fix: 修复爬虫问题

.env.example

@@ -1,4 +1,8 @@
 # Mapbox 地图令牌 (波斯湾区域展示)
 # 免费申请: https://account.mapbox.com/access-tokens/
-# 复制本文件为 .env 并填入你的 token
 VITE_MAPBOX_ACCESS_TOKEN=your_mapbox_public_token_here
+
+# 阿里云 DashScope API Key（爬虫 AI 提取用，不设则用规则或 Ollama）
+# 在 crawler 目录或系统环境变量中设置，例如:
+# export DASHSCOPE_API_KEY=sk-xxx
+DASHSCOPE_API_KEY=

Dockerfile.crawler

@@ -10,7 +10,6 @@ COPY crawler ./
 
 ENV DB_PATH=/data/data.db
 ENV API_BASE=http://api:3001
-ENV CLEANER_AI_DISABLED=1
 ENV GDELT_DISABLED=1
 ENV RSS_INTERVAL_SEC=60
 

crawler/config.py

@@ -10,6 +10,9 @@ DB_PATH = os.environ.get("DB_PATH", str(PROJECT_ROOT / "server" / "data.db"))
 # Node API 地址（用于通知推送）
 API_BASE = os.environ.get("API_BASE", "http://localhost:3001")
 
+# 阿里云 DashScope API Key（用于 AI 提取面板数据，不设则回退到规则/Ollama）
+DASHSCOPE_API_KEY = os.environ.get("DASHSCOPE_API_KEY", "")
+
 # 抓取间隔（秒）
 CRAWL_INTERVAL = int(os.environ.get("CRAWL_INTERVAL", "300"))
 

