- 在 BidItem 实体中添加 priority 和 unit 字段,扩展数据结构 - 将爬虫浏览器模式改为非无头模式(headless: false)便于调试 - 为 ChngCrawler 添加人类行为模拟功能: * 模拟鼠标随机移动,增加移动步数和停顿时间 * 模拟人类滚动行为,包括随机滚动距离和停顿 * 添加 navigator 属性伪装,包括语言、插件等 - 在关键节点添加截图功能(bing.png, newPage.png, huaneng.png)用于调试 - 优化反检测策略,降低被目标网站识别为机器人的风险
234 lines
8.4 KiB
TypeScript
234 lines
8.4 KiB
TypeScript
import * as puppeteer from 'puppeteer';
|
||
import { Logger } from '@nestjs/common';
|
||
import { ChdtpResult } from './chdtp_target';
|
||
|
||
// 模拟人类鼠标移动
|
||
async function simulateHumanMouseMovement(page: puppeteer.Page) {
|
||
const viewport = page.viewport();
|
||
if (!viewport) return;
|
||
|
||
const movements = 5 + Math.floor(Math.random() * 5); // 5-10次随机移动
|
||
|
||
for (let i = 0; i < movements; i++) {
|
||
const x = Math.floor(Math.random() * viewport.width);
|
||
const y = Math.floor(Math.random() * viewport.height);
|
||
|
||
await page.mouse.move(x, y, {
|
||
steps: 10 + Math.floor(Math.random() * 20) // 10-30步,使移动更平滑
|
||
});
|
||
|
||
// 随机停顿 100-500ms
|
||
await new Promise(r => setTimeout(r, 100 + Math.random() * 400));
|
||
}
|
||
}
|
||
|
||
// 模拟人类滚动
|
||
async function simulateHumanScrolling(page: puppeteer.Page) {
|
||
const scrollCount = 3 + Math.floor(Math.random() * 5); // 3-7次滚动
|
||
|
||
for (let i = 0; i < scrollCount; i++) {
|
||
const scrollDistance = 100 + Math.floor(Math.random() * 400); // 100-500px
|
||
|
||
await page.evaluate((distance) => {
|
||
window.scrollBy({
|
||
top: distance,
|
||
behavior: 'smooth'
|
||
});
|
||
}, scrollDistance);
|
||
|
||
// 随机停顿 500-1500ms
|
||
await new Promise(r => setTimeout(r, 500 + Math.random() * 1000));
|
||
}
|
||
|
||
// 滚动回顶部
|
||
await page.evaluate(() => {
|
||
window.scrollTo({ top: 0, behavior: 'smooth' });
|
||
});
|
||
await new Promise(r => setTimeout(r, 1000));
|
||
}
|
||
|
||
export const ChngCrawler = {
|
||
name: '华能集团电子商务平台',
|
||
url: 'https://ec.chng.com.cn/ecmall/index.html#/purchase/home?top=0',
|
||
baseUrl: 'https://ec.chng.com.cn/ecmall/index.html',
|
||
|
||
async crawl(browser: puppeteer.Browser): Promise<ChdtpResult[]> {
|
||
const logger = new Logger('ChngCrawler');
|
||
let page = await browser.newPage();
|
||
// await page.setViewport({ width: 1920, height: 1080, deviceScaleFactor: 1 });
|
||
// await page.setViewport({ deviceScaleFactor: 1 });
|
||
const username = process.env.PROXY_USERNAME;
|
||
const password = process.env.PROXY_PASSWORD;
|
||
if (username && password) {
|
||
await page.authenticate({ username, password });
|
||
}
|
||
|
||
await page.evaluateOnNewDocument(() => {
|
||
Object.defineProperty(navigator, 'webdriver', { get: () => false });
|
||
Object.defineProperty(navigator, 'language', { get: () => "zh-CN"});
|
||
Object.defineProperty(navigator, 'plugins', { get: () => [1,2,3,4,5]});
|
||
});
|
||
|
||
await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Safari/537.36');
|
||
await page.setViewport({ width: 1920, height: 1080 });
|
||
|
||
const allResults: ChdtpResult[] = [];
|
||
let currentPage = 1;
|
||
const maxPages = 5;
|
||
|
||
try {
|
||
logger.log('Navigating to Bing...');
|
||
await page.goto('https://cn.bing.com', { waitUntil: 'networkidle2' });
|
||
|
||
logger.log('Searching for target site...');
|
||
const searchBoxSelector = 'input[name="q"]';
|
||
await page.waitForSelector(searchBoxSelector);
|
||
await page.type(searchBoxSelector, 'https://ec.chng.com.cn/');
|
||
await page.keyboard.press('Enter');
|
||
await page.waitForNavigation({ waitUntil: 'networkidle2' });
|
||
|
||
logger.log('Clicking search result...');
|
||
await page.screenshot({ path: 'bing.png' });
|
||
const firstResultSelector = '#b_results .b_algo h2 a';
|
||
await page.waitForSelector(firstResultSelector);
|
||
|
||
const newTargetPromise = browser.waitForTarget(target => target.opener() === page.target());
|
||
await page.click(firstResultSelector);
|
||
|
||
const newTarget = await newTargetPromise;
|
||
const newPage = await newTarget.page();
|
||
|
||
if (newPage) {
|
||
await newPage.screenshot({ path: 'newPage.png' });
|
||
await page.close();
|
||
page = newPage;
|
||
if (username && password) {
|
||
await page.authenticate({ username, password });
|
||
}
|
||
}
