
+                    waitUntil: options?.waitUntil || 'domcontentloaded',
+                    timeout: options?.timeout || this.config.timeout,
+                    referer: options?.referer,
+                });
+                // コンテンツを抽出
+                const result = await this.extractor.extract(page, options?.extractors);
+                // スクリーンショット（オプション）
+                let screenshot;
+                if (options?.screenshot) {
+                    screenshot = (await page.screenshot({ type: 'png' }));
+                }
+                const extPage = page;
+                return ok({
+                    url: extPage.url(),
+                    title: await extPage.title(),
+                    content: result.content,
+                    html: result.html,
+                    extractedData: result.extractedData,
+                    links: result.links,
+                    images: result.images,
+                    metadata: result.metadata,
+                    screenshot,
+                });
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('NAVIGATION_FAILED', `Failed to scrape ${url}: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * スクリプトを実行
+     */
+    async runScript(script) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            const results = [];
+            try {
+                for (const action of script.actions) {
+                    // 変数を置換
+                    const processedAction = this.processVariables(action, script.variables);
+                    const result = await this.executor.execute(page, processedAction);
+                    results.push(result);
+                    if (!result.success) {
+                        break;
+                    }
+                }
+                return ok(results);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('ACTION_FAILED', `Script execution failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * 単一アクションを実行
+     */
+    async executeAction(url, action) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                // まずURLに移動
+                await page.goto(url, {
+                    waitUntil: 'domcontentloaded',
+                    timeout: this.config.timeout,
+                });
+                // アクションを実行
+                const result = await this.executor.execute(page, action);
+                return ok(result);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('ACTION_FAILED', `Action execution failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * スクリーンショットを取得
+     */
+    async screenshot(url, options) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                await page.goto(url, {
+                    waitUntil: options?.waitUntil || 'networkidle2',
+                    timeout: this.config.timeout,
+                });
+                const screenshot = (await page.screenshot({
+                    fullPage: options?.fullPage ?? true,
+                    type: options?.type || 'png',
+                    quality: options?.type === 'jpeg' ? options?.quality : undefined,
+                }));
+                return ok(screenshot);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('ACTION_FAILED', `Screenshot failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * PDFを生成
+     */
+    async pdf(url, options) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                await page.goto(url, {
+                    waitUntil: options?.waitUntil || 'networkidle2',
+                    timeout: this.config.timeout,
+                });
+                const pdf = await page.pdf({
+                    format: options?.format || 'A4',
+                    landscape: options?.landscape || false,
+                    printBackground: options?.printBackground ?? true,
+                });
+                return ok(pdf);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('ACTION_FAILED', `PDF generation failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * 認証付きでページにアクセス
+     */
+    async scrapeWithAuth(url, loginUrl, credentials, loginSelectors) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            const extPage = page;