crawler/extractor_dashscope.py

@@ -0,0 +1,121 @@
+# -*- coding: utf-8 -*-
+"""
+阿里云 DashScope（通义千问）提取面板结构化数据
+从新闻文本中提取战损、报复指数、基地状态等，供 db_merge 落库
+API Key 通过环境变量 DASHSCOPE_API_KEY 配置
+"""
+import json
+import os
+import re
+from datetime import datetime, timezone
+from typing import Any, Dict, Optional
+
+from panel_schema import validate_category, validate_severity, validate_summary
+
+
+def _call_dashscope_extract(text: str, timeout: int = 15) -> Optional[Dict[str, Any]]:
+    """调用阿里云 DashScope 提取结构化数据"""
+    api_key = os.environ.get("DASHSCOPE_API_KEY", "").strip()
+    if not api_key or not text or len(str(text).strip()) < 10:
+        return None
+    try:
+        import dashscope
+        from http import HTTPStatus
+
+        dashscope.api_key = api_key
+
+        prompt = f"""从以下美伊/中东军事新闻中提取可明确推断的数值，输出 JSON。无依据的字段省略不写。
+
+要求：
+- summary: 1-2句中文事实摘要，≤80字
+- category: deployment|alert|intel|diplomatic|other
+- severity: low|medium|high|critical
+- 战损（仅当新闻明确提及数字时填写）:
+  us_personnel_killed, iran_personnel_killed, us_personnel_wounded, iran_personnel_wounded,
+  us_civilian_killed, iran_civilian_killed, us_civilian_wounded, iran_civilian_wounded,
+  us_bases_destroyed, iran_bases_destroyed, us_bases_damaged, iran_bases_damaged,
+  us_aircraft, iran_aircraft, us_warships, iran_warships, us_armor, iran_armor, us_vehicles, iran_vehicles
+- retaliation_sentiment: 0-100，仅当新闻涉及伊朗报复/反击情绪时
+- wall_street_value: 0-100，仅当新闻涉及美股/市场反应时
+- key_location_updates: 当新闻提及具体基地遭袭时，数组 [{{"name_keywords":"阿萨德|asad|assad","side":"us","status":"attacked","damage_level":1-3}}]
+
+原文：
+{str(text)[:800]}
+
+直接输出 JSON，不要其他解释："""
+
+        response = dashscope.Generation.call(
+            model="qwen-turbo",
+            messages=[{"role": "user", "content": prompt}],
+            result_format="message",
+            max_tokens=512,
+        )
+
+        if response.status_code != HTTPStatus.OK:
+            return None
+        raw = (response.output.get("choices", [{}])[0].get("message", {}).get("content", "") or "").strip()
+        raw = re.sub(r"^```\w*\s*|\s*```$", "", raw)
+        return json.loads(raw)
+    except Exception:
+        return None
+
+
+def extract_from_news(text: str, timestamp: Optional[str] = None) -> Dict[str, Any]:
+    """
+    从新闻文本提取结构化数据，符合面板 schema
+    返回: { situation_update?, combat_losses_delta?, retaliation?, wall_street?, key_location_updates? }
+    """
+    ts = timestamp or datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%S.000Z")
+    out: Dict[str, Any] = {}
+    parsed = _call_dashscope_extract(text)
+    if not parsed:
+        return out
+
+    if parsed.get("summary"):
+        out["situation_update"] = {
+            "summary": validate_summary(str(parsed["summary"])[:120], 120),
+            "category": validate_category(str(parsed.get("category", "other")).lower()),
+            "severity": validate_severity(str(parsed.get("severity", "medium")).lower()),
+            "timestamp": ts,
+        }
+
+    loss_us = {}
+    loss_ir = {}
+    for k in ["personnel_killed", "personnel_wounded", "civilian_killed", "civilian_wounded",
+              "bases_destroyed", "bases_damaged", "aircraft", "warships", "armor", "vehicles"]:
+        uk, ik = f"us_{k}", f"iran_{k}"
+        if uk in parsed and isinstance(parsed[uk], (int, float)):
+            loss_us[k] = max(0, int(parsed[uk]))
+        if ik in parsed and isinstance(parsed[ik], (int, float)):
+            loss_ir[k] = max(0, int(parsed[ik]))
+    if loss_us or loss_ir:
+        out["combat_losses_delta"] = {}
+        if loss_us:
+            out["combat_losses_delta"]["us"] = loss_us
+        if loss_ir:
+            out["combat_losses_delta"]["iran"] = loss_ir
+
+    if "retaliation_sentiment" in parsed:
+        v = parsed["retaliation_sentiment"]
+        if isinstance(v, (int, float)) and 0 <= v <= 100:
+            out["retaliation"] = {"value": int(v), "time": ts}
+
+    if "wall_street_value" in parsed:
+        v = parsed["wall_street_value"]
+        if isinstance(v, (int, float)) and 0 <= v <= 100:
+            out["wall_street"] = {"time": ts, "value": int(v)}
+
+    if "key_location_updates" in parsed and isinstance(parsed["key_location_updates"], list):
+        valid = []
+        for u in parsed["key_location_updates"]:
+            if isinstance(u, dict) and u.get("name_keywords") and u.get("side") in ("us", "iran"):
+                valid.append({
+                    "name_keywords": str(u["name_keywords"]),
+                    "side": u["side"],
+                    "status": str(u.get("status", "attacked"))[:20],
+                    "damage_level": min(3, max(1, int(u["damage_level"]))) if isinstance(u.get("damage_level"), (int, float)) else 2,
+                })
+        if valid:
+            out["key_location_updates"] = valid
+
+    return out