|
||
// 模拟人类行为
|
||
logger.log('Simulating human mouse movements...');
|
||
await simulateHumanMouseMovement(page);
|
||
|
||
logger.log('Simulating human scrolling...');
|
||
await simulateHumanScrolling(page);
|
||
|
||
await page.waitForNavigation({ waitUntil: 'domcontentloaded' }).catch(() => {});
|
||
// 模拟人类行为
|
||
logger.log('Simulating human mouse movements...');
|
||
await simulateHumanMouseMovement(page);
|
||
|
||
logger.log('Simulating human scrolling...');
|
||
await simulateHumanScrolling(page);
|
||
|
||
|
||
// PAUSE 15 SECONDS as requested
|
||
logger.log('Pausing 15 seconds before looking for "采购专栏"...');
|
||
await new Promise(r => setTimeout(r, 15000));
|
||
await page.screenshot({ path: 'huaneng.png' });
|
||
|
||
logger.log('Looking for "采购专栏" link...');
|
||
await page.waitForFunction(() => {
|
||
const divs = Array.from(document.querySelectorAll('div.text'));
|
||
return divs.some(div => div.textContent && div.textContent.includes('采购专栏'));
|
||
}, { timeout: 60000 });
|
||
|
||
const purchaseTargetPromise = browser.waitForTarget(target => target.opener() === page.target(), { timeout: 15000 }).catch(() => null);
|
||
|
||
await page.evaluate(() => {
|
||
const divs = Array.from(document.querySelectorAll('div.text'));
|
||
const target = divs.find(div => div.textContent && div.textContent.includes('采购专栏')) as HTMLElement;
|
||
if (target) target.click();
|
||
});
|
||
|
||
const purchaseTarget = await purchaseTargetPromise;
|
||
if (purchaseTarget) {
|
||
const pPage = await purchaseTarget.page();
|
||
if (pPage) {
|
||
logger.log('Switched to Purchase Page tab.');
|
||
page = pPage;
|
||
if (username && password) {
|
||
await page.authenticate({ username, password });
|
||
}
|
||
await new Promise(r => setTimeout(r, 5000));
|
||
}
|
||
}
|
||
|
||
logger.log(`Active URL: ${page.url()}`);
|
||
|
||
// 模拟人类行为
|
||
logger.log('Simulating human mouse movements...');
|
||
await simulateHumanMouseMovement(page);
|
||
|
||
logger.log('Simulating human scrolling...');
|
||
await simulateHumanScrolling(page);
|
||
|
||
while (currentPage <= maxPages) {
|
||
logger.log(`Processing page ${currentPage}...`);
|
||
|
||
// Wait for table rows to load
|
||
await page.waitForFunction(() => {
|
||
return document.querySelectorAll('tr.ant-table-row').length > 0;
|
||
}, { timeout: 60000 }).catch(() => logger.warn('Content not found. Site might be slow.'));
|
||
|
||
const pageResults = await page.evaluate((baseUrl) => {
|
||
// Extract from table rows
|
||
const items = Array.from(document.querySelectorAll('tr.ant-table-row'));
|
||
return items.map(item => {
|
||
const titleSpan = item.querySelector('span.list-text');
|
||
const dateCell = item.querySelector('td.ant-table-row-cell-break-word p');
|
||
|
||
if (titleSpan && dateCell) {
|
||
const title = titleSpan.textContent?.trim() || '';
|
||
const dateStr = dateCell.textContent?.trim() || '';
|
||
|
||
if (title.length < 5) return null; // Filter noise
|
||
|
||
// URL is not directly available in the table, need to construct from data-row-key
|
||
const rowKey = item.getAttribute('data-row-key');
|
||
const url = rowKey ? `${baseUrl}#/purchase/detail?id=${rowKey}` : '';
|
||
|
||
return {
|
||
title,
|
||
dateStr,
|
||
url
|
||
};
|
||
}
|
||
return null;
|
||
}).filter(i => i !== null);
|
||
}, this.baseUrl);
|
||
|
||
if (pageResults.length === 0) {
|
||
logger.warn(`No results found on page ${currentPage}. Extraction failed.`);
|
||
break;
|
||
}
|
||
|
||
allResults.push(...pageResults.map(r => ({
|
||
title: r!.title,
|
||
publishDate: new Date(r!.dateStr),
|
||
url: r!.url
|
||
})));
|
||
|
||
logger.log(`Extracted ${pageResults.length} items.`);
|
||
|
||
// Pagination: look for the "right" icon SVG
|
||
const nextButton = await page.$('svg[data-icon="right"]');
|
||
if (!nextButton) break;
|
||
|
||
await nextButton.click();
|
||
await new Promise(r => setTimeout(r, 5000));
|
||
currentPage++;
|
||
}
|
||
|
||
return allResults;
|
||
|
||
} catch (error) {
|
||
logger.error(`Crawl failed: ${error.message}`);
|
||
return allResults;
|
||
} finally {
|
||
if (page) await page.close();
|
||
}
|
||
},
|
||
|
||
extract() { return []; }
|
||
}; |