Merge pull request #3 from stardrophere/backend_optimize

Backend optimize

.dockerignore

@@ -0,0 +1,18 @@
+# 前端
+frontend/dist
+frontend/node_modules
+
+# 后端
+backend/.venv
+backend/.git
+backend/__pycache__
+backend/*.pyc
+backend/*.pyo
+backend/*.pyd
+backend/.pytest_cache
+backend/.mypy_cache
+backend/.cache
+backend/.env
+backend/*.log
+backend/dist
+backend/build

.github/workflows/docker-image.yml

@@ -1,35 +0,0 @@
-name: Docker Image CI
-
-on:
-  push:
-    tags: ['v*']
-
-jobs:
-  build:
-    runs-on: ubuntu-latest
-
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Logout Docker (avoid wrong credentials)
-        run: docker logout || true
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
-
-      - name: Install the latest version of uv
-        uses: astral-sh/setup-uv@v7
-
-      - name: uv lock
-        working-directory: backend
-        run: uv lock
-      
-      - name: Build Docker Image (with BuildKit)
-        working-directory: backend
-        run: |
-          docker build \
-            --progress=plain \
-            -t insightradar-backend:${{ github.ref_name }} \
-            -t insightradar-backend:latest \
-            .

.gitignore

@@ -37,13 +37,12 @@ MANIFEST
 *.manifest
 *.spec
 
-# uv
-*.lock
-
 # Installer logs
 pip-log.txt
 pip-delete-this-directory.txt
 
+**/logs/*
+
 # Unit test / coverage reports
 htmlcov/
 .tox/
@@ -190,3 +189,6 @@ cython_debug/
 
 **/data/*
 **/docker/*
+backend/app/static/*
+
+test*.*

backend/.dockerignore

@@ -1,13 +0,0 @@
-.venv
-.git
-__pycache__
-*.pyc
-*.pyo
-*.pyd
-.pytest_cache
-.mypy_cache
-.cache
-.env
-*.log
-dist
-build

backend/app/api/endpoints/auth.py

@@ -69,7 +69,7 @@ def _normalize_email(email: str) -> str:
 def _build_verification_email(code: str, purpose_text: str, expire_minutes: int) -> str:
     return f"""
     <div style="font-family: Arial, sans-serif; line-height: 1.6; color: #222;">
-        <h2 style="margin-bottom: 12px;">InsightRadar 邮箱验证</h2>
+        <h2 style="margin-bottom: 12px;">聚势智见邮箱验证</h2>
         <p>您的{purpose_text}验证码是：</p>
         <p style="font-size: 28px; font-weight: bold; letter-spacing: 4px; color: #0b57d0;">{code}</p>
         <p>该验证码在 {expire_minutes} 分钟内有效。请勿泄露给他人。</p>
@@ -203,7 +203,7 @@ async def send_register_code(
 
         await send_html_email(
             to_email=email,
-            subject=f"【{code}】InsightRadar 注册验证码",
+            subject=f"【{code}】聚势智见 注册验证码",
             html_content=_build_verification_email(
                 code, "注册", REGISTER_CODE_EXPIRE_MINUTES
             ),
@@ -241,7 +241,7 @@ async def send_login_code(
 
         await send_html_email(
             to_email=email,
-            subject=f"【{code}】InsightRadar 登录验证码",
+            subject=f"【{code}】聚势智见 登录验证码",
             html_content=_build_verification_email(
                 code, "登录", LOGIN_CODE_EXPIRE_MINUTES
             ),

backend/app/crud/crud_source.py

@@ -2,6 +2,8 @@
 """
 信息源 CRUD：对 InfoSource 的增删改查，供 API 与爬虫使用
 """
+from sqlite3 import IntegrityError
+
 from sqlalchemy.orm import Session
 from typing import List, Optional
 