crawler/news_storage.py

@@ -0,0 +1,141 @@
+# -*- coding: utf-8 -*-
+"""
+资讯内容独立存储，支持历史去重
+爬虫拉回数据 → 计算 content_hash → 若已存在则跳过（去重）→ 新数据落库 news_content
+"""
+import hashlib
+import os
+import re
+import sqlite3
+from datetime import datetime, timezone
+from typing import List, Optional, Tuple
+
+from config import DB_PATH
+
+
+def _to_utc_iso(dt: datetime) -> str:
+    if dt.tzinfo:
+        dt = dt.astimezone(timezone.utc)
+    return dt.strftime("%Y-%m-%dT%H:%M:%S.000Z")
+
+
+def _normalize_for_hash(text: str) -> str:
+    """归一化文本用于生成去重 hash"""
+    if not text:
+        return ""
+    t = re.sub(r"\s+", " ", str(text).strip().lower())[:600]
+    return re.sub(r"[\x00-\x1f]", "", t)
+
+
+def content_hash(title: str, summary: str, url: str) -> str:
+    """根据标题、摘要、URL 生成去重 hash，相似内容视为重复"""
+    raw = _normalize_for_hash(title) + "|" + _normalize_for_hash(summary) + "|" + (url or "").strip()
+    return hashlib.sha256(raw.encode("utf-8")).hexdigest()[:32]
+
+
+def _ensure_table(conn: sqlite3.Connection) -> None:
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS news_content (
+            id TEXT PRIMARY KEY,
+            content_hash TEXT NOT NULL UNIQUE,
+            title TEXT NOT NULL,
+            summary TEXT NOT NULL,
+            url TEXT NOT NULL DEFAULT '',
+            source TEXT NOT NULL DEFAULT '',
+            published_at TEXT NOT NULL,
+            category TEXT NOT NULL DEFAULT 'other',
+            severity TEXT NOT NULL DEFAULT 'medium',
+            created_at TEXT NOT NULL DEFAULT (datetime('now'))
+        )
+    """)
+    try:
+        conn.execute("CREATE UNIQUE INDEX IF NOT EXISTS idx_news_content_hash ON news_content(content_hash)")
+    except sqlite3.OperationalError:
+        pass
+    try:
+        conn.execute("CREATE INDEX IF NOT EXISTS idx_news_content_pub ON news_content(published_at DESC)")
+    except sqlite3.OperationalError:
+        pass
+    conn.commit()
+
+
+def exists_by_hash(conn: sqlite3.Connection, h: str) -> bool:
+    row = conn.execute("SELECT 1 FROM news_content WHERE content_hash = ? LIMIT 1", (h,)).fetchone()
+    return row is not None
+
+
+def insert_news(
+    conn: sqlite3.Connection,
+    *,
+    title: str,
+    summary: str,
+    url: str = "",
+    source: str = "",
+    published: datetime,
+    category: str = "other",
+    severity: str = "medium",
+) -> Optional[str]:
+    """
+    插入资讯，若 content_hash 已存在则跳过（去重）
+    返回: 新插入的 id，或 None 表示重复跳过
+    """
+    _ensure_table(conn)
+    h = content_hash(title, summary, url)
+    if exists_by_hash(conn, h):
+        return None
+    uid = "nc_" + hashlib.sha256(f"{h}{datetime.utcnow().isoformat()}".encode()).hexdigest()[:14]
+    ts = _to_utc_iso(published)
+    conn.execute(
+        """INSERT INTO news_content (id, content_hash, title, summary, url, source, published_at, category, severity)
+           VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)""",
+        (uid, h, (title or "")[:500], (summary or "")[:2000], (url or "")[:500], (source or "")[:100], ts, category, severity),
+    )
+    conn.commit()
+    return uid
+
+
+def save_and_dedup(items: List[dict], db_path: Optional[str] = None) -> Tuple[List[dict], int]:
+    """
+    去重后落库 news_content
+    items: [{"title","summary","url","published","category","severity","source"?}, ...]
+    返回: (通过去重的新项列表, 实际新增条数)
+    """
+    path = db_path or DB_PATH
+    if not os.path.exists(path):
+        return [], 0
+    conn = sqlite3.connect(path, timeout=10)
+    try:
+        _ensure_table(conn)
+        new_items: List[dict] = []
+        count = 0
+        for u in items:
+            title = (u.get("title") or "")[:500]
+            summary = (u.get("summary") or u.get("title") or "")[:2000]
+            url = (u.get("url") or "")[:500]
+            source = (u.get("source") or "")[:100]
+            pub = u.get("published")
+            if isinstance(pub, str):
+                try:
+                    pub = datetime.fromisoformat(pub.replace("Z", "+00:00"))
+                except ValueError:
+                    pub = datetime.now(timezone.utc)
+            elif pub is None:
+                pub = datetime.now(timezone.utc)
+            cat = u.get("category", "other")
+            sev = u.get("severity", "medium")
+            uid = insert_news(
+                conn,
+                title=title,
+                summary=summary,
+                url=url,
+                source=source,
+                published=pub,
+                category=cat,
+                severity=sev,
+            )
+            if uid:
+                count += 1
+                new_items.append({**u, "news_id": uid})
+        return new_items, count
+    finally:
+        conn.close()

crawler/realtime_conflict_service.py

@@ -283,7 +283,7 @@ def _rss_to_gdelt_fallback() -> None:
 