+            try {
+                // ログインページに移動
+                await extPage.goto(loginUrl, {
+                    waitUntil: 'networkidle2',
+                    timeout: this.config.timeout,
+                });
+                // 認証情報を入力
+                await page.type(loginSelectors.username, credentials.username);
+                await page.type(loginSelectors.password, credentials.password);
+                // ログインボタンをクリック
+                await Promise.all([
+                    page.waitForSelector('body'), // ナビゲーション待機の代替
+                    page.click(loginSelectors.submit),
+                ]);
+                // 少し待機
+                await new Promise(resolve => setTimeout(resolve, 2000));
+                // ログイン成功を確認
+                if (loginSelectors.successIndicator) {
+                    await page.waitForSelector(loginSelectors.successIndicator, {
+                        timeout: 10000,
+                    });
+                }
+                // 目的のURLに移動
+                await extPage.goto(url, {
+                    waitUntil: 'networkidle2',
+                    timeout: this.config.timeout,
+                });
+                // コンテンツを抽出
+                const result = await this.extractor.extract(page);
+                return ok({
+                    url: extPage.url(),
+                    title: await extPage.title(),
+                    content: result.content,
+                    html: result.html,
+                    links: result.links,
+                    images: result.images,
+                    metadata: result.metadata,
+                });
+            }
+            finally {
+                await extPage.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('AUTHENTICATION_FAILED', `Authentication failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * 無限スクロールを処理
+     */
+    async scrapeInfiniteScroll(url, options) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            const extPage = page;
+            const maxScrolls = options?.maxScrolls || 10;
+            const scrollDelay = options?.scrollDelay || 1000;
+            try {
+                await extPage.goto(url, {
+                    waitUntil: 'networkidle2',
+                    timeout: this.config.timeout,
+                });
+                let previousHeight = 0;
+                let scrollCount = 0;
+                while (scrollCount < maxScrolls) {
+                    // ページの高さを取得
+                    const currentHeight = await page.evaluate('document.body.scrollHeight');
+                    if (currentHeight === previousHeight) {
+                        break;
+                    }
+                    previousHeight = currentHeight;
+                    // 下までスクロール
+                    await page.evaluate('window.scrollTo(0, document.body.scrollHeight)');
+                    // 待機
+                    await new Promise(resolve => setTimeout(resolve, scrollDelay));
+                    scrollCount++;
+                }
+                // コンテンツを抽出
+                const result = await this.extractor.extract(page);
+                return ok({
+                    url: extPage.url(),
+                    title: await extPage.title(),
+                    content: result.content,
+                    html: result.html,
+                    links: result.links,
+                    images: result.images,
+                    metadata: result.metadata,
+                });
+            }
+            finally {
+                await extPage.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('NAVIGATION_FAILED', `Infinite scroll scraping failed: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * セッションを保存
+     */
+    async saveSession(name) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                const session = await this.sessionManager.save(page, name);
+                return ok(session);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('ACTION_FAILED', `Failed to save session: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * セッションを復元