@@ -22,10 +24,20 @@ def get_multi(db: Session, skip: int = 0, limit: int = 100) -> List[InfoSource]:
 def create(db: Session, obj_in: InfoSourceCreate) -> InfoSource:
     """创建新的信息源"""
     db_obj = InfoSource(**obj_in.model_dump())
+    exits =db.query(InfoSource).filter(InfoSource.source_name == db_obj.source_name).first()
+    if exits:
+        db.close()
+        return db_obj
+    try:
         db.add(db_obj)
         db.commit()
         db.refresh(db_obj)
         return db_obj
+    except IntegrityError:
+        db.rollback()
+    finally:
+        db.close()
+        return db_obj
 
 
 def update(db: Session, db_obj: InfoSource, obj_in: InfoSourceUpdate) -> InfoSource:

backend/app/initialize.py

@@ -1,17 +1,15 @@
-import requests
 import json
 
-# 请将此处的 URL 替换为您实际的 API 基础域名
+from app.database import SessionLocal
-api_url = "http://10.252.130.135:8000/api/v1/sources/"
+from app.crud.crud_source import create
+from app.models.models import SourceType
+from app.schemas.source_schema import InfoSourceCreate
 
-# 请求头
-headers = {
-    "Content-Type": "application/json",
-    # "Authorization": "Bearer YOUR_TOKEN" # 如果接口需要鉴权，请取消注释并填入 Token
-}
 
-# 解析后的数据源列表
+def init():
-sources_data = [
+
+    # 解析后的数据源列表
+    sources_data = [
         {"name": "今日头条", "url": "toutiao"},
         {"name": "百度热搜", "url": "baidu"},
         {"name": "华尔街见闻", "url": "wallstreetcn-hot"},
@@ -23,24 +21,22 @@ sources_data = [
         {"name": "微博", "url": "weibo"},
         {"name": "抖音", "url": "douyin"},
         {"name": "知乎", "url": "zhihu"}
-]
+    ]
 
-# 遍历数据并发送 POST 请求
+    # 遍历数据并发送 POST 请求
-for item in sources_data:
+    for item in sources_data:
-    payload = {
-        "source_name": item["name"],
-        "source_type": "HOT_TREND",
-        "home_url": item["url"],
-        "is_enabled": True
-    }
           
         try:
-        response = requests.post(api_url, headers=headers, data=json.dumps(payload))
+            
-        if response.status_code in (200, 201):
+            with SessionLocal() as db:
-            print(f"✅ 成功创建: {item['name']}")
+
-        else:
+                create(db, InfoSourceCreate(
-            print(f"❌ 创建失败: {item['name']} - 状态码: {response.status_code} - 详情: {response.text}")
+                    source_name=item["name"],
+                    source_type=SourceType.HOT_TREND,
+                    home_url=item["url"],
+                    is_enabled=True
+                ))
+                print(f"创建订阅源{item['name']}")
+
         except Exception as e:
             print(f"⚠️ 请求异常: {item['name']} - 错误: {e}")
-
-print("执行完毕！")

backend/app/main.py

@@ -1,8 +1,11 @@
 # app/main.py
 import logging
 import os
+from pathlib import Path
+from fastapi.responses import FileResponse, HTMLResponse, JSONResponse
+import httpx
 from contextlib import asynccontextmanager
-from fastapi import FastAPI
+from fastapi import FastAPI, HTTPException, Request, staticfiles
 from fastapi.middleware.cors import CORSMiddleware
 from dotenv import load_dotenv
 
@@ -21,11 +24,11 @@ from app.services.summary_service import generate_unified_summaries
 from app.services.delivery_service import check_and_deliver
 from app.database import engine
 from app.models.models import Base
+from app.initialize import init
 
 # 路由总线
 from app.api.router import api_router
 
-load_dotenv()
 CRAWL_INTERVAL = int(os.getenv("CRAWL_INTERVAL_MINUTES", 10))
 SUMMARY_INTERVAL = int(os.getenv("SUMMARY_INTERVAL_MINUTES", 30))
 