 
 # ==========================
-# RSS 新闻抓取（补充 situation_update + AI 提取面板数据）
+# RSS 新闻抓取：资讯落库(去重) → AI 提取 → 面板数据落库 → 通知前端
 # ==========================
 LAST_FETCH = {"items": 0, "inserted": 0, "error": None}
 
@@ -292,6 +292,7 @@ def fetch_news() -> None:
     try:
         from scrapers.rss_scraper import fetch_all
         from db_writer import write_updates
+        from news_storage import save_and_dedup
         from translate_utils import translate_to_chinese
         from cleaner_ai import clean_news_for_panel
         from cleaner_ai import ensure_category, ensure_severity
@@ -304,36 +305,44 @@ def fetch_news() -> None:
             it["summary"] = clean_news_for_panel(raw_summary or raw_title, max_len=120)
             it["category"] = ensure_category(it.get("category", "other"))
             it["severity"] = ensure_severity(it.get("severity", "medium"))
-        n = write_updates(items) if items else 0
+            it["source"] = it.get("source") or "rss"
+        # 1. 历史去重：资讯内容落库 news_content（独立表，便于后续消费）
+        new_items, n_news = save_and_dedup(items, db_path=DB_PATH)
+        # 2. 面板展示：新增资讯写入 situation_update（供前端 recentUpdates）
+        n_panel = write_updates(new_items) if new_items else 0
         LAST_FETCH["items"] = len(items)
-        LAST_FETCH["inserted"] = n
-        if items:
-            _extract_and_merge_panel_data(items)
+        LAST_FETCH["inserted"] = n_news
+        # 3. AI 提取 + 合并到 combat_losses / key_location 等
+        if new_items:
+            _extract_and_merge_panel_data(new_items)
         # GDELT 禁用时用 RSS 填充 gdelt_events，使地图有冲突点
         if GDELT_DISABLED:
             _rss_to_gdelt_fallback()
-        # 每次抓取完成都通知 Node 更新时间戳，便于「实时更新」显示
         _notify_node()
-        print(f"[{datetime.now().strftime('%H:%M:%S')}] RSS 抓取 {len(items)} 条，新增入库 {n} 条")
+        print(f"[{datetime.now().strftime('%H:%M:%S')}] RSS 抓取 {len(items)} 条，去重后新增 {n_news} 条资讯，面板 {n_panel} 条")
     except Exception as e:
         LAST_FETCH["error"] = str(e)
         print(f"[{datetime.now().strftime('%H:%M:%S')}] 新闻抓取失败: {e}")
 
 
 def _extract_and_merge_panel_data(items: list) -> None:
-    """对新闻做 AI/规则 提取，合并到 combat_losses / retaliation / wall_street_trend 等表"""
+    """AI 分析提取面板相关数据，清洗后落库"""
     if not items or not os.path.exists(DB_PATH):
         return
     try:
         from db_merge import merge
-        if os.environ.get("CLEANER_AI_DISABLED", "0") == "1":
+        use_dashscope = bool(os.environ.get("DASHSCOPE_API_KEY", "").strip())
+        if use_dashscope:
+            from extractor_dashscope import extract_from_news
+            limit = 10
+        elif os.environ.get("CLEANER_AI_DISABLED", "0") == "1":
             from extractor_rules import extract_from_news
+            limit = 25
         else:
             from extractor_ai import extract_from_news
+            limit = 10
         from datetime import timezone
         merged_any = False
-        # 规则模式可多处理几条（无 Ollama）；AI 模式限制 5 条避免调用过多
-        limit = 25 if os.environ.get("CLEANER_AI_DISABLED", "0") == "1" else 10
         for it in items[:limit]:
             text = (it.get("title", "") or "") + " " + (it.get("summary", "") or "")
             if len(text.strip()) < 20:

