tripadviser_scraper/src/lib/utils.ts

/**
 * Utility class for common WebDriver operations
 */

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';

export class WebDriverUtils {
    /**
     * Wait for a specified number of seconds
     * @param seconds Number of seconds to wait
     * @returns Promise that resolves after the specified time
     */
    static async wait(driver?: WebDriver): Promise<void> {
        const seconds = Math.floor(Math.random() * 1000) % 3 + 3;
        console.log(`Scrolling to bottom for ${seconds} seconds...`);

        const endTime = Date.now() + seconds * 1000;

        let scrollCounter = 0;

        while (Date.now() < endTime) {

            /*
            try {
                if (driver) {
                    await driver.executeScript(`
                        window.scrollBy(0, 10);
                    `);

                    scrollCounter++;
                }
            } catch (error) {
                console.warn('Scroll failed:', error);
            }
            */

            // Wait a little between scrolls
            await new Promise(resolve => setTimeout(resolve, 500));
        }
    }

    /**
     * Wait for an element to be located on the page
     * @param driver WebDriver instance
     * @param selector CSS selector for the element
     * @param timeoutMs Timeout in milliseconds (default: 10000)
     * @returns Promise that resolves when the element is found
     */
    static async waitForElement(driver: WebDriver, selector: string, timeoutMs: number = 10000) {
        console.log(`Waiting for element: ${selector}`);
        await driver.wait(until.elementLocated(By.css(selector)), timeoutMs);
    }
}

export function saveContactInfoToCSV(city: string, contactInfo: ContactInfo, filePath: string): void {
    const headers = 'City,Website URL,Email\n';
    const line = `"${city}","${contactInfo.name}","${contactInfo.websiteUrl}","${contactInfo.email}"\n`;

    if (!existsSync(filePath)) {
        writeFileSync(filePath, headers + line);
    } else {
        appendFileSync(filePath, line);
    }

    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 useChrome(initialUrl: string = "https://www.tripadvisor.com/"): 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(initialUrl);

        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;
}