feat: 全面优化爬虫系统和数据处理能力

- 增强数据重试机制:对数据为0的爬虫自动重试,提高数据完整性
- 优化前端筛选逻辑:改进日期筛选,只限制开始时间,更灵活的数据查看
- 新增最近数据接口:添加 /api/bids/recent 获取30天内最新招标数据
- 改进统计展示:实时显示筛选结果数量,优化用户体验
- 完善日志系统:确保日志目录自动创建,避免启动错误
- 增强独立脚本:使用自定义logger,完善错误处理和程序关闭
- 优化主程序:集成自定义日志服务,统一日志格式
- 扩展npm脚本:新增 web 命令用于构建前端
- 改进大唐爬虫:延长等待时间到60秒,提高页面加载成功率
- 优化数据筛选:今日招标改为使用独立接口,提升性能
This commit is contained in:
dmy
2026-01-12 12:28:37 +08:00
parent 3e6456e120
commit 1b28a3462a
10 changed files with 104 additions and 39 deletions

View File

@@ -32,6 +32,9 @@ export class BidCrawlerService {
// 统计结果
const crawlResults: Record<string, { success: number; error?: string }> = {};
// 记录数据为0的爬虫用于重试
const zeroDataCrawlers: any[] = [];
// 从环境变量读取代理配置
const proxyHost = this.configService.get<string>('PROXY_HOST');
const proxyPort = this.configService.get<string>('PROXY_PORT');
@@ -83,6 +86,11 @@ export class BidCrawlerService {
// 记录成功数量
crawlResults[crawler.name] = { success: results.length };
// 如果数据为0记录下来用于重试
if (results.length === 0) {
zeroDataCrawlers.push(crawler);
}
for (const item of results) {
await this.bidsService.createOrUpdate({
title: item.title,
@@ -98,6 +106,45 @@ export class BidCrawlerService {
crawlResults[crawler.name] = { success: 0, error: err.message };
}
}
// 对数据为0的爬虫进行重试
if (zeroDataCrawlers.length > 0) {
this.logger.log(`Retrying ${zeroDataCrawlers.length} crawlers with zero data...`);
for (const crawler of zeroDataCrawlers) {
this.logger.log(`Retrying: ${crawler.name}`);
// 检查是否超时
const elapsedTime = Date.now() - startTime;
if (elapsedTime > maxExecutionTime) {
this.logger.warn(`⚠️ Crawl task exceeded maximum execution time of 3 hours. Stopping retry...`);
this.logger.warn(`⚠️ Total elapsed time: ${Math.floor(elapsedTime / 1000 / 60)} minutes`);
break;
}
try {
const results = await crawler.crawl(browser);
this.logger.log(`Retry extracted ${results.length} items from ${crawler.name}`);
// 更新统计结果
crawlResults[crawler.name] = { success: results.length };
for (const item of results) {
await this.bidsService.createOrUpdate({
title: item.title,
url: item.url,
publishDate: item.publishDate,
source: crawler.name,
unit: '',
});
}
} catch (err) {
this.logger.error(`Error retrying ${crawler.name}: ${err.message}`);
// 记录错误信息
crawlResults[crawler.name] = { success: 0, error: err.message };
}
}
}
} catch (error) {
this.logger.error(`Crawl task failed: ${error.message}`);
} finally {