@@ -42,6 +45,10 @@ async def lifespan(app: FastAPI):
     Base.metadata.create_all(bind=engine)
     logging.info("数据库表初始化完成！")
     
+    logging.info("初始化订阅源")
+    init()
+    logging.info("订阅源初始化完毕")
+    
     # 2. 配置并启动定时任务
     scheduler.add_job(
         fetch_and_save_trending_data,
@@ -106,6 +113,24 @@ app.add_middleware(
 # 版本控制
 app.include_router(api_router, prefix="/api/v1")
 
+# 只需要保留API的优先匹配，catch_all可以简化成这样
+@app.get("/api/{full_path:path}")
+async def api_not_found(full_path: str):
+    return {"detail": "API Not Found"}
+
+staticPath = staticfiles.StaticFiles(directory="app/static", html=True)
+
+# 把目录改成static对应我们放dist内容的路径就可以
+app.mount("/", staticPath, name="static")
+
+INDEX_HTML = Path("app/static/index.html").read_text(encoding="utf-8")
+
+@app.exception_handler(404)
+async def not_found_handler(request: Request, exc: HTTPException):
+    # 如果是API路径才返回404，前端路径走catch-all不会进这里
+    if request.url.path.startswith("/api/"):
+        return JSONResponse({"detail": "Not Found"}, status_code=404)
+    return HTMLResponse(INDEX_HTML)
 
 # 健康检查
 @app.get("/", tags=["健康检查"])

backend/app/models/models.py

@@ -94,6 +94,10 @@ class InfoSource(Base):
     created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
     updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow, onupdate=utcnow)
     
+    __table_args__ = (
+        UniqueConstraint("source_name", name="uix_source_name"),
+    )
+
 
 # ==========================================
 # 模块二：AI 语义聚类中枢 (大事件池)
@@ -176,8 +180,7 @@ class NewsArticle(Base):
 
     id: Mapped[int] = mapped_column(BigIntType, primary_key=True, autoincrement=True)
     source_id: Mapped[int] = mapped_column(ForeignKey("info_sources.id"), comment="所属信息源ID")
-    unified_event_id: Mapped[Optional[int]] = mapped_column(ForeignKey("unified_events.id"),
+    unified_event_id: Mapped[Optional[int]] = mapped_column(ForeignKey("unified_events.id"), comment="深度文章也可归入大事件分析")
-                                                            comment="深度文章也可归入大事件分析")
 
     external_id: Mapped[str] = mapped_column(String(32), comment="RSS原文<guid>生成的MD5防重指纹")
     title_embedding: Mapped[Optional[str]] = mapped_column(Text, comment="新闻标题/摘要的语义向量")
@@ -214,8 +217,7 @@ class HeadlineRevision(Base):
     previous_headline: Mapped[str] = mapped_column(String(255), comment="修改前的旧标题")
     revised_headline: Mapped[str] = mapped_column(String(255), comment="修改后的新标题")
 
-    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow,
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow, comment="系统发现被修改的时间")
-                                                 comment="系统发现被修改的时间")
 
 
 class RankingLog(Base):
@@ -235,8 +237,7 @@ class RankingLog(Base):
     # 当时它在第几名
     ranking_position: Mapped[int] = mapped_column(Integer, comment="当时抓取时的排名名次")
     # 爬虫看到它的那一瞬间的时间
-    observed_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow,
+    observed_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow, comment="观察到该名次的准确时间")
-                                                  comment="观察到该名次的准确时间")
 
     created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
 
@@ -307,13 +308,11 @@ class AppUser(Base):
 
     nickname: Mapped[Optional[str]] = mapped_column(String(100), comment="用户展示昵称")
     avatar_url: Mapped[Optional[str]] = mapped_column(String(500), comment="用户头像地址")
-    gender: Mapped[GenderType] = mapped_column(Enum(GenderType), default=GenderType.UNKNOWN,
+    gender: Mapped[GenderType] = mapped_column(Enum(GenderType), default=GenderType.UNKNOWN, comment="用户性别(用于AI调整行文语气)")
-                                               comment="用户性别(用于AI调整行文语气)")
 
