修复一个卡死问题

search/server.go

@@ -11,6 +11,7 @@ import (
 	"net/url"     // URL 解析
 	"regexp"       // 正则表达式（site: 过滤语法）
 	"sort"         // 排序
+	"strconv"      // 字符串转整数
 	"strings"       // 字符串操作
 	"sync"         // 互斥锁（保护并发切片写入）
 	"time"         // 时间戳
@@ -46,6 +47,8 @@ func New(db *storage.DB, infoSvc *info.Service, a *analyzer.Analyzer) *Server {
 func (s *Server) Handler() http.Handler {
 	mux := http.NewServeMux()
 	mux.HandleFunc("/search", s.handleSearch)
+	mux.HandleFunc("/admin/recent", s.handleAdminRecent)
+	mux.HandleFunc("/admin/stats", s.handleAdminStats)
 	return mux
 }
 
@@ -55,6 +58,149 @@ func (s *Server) ListenAndServe(addr string) error {
 	return http.ListenAndServe(addr, s.Handler())
 }
 
+// ---- Admin 接口 ----
+
+// recentItem 是 /admin/recent 接口返回的单条记录。
+type recentItem struct {
+	URL         string             `json:"url"`
+	Title       string             `json:"title"`
+	Description string             `json:"description"`
+	Domain      string             `json:"domain"`
+	Language    map[string]float64 `json:"language"`
+	WordCount   int                `json:"word_count"`
+	CrawledAt   int64              `json:"crawled_at"`
+}
+
+// handleAdminRecent 返回最近爬取的条目列表，按爬取时间倒序。
+// 参数：limit（默认50，最大200）。
+func (s *Server) handleAdminRecent(w http.ResponseWriter, r *http.Request) {
+	w.Header().Set("Access-Control-Allow-Origin", "*")
+	w.Header().Set("Content-Type", "application/json; charset=utf-8")
+
+	limit := 50
+	if l := r.URL.Query().Get("limit"); l != "" {
+		if v, err := strconv.Atoi(l); err == nil && v > 0 {
+			limit = v
+		}
+	}
+	if limit > 200 {
+		limit = 200
+	}
+
+	type entry struct {
+		url      string
+		snippet  *storage.SnippetEntry
+		siteInfo *storage.SiteInfo
+	}
+
+	var items []entry
+	s.db.ForEachSnippet(func(url string, snippet *storage.SnippetEntry) error {
+		siteInfo, _ := s.db.GetSiteInfo(netloc(url))
+		items = append(items, entry{url, snippet, siteInfo})
+		return nil
+	})
+
+	// 按时间倒序
+	sort.Slice(items, func(i, j int) bool {
+		return items[i].snippet.Timestamp > items[j].snippet.Timestamp
+	})
+
+	if len(items) > limit {
+		items = items[:limit]
+	}
+
+	result := make([]recentItem, 0, len(items))
+	for _, e := range items {
+		lang := e.siteInfo.Languages
+		if lang == nil {
+			lang = make(map[string]float64)
+		}
+		desc := e.snippet.Description
+		if len(desc) > 200 {
+			desc = desc[:200]
+		}
+		result = append(result, recentItem{
+			URL:         e.url,
+			Title:       e.snippet.Title,
+			Description: desc,
+			Domain:      netloc(e.url),
+			Language:    lang,
+			WordCount:  len(e.snippet.Text),
+			CrawledAt:  e.snippet.Timestamp,
+		})
+	}
+
+	resp := map[string]any{
+		"items": result,
+		"total": len(items),
+	}
+	json.NewEncoder(w).Encode(resp)
+}
+
+// handleAdminStats 返回全局统计：域名分布、语种分布、总 URL 数、总词数。
+func (s *Server) handleAdminStats(w http.ResponseWriter, r *http.Request) {
+	w.Header().Set("Access-Control-Allow-Origin", "*")
+	w.Header().Set("Content-Type", "application/json; charset=utf-8")
+
+	domainCount := make(map[string]int)
+	langCount := make(map[string]int)
+	totalWords := 0
+	total := 0
+
+	s.db.ForEachSnippet(func(url string, snippet *storage.SnippetEntry) error {
+		total++
+		domain := netloc(url)
+		domainCount[domain]++
+		totalWords += len(snippet.Text)
+
+		siteInfo, _ := s.db.GetSiteInfo(domain)
+		if siteInfo != nil {
+			for lang, ratio := range siteInfo.Languages {
+				if ratio > 0.1 {
+					langCount[lang]++
+				}
+			}
+		}
+		return nil
+	})
+
+	// 排序取 Top
+	type kv struct{ k string; v int }
+	topDomains := make([]kv, 0, len(domainCount))
+	for k, v := range domainCount {
+		topDomains = append(topDomains, kv{k, v})
+	}
+	sort.Slice(topDomains, func(i, j int) bool { return topDomains[i].v > topDomains[j].v })
+	if len(topDomains) > 20 {
+		topDomains = topDomains[:20]
+	}
+	topLangs := make([]kv, 0, len(langCount))
+	for k, v := range langCount {
+		topLangs = append(topLangs, kv{k, v})
+	}
+	sort.Slice(topLangs, func(i, j int) bool { return topLangs[i].v > topLangs[j].v })
+	if len(topLangs) > 10 {
+		topLangs = topLangs[:10]
+	}
+
+	domainsMap := make(map[string]int)
+	for _, kv := range topDomains {
+		domainsMap[kv.k] = kv.v
+	}
+	langsMap := make(map[string]int)
+	for _, kv := range topLangs {
+		langsMap[kv.k] = kv.v
+	}
+
+	resp := map[string]any{
+		"total_urls":   total,
+		"total_words":  totalWords,
+		"domains":      domainsMap,
+		"languages":    langsMap,
+	}
+	json.NewEncoder(w).Encode(resp)
+}
+
 // ---- 搜索处理器 ----
 
 // searchResponse 是搜索 API 的 JSON 响应结构。