crawler/requirements.txt

@@ -3,3 +3,4 @@ feedparser>=6.0.0
 fastapi>=0.109.0
 uvicorn>=0.27.0
 deep-translator>=1.11.0
+dashscope>=1.20.0

docker-compose.yml

@@ -22,9 +22,9 @@ services:
     environment:
       - DB_PATH=/data/data.db
       - API_BASE=http://api:3001
-      - CLEANER_AI_DISABLED=1
       - GDELT_DISABLED=1
       - RSS_INTERVAL_SEC=60
+      - DASHSCOPE_API_KEY=${DASHSCOPE_API_KEY:-}
     volumes:
       - app-data:/data
     depends_on:

docs/CRAWLER_PIPELINE.md

@@ -0,0 +1,65 @@
+# 爬虫数据流水线
+
+## 数据流
+
+```
+RSS 抓取
+    ↓ 翻译、清洗
+    ↓ news_storage.save_and_dedup() → 历史去重
+    ↓
+news_content（资讯独立表，供后续消费）
+    ↓
+    ↓ 去重后的新数据
+    ↓
+situation_update（面板展示用）
+    ↓
+    ↓ AI 提取（阿里云 DashScope）
+    ↓
+combat_losses / retaliation / key_location / wall_street_trend
+    ↓
+    ↓ notify Node
+    ↓
+前端 WebSocket + 轮询
+```
+
+## 阿里云 DashScope API Key
+
+设置环境变量 `DASHSCOPE_API_KEY` 后，爬虫使用阿里云通义千问进行 AI 提取。不设置时回退到规则提取（`extractor_rules`）或 Ollama（若可用）。
+
+```bash
+# 本地
+export DASHSCOPE_API_KEY=sk-xxx
+
+# Docker
+docker compose up -d -e DASHSCOPE_API_KEY=sk-xxx
+# 或在 .env 中写入 DASHSCOPE_API_KEY=sk-xxx
+```
+
+## 表说明
+
+| 表 | 用途 |
+|----|------|
+| `news_content` | 资讯原文，独立存储，支持去重（content_hash），供后续消费 |
+| `situation_update` | 面板「近期更新」展示 |
+| `combat_losses` | 战损数据（AI/规则提取） |
+| `key_location` | 基地状态 |
+| `gdelt_events` | 地图冲突点 |
+
+## 去重逻辑
+
+根据 `content_hash = sha256(normalize(title) + normalize(summary) + url)` 判断，相同或高度相似内容视为重复，不入库。
+
+## 消费资讯
+
+- HTTP: `GET /api/news?limit=50`
+- 调试: `/db` 面板查看 `news_content` 表
+
+## 链路验证
+
+运行脚本一键检查全链路：
+
+```bash
+./scripts/verify-pipeline.sh
+```
+
+支持环境变量覆盖：`API_URL`、`CRAWLER_URL`

package.json

@@ -14,7 +14,9 @@
     "build": "vite build",
     "typecheck": "tsc --noEmit",
     "lint": "eslint .",
-    "preview": "vite preview"
+    "preview": "vite preview",
+    "verify": "./scripts/verify-pipeline.sh",
+    "verify:full": "./scripts/verify-pipeline.sh --start-crawler"
   },
   "dependencies": {
     "better-sqlite3": "^11.6.0",