     # 极其强大：一个万能收纳箱！前端未来想加任何诸如“夜间模式”、“字体变大”的开关，
     # 全部丢进这个 JSON 字段即可，从此免去手动修改后端表结构的麻烦。
-    metadata_: Mapped[Optional[Any]] = mapped_column("metadata", JSON,
+    metadata_: Mapped[Optional[Any]] = mapped_column("metadata", JSON, comment="JSON扩展字段: 存放灵活多变的前端用户偏好设置")
-                                                     comment="JSON扩展字段: 存放灵活多变的前端用户偏好设置")
 
     # 时区对于定时推送系统极其重要！保证纽约的用户和北京的用户都能在早晨8点收到新闻。
     timezone: Mapped[str] = mapped_column(String(50), default="Asia/Shanghai", comment="用户所在地时区")
@@ -401,8 +400,7 @@ class DeliveryHistory(Base):
     # 记录这次推送是彻底成功了，还是由于渠道网络问题失败了
     status: Mapped[TaskStatus] = mapped_column(Enum(TaskStatus), comment="最终推送结果状态")
 
-    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow,
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow, comment="记录或实际推送的准确时间")
-                                                 comment="记录或实际推送的准确时间")
 
 
 # ==========================================

backend/app/prompts/digest_email_template.py

@@ -86,7 +86,7 @@ body{{margin:0;padding:0;background:#0d1117;color:#e6edf3;font-family:-apple-sys
 <body>
 <div class="container">
   <div class="header">
-    <h1>InsightRadar · 热点快报</h1>
+    <h1>聚势智见 · 热点快报</h1>
     <p>{delivery_time} · 为你精选了 {event_count} 条事件</p>
     <span class="mode-badge {mode_badge_class}">{mode_label}</span>
   </div>
@@ -94,8 +94,8 @@ body{{margin:0;padding:0;background:#0d1117;color:#e6edf3;font-family:-apple-sys
   {event_cards_html}
 
   <div class="footer">
-    <p>此邮件由 InsightRadar 自动推送。</p>
+    <p>此邮件由 聚势智见自动推送。</p>
-    <p>如需调整推送设置，请登录 <a href="{app_url}">InsightRadar 控制台</a></p>
+    <p>如需调整推送设置，请登录 <a href="{app_url}">聚势智见 控制台</a></p>
   </div>
 </div>
 </body>

backend/app/services/delivery_service.py

@@ -377,7 +377,7 @@ def _prepare_user_push(db: Session, user: AppUser, schedule: UserDeliverySchedul
     return _PendingPush(
         user_id=user_id,
         email_targets=[ep.channel_account for ep in email_endpoints],
-        subject=f"InsightRadar {subject_suffix} · {time_str}",
+        subject=f"聚势智见 {subject_suffix} · {time_str}",
         html_body=html_body,
         event_ids=event_ids,
     )

backend/app/services/fetcher_service.py

@@ -26,7 +26,7 @@ SIMILARITY_THRESHOLD = float(os.getenv("SIMILARITY_THRESHOLD", 0.72))
 API_BASE_URL = os.getenv("API_BASE_URL", "https://newsnow.busiyi.world/api/s")
 EMBEDDING_MODEL_PATH = os.getenv("EMBEDDING_MODEL_PATH", "")
 
