feat: 添加爬虫统计信息页面和功能

新增爬虫统计信息页面，展示各来源的爬取数量、最新更新时间、错误信息等统计指标
后端添加爬虫统计信息存储和查询接口，记录每次爬取的结果
支持按关键词过滤招标信息查询

frontend/package.json

@@ -6,7 +6,7 @@
   "scripts": {
     "dev": "vite",
     "build": "vue-tsc -b && vite build",
-    "build:watch": "vue-tsc -b && vite build --watch",
+    "build:watch": "concurrently \"vue-tsc -b --watch\" \"vite build --watch\"",
     "preview": "vite preview"
   },
   "dependencies": {

frontend/src/App.vue

@@ -26,6 +26,10 @@
           <el-icon><Setting /></el-icon>
           <span>Keywords</span>
         </el-menu-item>
+        <el-menu-item index="5">
+          <el-icon><Connection /></el-icon>
+          <span>Crawl Info</span>
+        </el-menu-item>
       </el-menu>
     </el-aside>
 
@@ -66,6 +70,10 @@
           :loading="loading"
           @refresh="fetchData"
         />
+
+        <CrawlInfo
+          v-if="activeIndex === '5'"
+        />
       </el-main>
     </el-container>
   </el-container>
@@ -75,11 +83,12 @@
 import { ref, onMounted } from 'vue'
 import axios from 'axios'
 import { ElMessage } from 'element-plus'
-import { DataBoard, Document, Setting, MagicStick } from '@element-plus/icons-vue'
+import { DataBoard, Document, Setting, MagicStick, Connection } from '@element-plus/icons-vue'
 import Dashboard from './components/Dashboard.vue'
 import DashboardAI from './components/Dashboard-AI.vue'
 import Bids from './components/Bids.vue'
 import Keywords from './components/Keywords.vue'
+import CrawlInfo from './components/CrawlInfo.vue'
 
 const activeIndex = ref('1')
 const bids = ref<any[]>([])

frontend/src/components/CrawlInfo.vue

@@ -0,0 +1,135 @@
+<template>
+  <div class="crawl-info">
+    <el-card>
+      <template #header>
+        <div class="card-header">
+          <span>爬虫统计信息</span>
+          <el-button type="primary" size="small" @click="fetchCrawlStats" :loading="loading">
+            <el-icon><Refresh /></el-icon>
+            刷新
+          </el-button>
+        </div>
+      </template>
+
+      <el-table :data="crawlStats" stripe style="width: 100%" v-loading="loading">
+        <el-table-column prop="source" label="爬虫来源" width="200" />
+        <el-table-column prop="count" label="本次获取数量" width="120" sortable />
+        <el-table-column label="最近更新时间" width="180">
+          <template #default="{ row }">
+            {{ formatDate(row.latestUpdate) }}
+          </template>
+        </el-table-column>
+        <el-table-column label="最新工程时间" width="180">
+          <template #default="{ row }">
+            {{ formatDate(row.latestPublishDate) }}
+          </template>
+        </el-table-column>
+        <el-table-column label="状态" width="100">
+          <template #default="{ row }">
+            <el-tag :type="row.error ? 'danger' : (row.count > 0 ? 'success' : 'info')">
+              {{ row.error ? '出错' : (row.count > 0 ? '正常' : '无数据') }}
+            </el-tag>
+          </template>
+        </el-table-column>
+        <el-table-column label="错误信息" min-width="200">
+          <template #default="{ row }">
+            <span v-if="row.error" style="color: #f56c6c">{{ row.error }}</span>
+            <span v-else>-</span>
+          </template>
+        </el-table-column>
+      </el-table>
+
+      <div class="summary" v-if="crawlStats.length > 0">
+        <el-descriptions :column="3" border>
+          <el-descriptions-item label="爬虫来源总数">
+            {{ crawlStats.length }}
+          </el-descriptions-item>
+          <el-descriptions-item label="本次获取总数">
+            {{ totalCount }}
+          </el-descriptions-item>
+          <el-descriptions-item label="有数据来源">
+            {{ activeSources }}
+          </el-descriptions-item>
+          <el-descriptions-item label="出错来源">
+            {{ errorSources }}
+          </el-descriptions-item>
+        </el-descriptions>
+      </div>
+    </el-card>
+  </div>
+</template>
+
+<script setup lang="ts">
+import { ref, computed, onMounted } from 'vue'
+import axios from 'axios'
+import { ElMessage } from 'element-plus'
+import { Refresh } from '@element-plus/icons-vue'
+
+interface CrawlStat {
+  source: string
+  count: number
+  latestUpdate: string | null
+  latestPublishDate: string | null
+  error: string | null
+}
+
+const crawlStats = ref<CrawlStat[]>([])
+const loading = ref(false)
+
+const totalCount = computed(() => {
+  return crawlStats.value.reduce((sum, item) => sum + item.count, 0)
+})
+
+const activeSources = computed(() => {
+  return crawlStats.value.filter(item => item.count > 0).length
+})
+
+const errorSources = computed(() => {
+  return crawlStats.value.filter(item => item.error).length
+})
+
+const formatDate = (dateStr: string | null) => {
+  if (!dateStr) return '-'
+  const date = new Date(dateStr)
+  return date.toLocaleString('zh-CN', {
+    year: 'numeric',
+    month: '2-digit',
+    day: '2-digit',
+    hour: '2-digit',
+    minute: '2-digit'
+  })
+}
+
+const fetchCrawlStats = async () => {
+  loading.value = true
+  try {
+    const res = await axios.get('/api/bids/crawl-info-stats')
+    crawlStats.value = res.data
+  } catch (error) {
+    console.error('Failed to fetch crawl stats:', error)
+    ElMessage.error('获取爬虫统计信息失败')
+  } finally {
+    loading.value = false
+  }
+}
+
+onMounted(() => {
+  fetchCrawlStats()
+})
+</script>
+
+<style scoped>
+.crawl-info {
+  padding: 20px;
+}
+
+.card-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+}
+
+.summary {
+  margin-top: 20px;
+}
+</style>

