196 lines
6.5 KiB
TypeScript
196 lines
6.5 KiB
TypeScript
|
import * as puppeteer from 'puppeteer';
|
|||
|
|
import { Logger } from '@nestjs/common';
|
|||
|
|
|
|||
|
|
// 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 interface CnoocResult {
|
|||
|
|
title: string;
|
|||
|
|
publishDate: Date;
|
|||
|
|
url: string;
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
export const CnoocCrawler = {
|
|||
|
|
name: '中海油招标平台',
|
|||
|
|
url: 'https://buy.cnooc.com.cn/cbjyweb/001/001001/moreinfo.html',
|
|||
|
|
baseUrl: 'https://buy.cnooc.com.cn',
|
|||
|
|
|
|||
|
|
async crawl(browser: puppeteer.Browser): Promise<CnoocResult[]> {
|
|||
|
|
const logger = new Logger('CnoocCrawler');
|
|||
|
|
const page = await browser.newPage();
|
|||
|
|
|
|||
|
|
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: CnoocResult[] = [];
|
|||
|
|
let currentPage = 1;
|
|||
|
|
const maxPages = 5;
|
|||
|
|
|
|||
|
|
try {
|
|||
|
|
logger.log(`Navigating to ${this.url}...`);
|
|||
|
|
await page.goto(this.url, { waitUntil: 'networkidle2', timeout: 60000 });
|
|||
|
|
|
|||
|
|
// 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}...`);
|
|||
|
|
|
|||
|
|
const content = await page.content();
|
|||
|
|
const pageResults = this.extract(content);
|
|||
|
|
|
|||
|
|
if (pageResults.length === 0) {
|
|||
|
|
logger.warn(`No results found on page ${currentPage}, stopping.`);
|
|||
|
|
break;
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
allResults.push(...pageResults);
|
|||
|
|
logger.log(`Extracted ${pageResults.length} items from page ${currentPage}`);
|
|||
|
|
|
|||
|
|
// 模拟人类行为 - 翻页前
|
|||
|
|
logger.log('Simulating human mouse movements before pagination...');
|
|||
|
|
await simulateHumanMouseMovement(page);
|
|||
|
|
|
|||
|
|
logger.log('Simulating human scrolling before pagination...');
|
|||
|
|
await simulateHumanScrolling(page);
|
|||
|
|
|
|||
|
|
// 查找下一页按钮 - 中海油使用特定的分页结构
|
|||
|
|
// 下一页链接格式: <a href="https://buy.cnooc.com.cn/cbjyweb/001/001001/2.html" class="pageLink">下页 ></a>
|
|||
|
|
const nextButtonSelector = 'a.pageLink[href*="/cbjyweb/001/001001/"]';
|
|||
|
|
const nextButton = await page.$(nextButtonSelector);
|
|||
|
|
|
|||
|
|
if (!nextButton) {
|
|||
|
|
logger.log('Next page button not found. Reached end of list.');
|
|||
|
|
break;
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
logger.log(`Navigating to page ${currentPage + 1}...`);
|
|||
|
|
|
|||
|
|
try {
|
|||
|
|
// 点击下一页按钮
|
|||
|
|
await nextButton.click();
|
|||
|
|
await new Promise(r => setTimeout(r, 3000)); // 等待页面加载
|
|||
|
|
} catch (navError) {
|
|||
|
|
logger.error(`Navigation to page ${currentPage + 1} failed: ${navError.message}`);
|
|||
|
|
break;
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
currentPage++;
|
|||
|
|
|
|||
|
|
// 模拟人类行为 - 翻页后
|
|||
|
|
logger.log('Simulating human mouse movements after pagination...');
|
|||
|
|
await simulateHumanMouseMovement(page);
|
|||
|
|
|
|||
|
|
logger.log('Simulating human scrolling after pagination...');
|
|||
|
|
await simulateHumanScrolling(page);
|
|||
|
|
|
|||
|
|
// Random delay between pages
|
|||
|
|
const delay = Math.floor(Math.random() * (3000 - 1000 + 1)) + 1000;
|
|||
|
|
await new Promise(resolve => setTimeout(resolve, delay));
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
return allResults;
|
|||
|
|
|
|||
|
|
} catch (error) {
|
|||
|
|
logger.error(`Failed to crawl ${this.name}: ${error.message}`);
|
|||
|
|
return allResults;
|
|||
|
|
} finally {
|
|||
|
|
await page.close();
|
|||
|
|
}
|
|||
|
|
},
|
|||
|
|
|
|||
|
|
extract(html: string): CnoocResult[] {
|
|||
|
|
const results: CnoocResult[] = [];
|
|||
|
|
/**
|
|||
|
|
* Regex groups for buy.cnooc.com.cn:
|
|||
|
|
* 1: URL (href属性)
|
|||
|
|
* 2: Title (font标签内的文本)
|
|||
|
|
* 3: Date (发布时间，格式：2026-01-12)
|
|||
|
|
*
|
|||
|
|
* HTML结构示例:
|
|||
|
|
* <li class="now-hd-items clearfix">
|
|||
|
|
* <a href="https://buy.cnooc.com.cn/cbjyweb/001/001001/20260112/1460280812582768641-zhy.html" target="_blank" class="now-link" title="...">
|
|||
|
|
* <font style="font-weight:bold">中海油服-物探事业部2026-2028年度海事许可办理及码头服务(二次)</font>
|
|||
|
|
* </a>
|
|||
|
|
* <span class="now-span" style="width:100px">2026-01-12</span>
|
|||
|
|
* </li>
|
|||
|
|
*/
|
|||
|
|
const regex = /<li class="now-hd-items clearfix">[\s\S]*?<a[^>]*href="([^"]*)"[^>]*>[\s\S]*?<font[^>]*>([^<]*)<\/font>[\s\S]*?<span class="now-span"[^>]*>\s*(\d{4}-\d{2}-\d{2})\s*<\/span>[\s\S]*?<\/li>/gs;
|
|||
|
|
|
|||
|
|
let match;
|
|||
|
|
while ((match = regex.exec(html)) !== null) {
|
|||
|
|
const url = match[1]?.trim();
|
|||
|
|
const title = match[2]?.trim();
|
|||
|
|
const dateStr = match[3]?.trim();
|
|||
|
|
|
|||
|
|
if (title && url) {
|
|||
|
|
results.push({
|
|||
|
|
title,
|
|||
|
|
publishDate: dateStr ? new Date(dateStr) : new Date(),
|
|||
|
|
url: url.startsWith('http') ? url : this.baseUrl + url
|
|||
|
|
});
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
return results;
|
|||
|
|
}
|
|||
|
|
};
|