-print("正在加载 BAAI/bge-m3 向量模型...")
+print("正在加载模型...")
 # 全局单例
 embedder_model = SentenceTransformer(EMBEDDING_MODEL_PATH, local_files_only=True)
 print("模型加载完成。")

backend/app/services/matching_service.py

@@ -1,6 +1,6 @@
 """
 匹配服务：根据用户兴趣关键词（精确 + 语义）推荐事件
-打分融合：匹配分 + 标签相关度 + 热度 + 新鲜度加成
+打分融合：标签/标题匹配分 + 标签相关度 + 热度 + 新鲜度加成
 """
 import os
 from dataclasses import dataclass
@@ -14,7 +14,7 @@ from app.models.models import ExtractedTopic, TargetType, UnifiedEvent, UserTopi
 from app.services.fetcher_service import embedder_model
 
 
-# 语义匹配阈值：用户关键词和事件标签向量相似度达到该值才计入语义命中
+# 语义匹配阈值：用户关键词和事件标签/标题向量相似度达到该值才计入语义命中
 DEFAULT_PREFERENCE_SEMANTIC_THRESHOLD = 0.78
 PREFERENCE_SEMANTIC_THRESHOLD = float(
     os.getenv("PREFERENCE_SEMANTIC_THRESHOLD", str(DEFAULT_PREFERENCE_SEMANTIC_THRESHOLD))
@@ -41,6 +41,31 @@ def _normalize_text(text: str) -> str:
     return text.strip().casefold()
 
 
+def _find_exact_preference_match(
+    target_text: str,
+    normalized_preferences: list[tuple[str, str]],
+) -> str | None:
+    """
+    判断目标文本是否与某个用户兴趣词形成“精确命中”。
+    命中条件：
+    1. 标准化后完全相等
+    2. 二者互为包含关系
+    返回命中的原始兴趣词，未命中则返回 None。
+    """
+    normalized_target = _normalize_text(target_text)
+    if not normalized_target:
+        return None
+
+    for raw_pref, normalized_pref in normalized_preferences:
+        if not normalized_pref:
+            continue
+        if normalized_target == normalized_pref:
+            return raw_pref
+        if normalized_pref in normalized_target or normalized_target in normalized_pref:
+            return raw_pref
+    return None
+
+
 _EMBEDDING_CACHE: dict[str, np.ndarray] = {}
 MAX_CACHE_SIZE = 10000
 
@@ -86,6 +111,26 @@ def _build_keyword_embedding_map(keywords: list[str]) -> dict[str, np.ndarray]:
     return result
 
 
+def _find_best_semantic_match(
+    target_text: str,
+    target_vec_map: dict[str, np.ndarray],
+    pref_vec_map: dict[str, np.ndarray],
+) -> tuple[str | None, float]:
+    """返回与目标文本最接近的兴趣词及其余弦相似度。"""
+    target_vec = target_vec_map.get(target_text)
+    if target_vec is None:
+        return None, -1.0
+
+    best_pref = None
+    best_sim = -1.0
+    for pref_keyword, pref_vec in pref_vec_map.items():
+        sim = float(np.dot(target_vec, pref_vec))
+        if sim > best_sim:
+            best_sim = sim
+            best_pref = pref_keyword
+    return best_pref, best_sim
+
+
 def _ensure_aware(dt: datetime) -> datetime:
     """SQLite 读出的 datetime 不带时区信息，统一补上 UTC 后才能和 utcnow() 做减法。"""
     if dt.tzinfo is None:
@@ -116,8 +161,8 @@ def recommend_events_for_user(
 ) -> list[MatchedEventResult]:
     """
     用户兴趣推荐主流程：
-    1) 精确匹配：用户词 == EVENT 标签
+    1) 精确匹配：用户词 vs EVENT 标签/标题
-    2) 语义匹配：用户词向量 vs EVENT 标签向量（超过阈值）
+    2) 语义匹配：用户词向量 vs EVENT 标签/标题向量（超过阈值）
     3) 打分融合：匹配分 + 标签相关度 + 热度 + 新鲜度
     """
     final_limit = max(1, min(limit, PREFERENCE_RECOMMEND_MAX_LIMIT))
@@ -167,8 +212,6 @@ def recommend_events_for_user(
         )
         .all()
     )
-    if not topic_rows:
-        return []
 