scripts/verify-pipeline.sh

@@ -0,0 +1,124 @@
+#!/usr/bin/env bash
+# 验证爬虫 → 数据库 → API → 前端 全链路
+# 用法: ./scripts/verify-pipeline.sh [--start-crawler]
+set -e
+
+API_URL="${API_URL:-http://localhost:3001}"
+CRAWLER_URL="${CRAWLER_URL:-http://localhost:8000}"
+START_CRAWLER=false
+[[ "${1:-}" = "--start-crawler" ]] && START_CRAWLER=true
+
+PROJECT_ROOT="$(cd "$(dirname "$0")/.." && pwd)"
+
+echo "=========================================="
+echo "US-Iran 态势面板 链路验证"
+echo "API: $API_URL | Crawler: $CRAWLER_URL"
+echo "=========================================="
+echo ""
+
+# 可选：启动爬虫
+if $START_CRAWLER; then
+  echo "[0/6] 启动爬虫..."
+  if curl -sf "$CRAWLER_URL/crawler/status" >/dev/null 2>&1; then
+    echo "  ✓ 爬虫已在运行"
+  else
+    cd "$PROJECT_ROOT/crawler"
+    python3 -c "import uvicorn" 2>/dev/null || { echo "  需安装: pip install uvicorn"; exit 1; }
+    uvicorn realtime_conflict_service:app --host 127.0.0.1 --port 8000 &
+    echo "  等待爬虫就绪..."
+    for i in $(seq 1 15); do
+      sleep 2
+      if curl -sf "$CRAWLER_URL/crawler/status" >/dev/null 2>&1; then
+        echo "  ✓ 爬虫已启动"
+        echo "  等待首次 RSS 抓取（约 70 秒）..."
+        sleep 70
+        break
+      fi
+    done
+    if ! curl -sf "$CRAWLER_URL/crawler/status" >/dev/null 2>&1; then
+      echo "  ✗ 爬虫启动超时"
+      exit 1
+    fi
+  fi
+  echo ""
+fi
+
+# 1. API 健康检查
+echo "[1/6] API 健康检查..."
+if curl -sf "$API_URL/api/health" > /dev/null; then
+  echo "  ✓ API 正常"
+else
+  echo "  ✗ API 无响应，请先运行: npm run api"
+  exit 1
+fi
+
+# 2. 态势数据
+echo "[2/6] 态势数据..."
+SIT=$(curl -sf "$API_URL/api/situation" 2>/dev/null || echo "{}")
+if echo "$SIT" | grep -q "lastUpdated"; then
+  echo "  ✓ 态势数据可读"
+  LAST=$(echo "$SIT" | grep -o '"lastUpdated":"[^"]*"' | head -1)
+  echo "    $LAST"
+else
+  echo "  ✗ 态势数据异常"
+  exit 1
+fi
+
+# 3. 爬虫状态
+echo "[3/6] 爬虫状态..."
+CRAWLER=$(curl -sf "$CRAWLER_URL/crawler/status" 2>/dev/null || echo "{}")
+if echo "$CRAWLER" | grep -q "db_path\|db_exists"; then
+  echo "  ✓ 爬虫服务可访问"
+  if command -v jq &>/dev/null; then
+    CNT=$(echo "$CRAWLER" | jq -r '.situation_update_count // "?"')
+    echo "    situation_update 条数: $CNT"
+  fi
+else
+  echo "  ⚠ 爬虫未启动或不可达（可选，需单独运行爬虫）"
+fi
+
+# 4. 资讯表
+echo "[4/6] 资讯表 news_content..."
+NEWS=$(curl -sf "$API_URL/api/news?limit=3" 2>/dev/null || echo '{"items":[]}')
+if echo "$NEWS" | grep -q '"items"'; then
+  if command -v jq &>/dev/null; then
+    N=$(echo "$NEWS" | jq '.items | length')
+    echo "  ✓ 最近 $N 条资讯"
+  else
+    echo "  ✓ 资讯接口可读"
+  fi
+else
+  echo "  ⚠ news_content 可能为空（爬虫未跑或刚启动）"
+fi
+
+# 5. 战损数据
+echo "[5/6] 战损数据 combat_losses..."
+if echo "$SIT" | grep -q "personnelCasualties"; then
+  echo "  ✓ 战损字段存在"
+  if command -v jq &>/dev/null; then
+    US_K=$(echo "$SIT" | jq -r '.usForces.combatLosses.personnelCasualties.killed // "?"')
+    IR_K=$(echo "$SIT" | jq -r '.iranForces.combatLosses.personnelCasualties.killed // "?"')
+    echo "    美军阵亡: $US_K | 伊朗阵亡: $IR_K"
+  fi
+else
+  echo "  ✗ 战损结构异常"
+fi
+
+# 6. 通知接口（仅验证可调用）
+echo "[6/6] 通知接口 POST /api/crawler/notify..."
+NOTIFY=$(curl -sf -X POST "$API_URL/api/crawler/notify" 2>/dev/null || echo "{}")
+if echo "$NOTIFY" | grep -q '"ok"'; then
+  echo "  ✓ 通知接口正常"
+else
+  echo "  ⚠ 通知接口可能异常"
+fi
+
+echo ""
+echo "=========================================="
+echo "验证完成。"
+echo ""
+echo "建议："
+echo "  - 访问 $API_URL/db 查看各表数据"
+echo "  - 爬虫未启动时: ./scripts/verify-pipeline.sh --start-crawler"
+echo "  - 或手动启动: cd crawler && uvicorn realtime_conflict_service:app --port 8000"
+echo "=========================================="