package.json

@@ -55,6 +55,7 @@
     "@types/jest": "^30.0.0",
     "@types/node": "^22.10.7",
     "@types/supertest": "^6.0.2",
+    "concurrently": "^9.2.1",
     "eslint": "^9.18.0",
     "eslint-config-prettier": "^10.0.1",
     "eslint-plugin-prettier": "^5.2.2",

src/app.module.ts

@@ -17,7 +17,7 @@ import { AiModule } from './ai/ai.module';
     ScheduleModule.forRoot(),
     ServeStaticModule.forRoot({
       rootPath: join(__dirname, '..', 'frontend', 'dist'),
-      exclude: ['/api/(.*)'],
+      exclude: ['/api/:path(*)'],
     }),
     LoggerModule,
     DatabaseModule,

src/bids/controllers/bid.controller.ts

@@ -26,7 +26,13 @@ export class BidsController {
   }
 
   @Get('by-date-range')
-  getByDateRange(@Query('startDate') startDate: string, @Query('endDate') endDate: string) {
+  getByDateRange(@Query('startDate') startDate: string, @Query('endDate') endDate?: string, @Query('keywords') keywords?: string) {
-    return this.bidsService.getBidsByDateRange(startDate, endDate);
+    const keywordsArray = keywords ? keywords.split(',') : undefined;
+    return this.bidsService.getBidsByDateRange(startDate, endDate, keywordsArray);
+  }
+
+  @Get('crawl-info-stats')
+  getCrawlInfoStats() {
+    return this.bidsService.getCrawlInfoAddStats();
   }
 }

src/bids/services/bid.service.ts

@@ -88,7 +88,7 @@ export class BidsService {
       .getMany();
   }
 
-  async getBidsByDateRange(startDate?: string, endDate?: string) {
+  async getBidsByDateRange(startDate?: string, endDate?: string, keywords?: string[]) {
     const qb = this.bidRepository.createQueryBuilder('bid');
 
     if (startDate) {
@@ -103,6 +103,49 @@ export class BidsService {
       qb.andWhere('bid.publishDate <= :endDate', { endDate: end });
     }
 
+    if (keywords && keywords.length > 0) {
+      const keywordConditions = keywords.map((keyword, index) => {
+        return `bid.title LIKE :keyword${index}`;
+      }).join(' OR ');
+      qb.andWhere(`(${keywordConditions})`, keywords.reduce((params, keyword, index) => {
+        params[`keyword${index}`] = `%${keyword}%`;
+        return params;
+      }, {}));
+    }
+
     return qb.orderBy('bid.publishDate', 'DESC').getMany();
   }
+
+  async getCrawlInfoAddStats() {
+    const { InjectRepository } = require('@nestjs/typeorm');
+    const { Repository } = require('typeorm');
+    const { CrawlInfoAdd } = require('../../crawler/entities/crawl-info-add.entity');
+
+    // 获取每个来源的最新一次爬虫记录
+    const query = `
+      SELECT
+        source,
+        count,
+        latestPublishDate,
+        error,
+        createdAt as latestUpdate
+      FROM crawl_info_add
+      WHERE id IN (
+        SELECT MAX(id)
+        FROM crawl_info_add
+        GROUP BY source
+      )
+      ORDER BY source ASC
+    `;
+
+    const results = await this.bidRepository.query(query);
+
+    return results.map((item: any) => ({
+      source: item.source,
+      count: item.count,
+      latestUpdate: item.latestUpdate,
+      latestPublishDate: item.latestPublishDate,
+      error: item.error,
+    }));
+  }
 }

src/crawler/crawler.module.ts

@@ -1,10 +1,12 @@
 import { Module } from '@nestjs/common';
+import { TypeOrmModule } from '@nestjs/typeorm';
 import { BidCrawlerService } from './services/bid-crawler.service';
 import { CrawlerController } from './crawler.controller';
 import { BidsModule } from '../bids/bids.module';
+import { CrawlInfoAdd } from './entities/crawl-info-add.entity';
 
