Merge branch 'main' of https://git.yasue.org/ken/tripadviser_scraper

src/index.ts

@@ -9,7 +9,7 @@ import chrome, { ServiceBuilder } from 'selenium-webdriver/chrome';
 import * as fs from 'fs';
 import * as path from 'path';
 import { getCities } from './lib/cities';
-import { WebDriverUtils, saveContactInfoToCSV } from './lib/utils';
+import { WebDriverUtils, saveContactInfoToCSV, useExistingChrome, disableCookiesInChrome } from './lib/utils';
 import * as UIActions from './lib/UIActions';
 import { randomUUID } from 'crypto';
 
@@ -23,17 +23,8 @@ async function visitCityPages(): Promise<void> {
 
     console.log('Connecting to existing Chrome browser...');
 
-    // Connect to an existing Chrome browser running in debug mode on port 9222
+    const driver = await useExistingChrome();
-    const options = new chrome.Options();
+    if (!driver) return;
-
-    // Set the debugger address to connect to the existing Chrome instance
-    options.debuggerAddress('localhost:9222');
-
-    // Create WebDriver instance that connects to the existing browser
-    const driver: WebDriver = await new Builder()
-        .forBrowser('chrome')
-        .setChromeOptions(options)
-        .build();
 
     // Visit each city's TripAdvisor page
     for (let i = 0; i < cities.length; i++) {

src/lib/utils.ts

@@ -2,7 +2,8 @@
  * Utility class for common WebDriver operations
  */
 
-import { WebDriver, By, until } from 'selenium-webdriver';
+import { Builder, By, until, WebDriver } from 'selenium-webdriver';
+import chrome from 'selenium-webdriver/chrome';
 import { writeFileSync, existsSync, appendFileSync } from 'fs';
 import * as path from 'path';
 import { ContactInfo } from './types';
@@ -44,3 +45,58 @@ export function saveContactInfoToCSV(city: string, contactInfo: ContactInfo, fil
 
     console.log(`Contact info saved to ${filePath}`);
 }
+
+export async function disableCookiesInChrome(): Promise<WebDriver | null> {
+    // Set Chrome options
+    const options = new chrome.Options();
+
+    // 1. Block all cookies
+    //options.setUserPreferences({
+    //    'profile.default_content_setting_values.cookies': 2, // 2 = Block all
+    //    'profile.block_third_party_cookies': true
+    //});
+
+    // 2. Optional: Launch in incognito for extra privacy
+    options.addArguments('--incognito');
+    options.addArguments('--start-maximized');
+
+    let driver: WebDriver | null = null;
+
+    try {
+        driver = await new Builder()
+            .forBrowser('chrome')
+            .setChromeOptions(options)
+            .build();
+
+        await driver.get('https://www.tripadvisor.com/');
+
+        console.log('Chrome launched with cookies disabled.');
+
+        // Optional: Verify cookies are blocked by trying to set/get a cookie
+        await driver.manage().addCookie({ name: 'test', value: '123' });
+        const cookies = await driver.manage().getCookies();
+        console.log('Cookies after trying to add:', cookies); // Should be empty or restricted
+
+        return driver;
+    } catch (error) {
+        console.error('Error:', error);
+        return driver;
+    }
+
+}
+
+export async function useExistingChrome(): Promise<WebDriver> {
+    // Connect to an existing Chrome browser running in debug mode on port 9222
+    const options = new chrome.Options();
+
+    // Set the debugger address to connect to the existing Chrome instance
+    options.debuggerAddress('localhost:9222');
+
+    // Create WebDriver instance that connects to the existing browser
+    const driver: WebDriver = await new Builder()
+        .forBrowser('chrome')
+        .setChromeOptions(options)
+        .build();
+
+    return driver;
+}