server/db.js

@@ -112,6 +112,21 @@ db.exec(`
     estimated_strike_count INTEGER NOT NULL DEFAULT 0,
     updated_at TEXT NOT NULL
   );
+
+  CREATE TABLE IF NOT EXISTS news_content (
+    id TEXT PRIMARY KEY,
+    content_hash TEXT NOT NULL UNIQUE,
+    title TEXT NOT NULL,
+    summary TEXT NOT NULL,
+    url TEXT NOT NULL DEFAULT '',
+    source TEXT NOT NULL DEFAULT '',
+    published_at TEXT NOT NULL,
+    category TEXT NOT NULL DEFAULT 'other',
+    severity TEXT NOT NULL DEFAULT 'medium',
+    created_at TEXT NOT NULL DEFAULT (datetime('now'))
+  );
+  CREATE INDEX IF NOT EXISTS idx_news_content_hash ON news_content(content_hash);
+  CREATE INDEX IF NOT EXISTS idx_news_content_published ON news_content(published_at DESC);
 `)
 
 // 迁移：为已有 key_location 表添加 type、region、status、damage_level 列

server/openapi.js

@@ -25,6 +25,15 @@ module.exports = {
         },
       },
     },
+    '/api/news': {
+      get: {
+        summary: '资讯内容',
+        description: '从 news_content 表读取，支持 ?limit=50 分页',
+        tags: ['资讯'],
+        parameters: [{ in: 'query', name: 'limit', schema: { type: 'integer', default: 50 } }],
+        responses: { 200: { description: 'items 数组' } },
+      },
+    },
     '/api/db/dashboard': {
       get: {
         summary: '数据库面板',
@@ -130,5 +139,5 @@ module.exports = {
       },
     },
   },
-  tags: [{ name: '态势' }, { name: '统计' }, { name: '反馈' }, { name: '调试' }, { name: '系统' }],
+  tags: [{ name: '态势' }, { name: '资讯' }, { name: '统计' }, { name: '反馈' }, { name: '调试' }, { name: '系统' }],
 }

server/routes.js

@@ -19,6 +19,7 @@ router.get('/db/dashboard', (req, res) => {
       'retaliation_current',
       'retaliation_history',
       'situation_update',
+      'news_content',
       'gdelt_events',
       'conflict_stats',
     ]
@@ -27,6 +28,7 @@ router.get('/db/dashboard', (req, res) => {
       feedback: 'created_at DESC',
       situation: 'updated_at DESC',
       situation_update: 'timestamp DESC',
+      news_content: 'published_at DESC',
       gdelt_events: 'event_time DESC',
       wall_street_trend: 'time DESC',
       retaliation_history: 'time DESC',
@@ -55,6 +57,17 @@ router.get('/db/dashboard', (req, res) => {
   }
 })
 
+// 资讯内容（独立表，供后续消费）
+router.get('/news', (req, res) => {
+  try {
+    const limit = Math.min(parseInt(req.query.limit, 10) || 50, 200)
+    const rows = db.prepare('SELECT id, title, summary, url, source, published_at, category, severity, created_at FROM news_content ORDER BY published_at DESC LIMIT ?').all(limit)
+    res.json({ items: rows })
+  } catch (err) {
+    res.status(500).json({ error: err.message })
+  }
+})
+
 router.get('/situation', (req, res) => {
   try {
     res.json(getSituation())

src/config.ts

@@ -8,3 +8,4 @@ export const config = {
   /** 是否显示滚动情报 */
   showNewsTicker: false,
 }
+