+     */
+    async loadSession(name) {
+        try {
+            await this.initialize();
+            const page = await this.createPage();
+            try {
+                await this.sessionManager.load(page, name);
+                return ok(undefined);
+            }
+            finally {
+                await page.close();
+            }
+        }
+        catch (error) {
+            return err(new BrowserAutomationError('SESSION_NOT_FOUND', `Failed to load session: ${error instanceof Error ? error.message : String(error)}`, error instanceof Error ? error : undefined));
+        }
+    }
+    /**
+     * 新しいページを作成
+     */
+    async createPage() {
+        if (!this.browser) {
+            throw new BrowserAutomationError('BROWSER_NOT_INITIALIZED', 'Browser not initialized. Call initialize() first.');
+        }
+        const page = (await this.browser.newPage());
+        // ビューポート設定
+        if (this.config.viewport) {
+            await page.setViewport(this.config.viewport);
+        }
+        // ユーザーエージェント設定
+        if (this.config.userAgent) {
+            await page.setUserAgent(this.config.userAgent);
+        }
+        // JavaScript無効化
+        if (this.config.disableJavaScript) {
+            await page.setJavaScriptEnabled(false);
+        }
+        // リクエストインターセプト（画像ブロック等）
+        if (this.config.disableImages) {
+            await page.setRequestInterception(true);
+            page.on('request', (req) => {
+                const request = req;
+                if (request.resourceType() === 'image') {
+                    request.abort();
+                }
+                else {
+                    request.continue();
+                }
+            });
+        }
+        return page;
+    }
+    /**
+     * 変数を置換
+     */
+    processVariables(action, variables) {
+        if (!variables)
+            return action;
+        const actionStr = JSON.stringify(action);
+        let processed = actionStr;
+        for (const [key, value] of Object.entries(variables)) {
+            processed = processed.replace(new RegExp(`\\$\\{${key}\\}`, 'g'), value);
+        }
+        return JSON.parse(processed);
+    }
+    /**
+     * 設定を取得
+     */
+    getConfig() {
+        return { ...this.config };
+    }
+    /**
+     * セッションマネージャーを取得
+     */
+    getSessionManager() {
+        return this.sessionManager;
+    }
+}
+// デフォルトエクスポート
+export default BrowserAutomation;
+//# sourceMappingURL=BrowserAutomation.js.map

package/dist/browser/BrowserAutomation.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"BrowserAutomation.js","sourceRoot":"","sources":["../../src/browser/BrowserAutomation.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,EAAE,EAAE,GAAG,EAAe,MAAM,0BAA0B,CAAC;AAehE,OAAO,EAAE,sBAAsB,EAAE,MAAM,YAAY,CAAC;AACpD,OAAO,EAAE,cAAc,EAAoB,MAAM,qBAAqB,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,uBAAuB,CAAC;AACzD,OAAO,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AA4CrD;;GAEG;AACH,MAAM,OAAO,sBAAuB,SAAQ,KAAK;IAE7B;IAEA;IAHlB,YACkB,IAAgC,EAChD,OAAe,EACC,KAAa;QAE7B,KAAK,CAAC,OAAO,CAAC,CAAC;QAJC,SAAI,GAAJ,IAAI,CAA4B;QAEhC,UAAK,GAAL,KAAK,CAAQ;QAG7B,IAAI,CAAC,IAAI,GAAG,wBAAwB,CAAC;IACvC,CAAC;CACF;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,OAAO,iBAAiB;IACpB,MAAM,CAAgB;IACtB,OAAO,GAAmB,IAAI,CAAC;IAC/B,QAAQ,CAAiB;IACzB,SAAS,CAAmB;IAC5B,cAAc,CAAiB;IAC/B,SAAS,GAA6B,IAAI,CAAC;IAEnD,YAAY,SAAiC,EAAE;QAC7C,IAAI,CAAC,MAAM,GAAG,EAAE,GAAG,sBAAsB,EAAE,GAAG,MAAM,EAAE,CAAC;QACvD,IAAI,CAAC,QAAQ,GAAG,IAAI,cAAc,EAAE,CAAC;QACrC,IAAI,CAAC,SAAS,GAAG,IAAI,gBAAgB,EAAE,CAAC;QACxC,IAAI,CAAC,cAAc,GAAG,IAAI,cAAc,EAAE,CAAC;IAC7C,CAAC;IAED;;OAEG;IACH,YAAY,CAAC,SAA4B;QACvC,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;IAC7B,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,UAAU;QACd,IAAI,IAAI,CAAC,OAAO;YAAE,OAAO;QAEzB,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACpB,cAAc;YACd,IAAI,CAAC;gBACH,qDAAqD;gBACrD,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,yBAAyB,CAAC,WAAW,CAAC,CAAC;gBACtE,