     # 组织事件标签映射：event_id -> [(tag, relevance_score), ...]
     event_topics: dict[int, list[tuple[str, float | None]]] = {}
@@ -177,10 +220,6 @@ def recommend_events_for_user(
             continue
         event_topics.setdefault(event_id, []).append((topic_keyword, relevance_score))
 
-    # 如果某事件没有标签，就不参与推荐
-    if not event_topics:
-        return []
-
     # 3. 批量编码用户词与标签词，减少模型调用次数
     unique_preference_keywords = list(dict.fromkeys(preference_keywords))
     unique_topic_keywords = list(dict.fromkeys([row[1] for row in topic_rows if row[1]]))
@@ -188,13 +227,21 @@ def recommend_events_for_user(
     topic_vec_map = _build_keyword_embedding_map(unique_topic_keywords)
 
     # 预先建立“标准化后用户词集合”，用于精确匹配
-    normalized_pref_set = {_normalize_text(word) for word in unique_preference_keywords}
+    normalized_preference_pairs = [
+        (word, _normalize_text(word))
+        for word in unique_preference_keywords
+        if _normalize_text(word)
+    ]
+    unique_event_titles = list(
+        dict.fromkeys(
+            [event.unified_title.strip() for event in events if event.unified_title and event.unified_title.strip()]
+        )
+    )
+    title_vec_map = _build_keyword_embedding_map(unique_event_titles)
 
     scored_results: list[MatchedEventResult] = []
     for event in events:
         topic_list = event_topics.get(event.id, [])
-        if not topic_list:
-            continue
 
         exact_hits: list[str] = []
         semantic_hits: list[dict[str, Any]] = []
@@ -202,37 +249,18 @@ def recommend_events_for_user(
 
         # 对每个事件标签做精确匹配或语义匹配
         for topic_keyword, topic_relevance in topic_list:
-            normalized_topic = _normalize_text(topic_keyword)
             topic_relevance_score = float(topic_relevance) if topic_relevance is not None else 50.0
 
             # 1) 精确命中（包括完全相等与包含关系）
-            matched_exact = False
+            matched_pref = _find_exact_preference_match(topic_keyword, normalized_preference_pairs)
-            if normalized_topic in normalized_pref_set:
+            if matched_pref is not None:
-                matched_exact = True
-            else:
-                for pref_word in normalized_pref_set:
-                    if pref_word and (pref_word in normalized_topic or normalized_topic in pref_word):
-                        matched_exact = True
-                        break
-
-            if matched_exact:
                 exact_hits.append(topic_keyword)
                 # 精确命中给较高基础分，标签自身相关度作为增益
                 score += 45.0 + topic_relevance_score * 0.2
                 continue
 
             # 2) 语义命中（未精确命中时再算）
-            topic_vec = topic_vec_map.get(topic_keyword)
+            best_pref, best_sim = _find_best_semantic_match(topic_keyword, topic_vec_map, pref_vec_map)
-            if topic_vec is None:
-                continue
-
-            best_pref = None
-            best_sim = -1.0
-            for pref_keyword, pref_vec in pref_vec_map.items():
-                sim = float(np.dot(topic_vec, pref_vec))
-                if sim > best_sim:
-                    best_sim = sim
-                    best_pref = pref_keyword
 
             if best_pref is not None and best_sim >= similarity_threshold:
                 semantic_hits.append(
@@ -245,6 +273,25 @@ def recommend_events_for_user(
                 # 语义命中分略低于精确命中，并由相似度放大
                 score += best_sim * 35.0 + topic_relevance_score * 0.12
 
