feat: 内容数据迁移至数据库，合规审查全链路打通

- 文章 HTML 存储从文件系统迁移至 articles 表，删除 releases 目录 - 合规审查从 DB 读取 HTML，审查结果写回 DB，通过后自动推进至待发布 - 新增 todayCount 筛选按钮，与系统概览统计数据一致 - 全屏预览修复：提升 z-index 超过侧边栏，添加退出全屏/关闭按钮 - 统一 '优化' → '审查' 命名，消除前后端术语不一致 - 调度器创作完成后自动触发审查（生成 → 审查 → 待发布） - 清理旧备份/调试文件、过期大纲和研究笔记
2026-05-13 17:33:56 +08:00
parent bc6a302e59
commit 233e23016c
234 changed files with 5670 additions and 10651 deletions
@@ -0,0 +1,103 @@
+#!/usr/bin/env python3
+"""
+热点趋势感知模块
+- LLM 生成当前领域热点话题
+- 可扩展接入外部热搜 API
+"""
+
+import json, datetime, logging, sys
+from pathlib import Path
+from typing import List, Dict
+
+PROJECT_ROOT = Path(__file__).parent.parent
+sys.path.insert(0, str(PROJECT_ROOT))
+sys.path.insert(0, str(PROJECT_ROOT / "platform" / "backend"))
+
+from app.core.nvidia_client import call_llm
+
+DATA_DIR = PROJECT_ROOT / "automation" / "data"
+TRENDS_FILE = DATA_DIR / "trends.json"
+LOGS_DIR = PROJECT_ROOT / "automation" / "logs"
+TODAY = datetime.datetime.now().strftime("%Y-%m-%d")
+
+logging.basicConfig(
+    level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=[logging.FileHandler(LOGS_DIR / f"trends_{TODAY}.log"), logging.StreamHandler()]
+)
+logger = logging.getLogger(__name__)
+
+DOMAINS = ["远程工作", "AI工具", "可持续生活", "知识管理", "数字生活", "科技人文"]
+
+def fetch_llm_trends() -> List[Dict]:
+    prompt = f"""你是社交媒体趋势分析师。列出今天（2026年5月）中文互联网上最热的10个话题，要求：
+
+1. 覆盖以下领域：{', '.join(DOMAINS)}
+2. 每个话题包含：领域、话题名称、热度原因（1句话）、相关热搜词（3个）
+3. 优先选择在知乎/微博/小红书上有讨论度的话题
+4. 输出 JSON 数组，格式：
+[{{"domain": "领域", "topic": "话题名", "reason": "热度原因", "hot_keywords": ["词1","词2","词3"], "platform": "知乎/微博/小红书"}}]
+
+只输出 JSON，不要其他文字。"""
+    try:
+        resp = call_llm(prompt, temperature=0.4, max_tokens=2000)
+        resp = resp.strip()
+        if resp.startswith("```"):
+            resp = resp.split("\n", 1)[1].rsplit("\n", 1)[0]
+        trends = json.loads(resp)
+        if isinstance(trends, list):
+            return trends
+    except Exception as e:
+        logger.warning(f"LLM 趋势获取失败: {e}")
+    return []
+
+def save_trends(trends: List[Dict]):
+    data = {
+        "date": TODAY,
+        "updated_at": datetime.datetime.now().isoformat(),
+        "trends": trends
+    }
+    DATA_DIR.mkdir(parents=True, exist_ok=True)
+    TRENDS_FILE.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding='utf-8')
+    logger.info(f"趋势数据已保存: {len(trends)} 条")
+
+def load_trends() -> List[Dict]:
+    if TRENDS_FILE.exists():
+        try:
+            data = json.loads(TRENDS_FILE.read_text(encoding='utf-8'))
+            if data.get("date") == TODAY:
+                return data.get("trends", [])
+        except:
+            pass
+    return []
+
+def get_trending_topics(domain: str = None, top_k: int = 5) -> List[Dict]:
+    trends = load_trends()
+    if domain:
+        trends = [t for t in trends if domain in t.get("domain", "") or t.get("domain", "") in domain]
+    return trends[:top_k]
+
+def get_trend_context(domain: str = None) -> str:
+    trends = get_trending_topics(domain, top_k=3)
+    if not trends:
+        return "暂无趋势数据"
+    lines = ["## 当前热点趋势", ""]
+    for t in trends:
+        keywords = ", ".join(t.get("hot_keywords", []))
+        lines.append(f"- **{t['topic']}**（{t.get('platform','')}）：{t.get('reason','')}")
+        if keywords:
+            lines.append(f"  热搜词：{keywords}")
+    return "\n".join(lines)
+
+def main():
+    logger.info("开始获取热点趋势...")
+    trends = fetch_llm_trends()
+    if trends:
+        save_trends(trends)
+        for t in trends:
+            print(f"  [{t.get('domain','?')}] {t['topic']} — {t.get('platform','')}")
+    else:
+        print("未获取到趋势数据")
+    print(f"完成，共 {len(trends)} 条")
+
+if __name__ == "__main__":
+    main()