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC,OAAO,IAAI,SAAS,CAAC;YAClD,CAAC;YAAC,MAAM,CAAC;gBACP,MAAM,IAAI,sBAAsB,CAC9B,yBAAyB,EACzB,6EAA6E,CAC9E,CAAC;YACJ,CAAC;QACH,CAAC;QAED,IAAI,CAAC,OAAO,GAAG,MAAM,IAAI,CAAC,SAAU,CAAC,MAAM,CAAC;YAC1C,QAAQ,EAAE,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK;YAC9C,IAAI,EAAE;gBACJ,cAAc;gBACd,0BAA0B;gBAC1B,yBAAyB;gBACzB,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;aAC5B;SACF,CAAC,CAAC;IACL,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,KAAK;QACT,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YACjB,MAAM,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;YAC3B,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC;QACtB,CAAC;IACH,CAAC;IAED;;OAEG;IACH,aAAa;QACX,OAAO,IAAI,CAAC,OAAO,KAAK,IAAI,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,MAAM,CACV,GAAW,EACX,OAGC;QAED,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,SAAS;gBACT,MAAO,IAAqB,CAAC,IAAI,CAAC,GAAG,EAAE;oBACrC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,kBAAkB;oBACnD,OAAO,EAAE,OAAO,EAAE,OAAO,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO;oBAChD,OAAO,EAAE,OAAO,EAAE,OAAO;iBAC1B,CAAC,CAAC;gBAEH,WAAW;gBACX,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC;gBAEvE,mBAAmB;gBACnB,IAAI,UAA8B,CAAC;gBACnC,IAAI,OAAO,EAAE,UAAU,EAAE,CAAC;oBACxB,UAAU,GAAG,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAW,CAAC;gBAClE,CAAC;gBAED,MAAM,OAAO,GAAG,IAAoB,CAAC;gBACrC,OAAO,EAAE,CAAC;oBACR,GAAG,EAAE,OAAO,CAAC,GAAG,EAAE;oBAClB,KAAK,EAAE,MAAM,OAAO,CAAC,KAAK,EAAE;oBAC5B,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,aAAa,EAAE,MAAM,CAAC,aAAa;oBACnC,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,UAAU;iBACX,CAAC,CAAC;YACL,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,mBAAmB,EACnB,oBAAoB,GAAG,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACpF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,SAAS,CACb,MAAqB;QAErB,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YACrC,MAAM,OAAO,GAAmB,EAAE,CAAC;YAEnC,IAAI,CAAC;gBACH,KAAK,MAAM,MAAM,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;oBACpC,QAAQ;oBACR,MAAM,eAAe,GAAG,IAAI,CAAC,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,SAAS,CAAC,CAAC;oBAExE,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,eAAe,CAAC,CAAC;oBAClE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;oBAErB,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;wBACpB,MAAM;oBACR,CAAC;gBACH,CAAC;gBAED,OAAO,EAAE,CAAC,OAAO,CAAC,CAAC;YACrB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,eAAe,EACf,4BAA4B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACpF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,aAAa,CACjB,GAAW,EACX,MAAqB;QAErB,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,WAAW;gBACX,MAAO,IAAqB,CAAC,IAAI,CAAC,GAAG,EAAE;oBACrC,SAAS,EAAE,kBAAkB;oBAC7B,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,WAAW;gBACX,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;gBAEzD,OAAO,EAAE,CAAC,MAAM,CAAC,CAAC;YACpB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,eAAe,EACf,4BAA4B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACpF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,UAAU,CACd,GAAW,EACX,OAEC;QAED,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,MAAO,IAAqB,CAAC,IAAI,CAAC,GAAG,EAAE;oBACrC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,cAAc;oBAC/C,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,MAAM,UAAU,GAAG,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC;oBACxC,QAAQ,EAAE,OAAO,EAAE,QAAQ,IAAI,IAAI;oBACnC,IAAI,EAAE,OAAO,EAAE,IAAI,IAAI,KAAK;oBAC5B,OAAO,EAAE,OAAO,EAAE,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC,SAAS;iBACjE,CAAC,CAAW,CAAC;gBAEd,OAAO,EAAE,CAAC,UAAU,CAAC,CAAC;YACxB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,eAAe,EACf,sBAAsB,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EAC9E,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,GAAG,CACP,GAAW,EACX,OAEC;QAED,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,MAAO,IAAqB,CAAC,IAAI,CAAC,GAAG,EAAE;oBACrC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,cAAc;oBAC/C,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC;oBACzB,MAAM,EAAE,OAAO,EAAE,MAAM,IAAI,IAAI;oBAC/B,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,KAAK;oBACtC,eAAe,EAAE,OAAO,EAAE,eAAe,IAAI,IAAI;iBAClD,CAAC,CAAC;gBAEH,OAAO,EAAE,CAAC,GAAG,CAAC,CAAC;YACjB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,eAAe,EACf,0BAA0B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EAClF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,cAAc,CAClB,GAAW,EACX,QAAgB,EAChB,WAA4B,EAC5B,cAA8B;QAE9B,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YACrC,MAAM,OAAO,GAAG,IAAoB,CAAC;YAErC,IAAI,CAAC;gBACH,aAAa;gBACb,MAAM,OAAO,CAAC,IAAI,CAAC,QAAQ,EAAE;oBAC3B,SAAS,EAAE,cAAc;oBACzB,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,UAAU;gBACV,MAAM,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,QAAQ,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC;gBAC/D,MAAM,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,QAAQ,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC;gBAE/D,eAAe;gBACf,MAAM,OAAO,CAAC,GAAG,CAAC;oBAChB,IAAI,CAAC,eAAe,CAAC,MAAM,CAAC,EAAE,eAAe;oBAC7C,IAAI,CAAC,KAAK,CAAC,cAAc,CAAC,MAAM,CAAC;iBAClC,CAAC,CAAC;gBAEH,OAAO;gBACP,MAAM,IAAI,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC;gBAExD,YAAY;gBACZ,IAAI,cAAc,CAAC,gBAAgB,EAAE,CAAC;oBACpC,MAAM,IAAI,CAAC,eAAe,CAAC,cAAc,CAAC,gBAAgB,EAAE;wBAC1D,OAAO,EAAE,KAAK;qBACf,CAAC,CAAC;gBACL,CAAC;gBAED,YAAY;gBACZ,MAAM,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE;oBACtB,SAAS,EAAE,cAAc;oBACzB,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,WAAW;gBACX,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;gBAElD,OAAO,EAAE,CAAC;oBACR,GAAG,EAAE,OAAO,CAAC,GAAG,EAAE;oBAClB,KAAK,EAAE,MAAM,OAAO,CAAC,KAAK,EAAE;oBAC5B,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;iBAC1B,CAAC,CAAC;YACL,CAAC;oBAAS,CAAC;gBACT,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;YACxB,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,uBAAuB,EACvB,0BAA0B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EAClF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,oBAAoB,CACxB,GAAW,EACX,OAIC;QAED,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YACrC,MAAM,OAAO,GAAG,IAAoB,CAAC;YACrC,MAAM,UAAU,GAAG,OAAO,EAAE,UAAU,IAAI,EAAE,CAAC;YAC7C,MAAM,WAAW,GAAG,OAAO,EAAE,WAAW,IAAI,IAAI,CAAC;YAEjD,IAAI,CAAC;gBACH,MAAM,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE;oBACtB,SAAS,EAAE,cAAc;oBACzB,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAO;iBAC7B,CAAC,CAAC;gBAEH,IAAI,cAAc,GAAG,CAAC,CAAC;gBACvB,IAAI,WAAW,GAAG,CAAC,CAAC;gBAEpB,OAAO,WAAW,GAAG,UAAU,EAAE,CAAC;oBAChC,YAAY;oBACZ,MAAM,aAAa,GAAG,MAAM,IAAI,CAAC,QAAQ,CACvC,4BAA4B,CACnB,CAAC;oBAEZ,IAAI,aAAa,KAAK,cAAc,EAAE,CAAC;wBACrC,MAAM;oBACR,CAAC;oBAED,cAAc,GAAG,aAAa,CAAC;oBAE/B,WAAW;oBACX,MAAM,IAAI,CAAC,QAAQ,CAAC,gDAAgD,CAAC,CAAC;oBAEtE,KAAK;oBACL,MAAM,IAAI,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,WAAW,CAAC,CAAC,CAAC;oBAE/D,WAAW,EAAE,CAAC;gBAChB,CAAC;gBAED,WAAW;gBACX,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;gBAElD,OAAO,EAAE,CAAC;oBACR,GAAG,EAAE,OAAO,CAAC,GAAG,EAAE;oBAClB,KAAK,EAAE,MAAM,OAAO,CAAC,KAAK,EAAE;oBAC5B,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;iBAC1B,CAAC,CAAC;YACL,CAAC;oBAAS,CAAC;gBACT,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;YACxB,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,mBAAmB,EACnB,oCAAoC,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EAC5F,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,WAAW,CACf,IAAY;QAEZ,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;gBAC3D,OAAO,EAAE,CAAC,OAAO,CAAC,CAAC;YACrB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,eAAe,EACf,2BAA2B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACnF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,WAAW,CAAC,IAAY;QAC5B,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAExB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;YAErC,IAAI,CAAC;gBACH,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;gBAC3C,OAAO,EAAE,CAAC,SAAS,CAAC,CAAC;YACvB,CAAC;oBAAS,CAAC;gBACT,MAAO,IAAqB,CAAC,KAAK,EAAE,CAAC;YACvC,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,GAAG,CACR,IAAI,sBAAsB,CACxB,mBAAmB,EACnB,2