smart-image-scraper-mcp 2.5.0 → 2.5.2

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "smart-image-scraper-mcp",
-  "version": "2.5.0",
+  "version": "2.5.2",
   "description": "全网智能图片抓取 MCP 服务器 - 支持 Bing/Google 图片搜索、验证和下载",
   "main": "src/index.js",
   "type": "module",

package/src/index.js

@@ -26,8 +26,12 @@ import logger from './infrastructure/logger.js';
 import { gracefulShutdown } from './infrastructure/gracefulShutdown.js';
 import { healthChecker } from './infrastructure/healthCheck.js';
 import { metrics } from './infrastructure/metrics.js';
+import { searchCache, validationCache } from './infrastructure/cache.js';
 import config from './config/index.js';
 
+// 请求计数器，用于定期清理
+let requestCount = 0;
+
 // 从 package.json 读取版本号
 const require = createRequire(import.meta.url);
 const packageJson = require('../package.json');
@@ -269,7 +273,16 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
       const totalTime = Date.now() - requestStartTime;
       logger.info(`[MCP] 请求完成: ${totalTime}ms, requestId=${result.requestId}`);
 
-      return {
+      // 每3个请求清理一次缓存，避免内存累积
+      requestCount++;
+      if (requestCount >= 3) {
+        searchCache.clear();
+        validationCache.clear();
+        requestCount = 0;
+        logger.info('[MCP] 缓存已清理');
+      }
+
+      const response = {
         content: [
           {
             type: 'text',
@@ -277,6 +290,8 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
           },
         ],
       };
+      
+      return response;
     } catch (innerError) {
       // 记录失败指标
       metrics.recordSearch(source, false, Date.now() - startTime);

package/src/providers/bingScraper.js

@@ -26,31 +26,14 @@ export class BingScraper extends BaseScraper {
     this.options = options;
     const pageSize = 35;
     
-    // 计算需要的页数，并行获取
-    const pagesNeeded = Math.min(Math.ceil(limit / pageSize), 5); // 最多 5 页
-    const offsets = Array.from({ length: pagesNeeded }, (_, i) => i * pageSize);
-    
-    logger.info(`[Bing] Searching "${keyword}" - ${pagesNeeded} pages in parallel`);
+    // 只获取第一页，避免触发速率限制
+    logger.info(`[Bing] Searching "${keyword}" - single page mode`);
 
     try {
-      // 并行获取所有页面
-      const pagePromises = offsets.map(offset => this._fetchPage(keyword, offset));
-      const results = await Promise.allSettled(pagePromises);
-      
-      // 合并结果
-      const urls = new Set();
-      for (const result of results) {
-        if (result.status === 'fulfilled') {
-          result.value.forEach(url => {
-            if (urls.size < limit) {
-              urls.add(url);
-            }
-          });
-        }
-      }
-
-      logger.info(`[Bing] Complete: ${urls.size} URLs for "${keyword}"`);
-      return Array.from(urls);
+      const urls = await this._fetchPage(keyword, 0);
+      const result = urls.slice(0, limit);
+      logger.info(`[Bing] Complete: ${result.length} URLs for "${keyword}"`);
+      return result;
     } catch (error) {
       logger.error(`Bing search error for "${keyword}"`, { message: error.message });
       return [];

package/src/providers/googleScraper.js

@@ -24,33 +24,15 @@ export class GoogleScraper extends BaseScraper {
    */
   async search(keyword, limit = 10, options = {}) {
     this.options = options;
-    const pageSize = 20;
     
-    // 计算需要的页数，并行获取
-    const pagesNeeded = Math.min(Math.ceil(limit / pageSize), 5); // 最多 5 页
-    const starts = Array.from({ length: pagesNeeded }, (_, i) => i * pageSize);
-    
-    logger.info(`[Google] Searching "${keyword}" - ${pagesNeeded} pages in parallel`);
+    // 只获取第一页，避免触发速率限制
+    logger.info(`[Google] Searching "${keyword}" - single page mode`);
 
     try {
-      // 并行获取所有页面
-      const pagePromises = starts.map(start => this._fetchPage(keyword, start));
-      const results = await Promise.allSettled(pagePromises);
-      
-      // 合并结果
-      const urls = new Set();
-      for (const result of results) {
-        if (result.status === 'fulfilled') {
-          result.value.forEach(url => {
-            if (urls.size < limit) {
-              urls.add(url);
-            }
-          });
-        }
-      }
-
-      logger.info(`[Google] Complete: ${urls.size} URLs for "${keyword}"`);
-      return Array.from(urls);
+      const urls = await this._fetchPage(keyword, 0);
+      const result = urls.slice(0, limit);
+      logger.info(`[Google] Complete: ${result.length} URLs for "${keyword}"`);
+      return result;
     } catch (error) {
       logger.error(`Google search error for "${keyword}"`, { message: error.message });
       return [];