 @Module({
-  imports: [BidsModule],
+  imports: [BidsModule, TypeOrmModule.forFeature([CrawlInfoAdd])],
   controllers: [CrawlerController],
   providers: [BidCrawlerService],
   exports: [BidCrawlerService],

src/crawler/entities/crawl-info-add.entity.ts

@@ -0,0 +1,22 @@
+import { Entity, PrimaryGeneratedColumn, Column, CreateDateColumn } from 'typeorm';
+
+@Entity('crawl_info_add')
+export class CrawlInfoAdd {
+  @PrimaryGeneratedColumn('uuid')
+  id: string;
+
+  @Column()
+  source: string;
+
+  @Column()
+  count: number;
+
+  @Column({ type: 'datetime', nullable: true })
+  latestPublishDate: Date | null;
+
+  @Column({ type: 'text', nullable: true })
+  error: string;
+
+  @CreateDateColumn()
+  createdAt: Date;
+}

src/crawler/services/bid-crawler.service.ts

@@ -1,7 +1,10 @@
 import { Injectable, Logger } from '@nestjs/common';
 import { ConfigService } from '@nestjs/config';
+import { InjectRepository } from '@nestjs/typeorm';
+import { Repository } from 'typeorm';
 import * as puppeteer from 'puppeteer';
 import { BidsService } from '../../bids/services/bid.service';
+import { CrawlInfoAdd } from '../entities/crawl-info-add.entity';
 import { ChdtpCrawler } from './chdtp_target';
 import { ChngCrawler } from './chng_target';
 import { SzecpCrawler } from './szecp_target';
@@ -22,6 +25,8 @@ export class BidCrawlerService {
   constructor(
     private bidsService: BidsService,
     private configService: ConfigService,
+    @InjectRepository(CrawlInfoAdd)
+    private crawlInfoRepository: Repository<CrawlInfoAdd>,
   ) {}
 
   async crawlAll() {
@@ -93,6 +98,14 @@ export class BidCrawlerService {
             zeroDataCrawlers.push(crawler);
           }
 
+          // 获取最新的发布日期
+          const latestPublishDate = results.length > 0
+            ? results.reduce((latest, item) => {
+                const itemDate = new Date(item.publishDate);
+                return itemDate > latest ? itemDate : latest;
+              }, new Date(0))
+            : null;
+
           for (const item of results) {
             await this.bidsService.createOrUpdate({
               title: item.title,
@@ -102,10 +115,16 @@ export class BidCrawlerService {
               unit: '',
             });
           }
+
+          // 保存爬虫统计信息到数据库
+          await this.saveCrawlInfo(crawler.name, results.length, latestPublishDate);
         } catch (err) {
           this.logger.error(`Error crawling ${crawler.name}: ${err.message}`);
           // 记录错误信息
           crawlResults[crawler.name] = { success: 0, error: err.message };
+
+          // 保存错误信息到数据库
+          await this.saveCrawlInfo(crawler.name, 0, null, err.message);
         }
       }
       
@@ -131,6 +150,14 @@ export class BidCrawlerService {
             // 更新统计结果
             crawlResults[crawler.name] = { success: results.length };
 
+            // 获取最新的发布日期
+            const latestPublishDate = results.length > 0
+              ? results.reduce((latest, item) => {
+                  const itemDate = new Date(item.publishDate);
+                  return itemDate > latest ? itemDate : latest;
+                }, new Date(0))
+              : null;
+
             for (const item of results) {
               await this.bidsService.createOrUpdate({
                 title: item.title,
@@ -140,10 +167,16 @@ export class BidCrawlerService {
                 unit: '',
               });
             }
+
+            // 更新爬虫统计信息到数据库
+            await this.saveCrawlInfo(crawler.name, results.length, latestPublishDate);
           } 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 };
+
+            // 更新错误信息到数据库
+            await this.saveCrawlInfo(crawler.name, 0, null, err.message);
           }
         }
       }
@@ -184,4 +217,24 @@ export class BidCrawlerService {
       this.logger.log('='.repeat(50));
     }
   }
+
+  private async saveCrawlInfo(
+    source: string,
+    count: number,
+    latestPublishDate: Date | null,
+    error?: string,
+  ) {
+    try {
+      const crawlInfo = this.crawlInfoRepository.create({
+        source,
+        count,
+        latestPublishDate,
+        error,
+      });
+      await this.crawlInfoRepository.save(crawlInfo);
+      this.logger.log(`Saved crawl info for ${source}: ${count} items`);
+    } catch (err) {
+      this.logger.error(`Failed to save crawl info for ${source}: ${err.message}`);
+    }
+  }
 }