BAA2B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACnF,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAC3C,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,UAAU;QACtB,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC;YAClB,MAAM,IAAI,sBAAsB,CAC9B,yBAAyB,EACzB,mDAAmD,CACpD,CAAC;QACJ,CAAC;QAED,MAAM,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAiB,CAAC;QAE5D,WAAW;QACX,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YACzB,MAAM,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC/C,CAAC;QAED,eAAe;QACf,IAAI,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;YAC1B,MAAM,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QACjD,CAAC;QAED,gBAAgB;QAChB,IAAI,IAAI,CAAC,MAAM,CAAC,iBAAiB,EAAE,CAAC;YAClC,MAAM,IAAI,CAAC,oBAAoB,CAAC,KAAK,CAAC,CAAC;QACzC,CAAC;QAED,wBAAwB;QACxB,IAAI,IAAI,CAAC,MAAM,CAAC,aAAa,EAAE,CAAC;YAC9B,MAAM,IAAI,CAAC,sBAAsB,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,CAAC,EAAE,CAAC,SAAS,EAAE,CAAC,GAAY,EAAE,EAAE;gBAClC,MAAM,OAAO,GAAG,GAA8E,CAAC;gBAC/F,IAAI,OAAO,CAAC,YAAY,EAAE,KAAK,OAAO,EAAE,CAAC;oBACvC,OAAO,CAAC,KAAK,EAAE,CAAC;gBAClB,CAAC;qBAAM,CAAC;oBACN,OAAO,CAAC,QAAQ,EAAE,CAAC;gBACrB,CAAC;YACH,CAAC,CAAC,CAAC;QACL,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACK,gBAAgB,CACtB,MAAqB,EACrB,SAAkC;QAElC,IAAI,CAAC,SAAS;YAAE,OAAO,MAAM,CAAC;QAE9B,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QACzC,IAAI,SAAS,GAAG,SAAS,CAAC;QAE1B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;YACrD,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,IAAI,MAAM,CAAC,SAAS,GAAG,KAAK,EAAE,GAAG,CAAC,EAAE,KAAK,CAAC,CAAC;QAC3E,CAAC;QAED,OAAO,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,SAAS;QACP,OAAO,EAAE,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;IAC5B,CAAC;IAED;;OAEG;IACH,iBAAiB;QACf,OAAO,IAAI,CAAC,cAAc,CAAC;IAC7B,CAAC;CACF;AAED,cAAc;AACd,eAAe,iBAAiB,CAAC"}

package/dist/browser/ContentExtractor.d.ts

@@ -0,0 +1,54 @@
+/**
+ * ContentExtractor - ページからコンテンツを抽出
+ *
+ * @requirement REQ-COLLECT-009
+ * @design DES-COLLECT-009-BrowserAutomation
+ */
+import type { ExtractionResult, ExtractorConfig } from './types.js';
+import type { BrowserPage } from './ActionExecutor.js';
+/**
+ * ページからコンテンツを抽出
+ */
+export declare class ContentExtractor {
+    /**
+     * ページからコンテンツを抽出
+     */
+    extract(page: BrowserPage, extractors?: ExtractorConfig[]): Promise<ExtractionResult>;
+    /**
+     * メインコンテンツを抽出
+     */
+    private extractMainContent;
+    /**
+     * HTML全体を取得
+     */
+    private extractHtml;
+    /**
+     * リンクを抽出
+     */
+    private extractLinks;
+    /**
+     * 画像を抽出
+     */
+    private extractImages;
+    /**
+     * メタデータを抽出
+     */
+    private extractMetadata;
+    /**
+     * カスタム抽出
+     */
+    private extractCustomData;
+    /**
+     * テキストコンテンツを抽出（シンプル版）
+     */
+    extractText(page: BrowserPage, selector: string): Promise<string>;
+    /**
+     * 属性を抽出
+     */
+    extractAttribute(page: BrowserPage, selector: string, attribute: string): Promise<string | null>;
+    /**
+     * 複数要素のテキストを抽出
+     */
+    extractAllText(page: BrowserPage, selector: string): Promise<string[]>;
+}
+//# sourceMappingURL=ContentExtractor.d.ts.map

package/dist/browser/ContentExtractor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ContentExtractor.d.ts","sourceRoot":"","sources":["../../src/browser/ContentExtractor.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,EACV,gBAAgB,EAChB,eAAe,EAIhB,MAAM,YAAY,CAAC;AACpB,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAEvD;;GAEG;AACH,qBAAa,gBAAgB;IAC3B;;OAEG;IACG,OAAO,CACX,IAAI,EAAE,WAAW,EACjB,UAAU,CAAC,EAAE,eAAe,EAAE,GAC7B,OAAO,CAAC,gBAAgB,CAAC;IAgC5B;;OAEG;YACW,kBAAkB;IAahC;;OAEG;YACW,WAAW;IAIzB;;OAEG;YACW,YAAY;IAa1B;;OAEG;YACW,aAAa;IAc3B;;OAEG;YACW,eAAe;IA2B7B;;OAEG;YACW,iBAAiB;IAiC/B;;OAEG;IACG,WAAW,CAAC,IAAI,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IAOvE;;OAEG;IACG,gBAAgB,CACpB,IAAI,EAAE,WAAW,EACjB,QAAQ,EAAE,MAAM,EAChB,SAAS,EAAE,MAAM,GAChB,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC;IAQzB;;OAEG;IACG,cAAc,CAAC,IAAI,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;CAM7E"}