+        # 标题也参与匹配，但权重低于结构化标签，避免长标题过度主导排序。
+        event_title = (event.unified_title or "").strip()
+        if event_title:
+            title_exact_pref = _find_exact_preference_match(event_title, normalized_preference_pairs)
+            if title_exact_pref is not None:
+                exact_hits.append(f"标题:{title_exact_pref}")
+                score += 30.0
+            else:
+                best_pref, best_sim = _find_best_semantic_match(event_title, title_vec_map, pref_vec_map)
+                if best_pref is not None and best_sim >= similarity_threshold:
+                    semantic_hits.append(
+                        {
+                            "preference_keyword": best_pref,
+                            "topic_keyword": f"标题:{best_pref}",
+                            "similarity": round(best_sim, 4),
+                        }
+                    )
+                    score += best_sim * 24.0
+
         # 如果精确和语义都没命中，直接跳过
         if not exact_hits and not semantic_hits:
             continue

backend/dockerfile

@@ -1,31 +0,0 @@
-FROM python:3.11-slim AS builder
-
-WORKDIR /insightradar
-
-COPY pyproject.toml uv.lock ./
-
-RUN --mount=type=cache,target=/root/.cache/uv \
-    pip install --no-cache-dir uv && \
-    uv sync --frozen --no-dev
-
-COPY app app
-COPY main.py main.py
-
-#-----------------------------------------------
-
-FROM python:3.11-slim
-
-WORKDIR /insightradar
-
-# 👇 复制虚拟环境
-COPY --from=builder /insightradar/.venv /insightradar/.venv
-
-COPY app app
-COPY main.py main.py
-
-# 👇 关键：用 venv 里的 python
-ENV PATH="/insightradar/.venv/bin:$PATH"
-
-EXPOSE 8000
-
-CMD ["python","main.py"]

backend/main.py

@@ -1,12 +1,18 @@
 # run.py
 import uvicorn
+import os
+from dotenv import load_dotenv
 
 if __name__ == "__main__":
+    
+    load_dotenv()
+    PORT = int(os.getenv("PORT", 8000))
+    
     # 启动服务
     uvicorn.run(
         app="app.main:app",
         host="0.0.0.0",
-        port=8000,
+        port=PORT,
         # reload=True,
         workers=1
     )

backend/pyproject.toml

@@ -49,7 +49,6 @@ dependencies = [
     "safetensors==0.7.0",
     "scikit-learn==1.8.0",
     "scipy==1.17.1",
-    "sentence-transformers==5.2.3",
     "shellingham==1.5.4",
     "sniffio==1.3.1",
     "sqlalchemy==2.0.48",
@@ -57,8 +56,6 @@ dependencies = [
     "sympy==1.14.0",
     "threadpoolctl==3.6.0",
     "tokenizers==0.22.2",
-    "torch==2.10.0",
-    "torchvision==0.25.0",
     "tqdm==4.67.3",
     "transformers==5.3.0",
     "typer==0.24.1",
@@ -68,4 +65,16 @@ dependencies = [
     "tzlocal==5.3.1",
     "urllib3==2.6.3",
     "uvicorn==0.41.0",
+    "torch==2.11.0+cpu",
+    "torchvision==0.26.0+cpu",
+    "torchaudio==2.11.0+cpu",
+    "sentence-transformers>=5.3.0",
 ]
+
+[[tool.uv.index]]
+name = "pytorch-cpu"
+url = "https://download.pytorch.org/whl/cpu"
+default = false
+
+[tool.uv]
+index-strategy = "unsafe-best-match"

dockerfile

@@ -0,0 +1,50 @@
+# ---------- 阶段1：前端编译（Node打包静态产物） ----------
+FROM node:22-alpine AS frontend-builder
+
+WORKDIR /frontend
+
+# 复制前端依赖，利用Docker缓存优化
+COPY frontend/package*.json ./
+RUN npm install --registry=https://registry.npmmirror.com
+
+# 复制前端代码，编译出静态产物
+COPY frontend/ .
+RUN npm run build
+
+# ---------- 阶段2：后端依赖构建（uv构建虚拟环境） ----------
+FROM python:3.11-slim AS backend-builder
+
+WORKDIR /backend
+
+# 安装uv，同步Python依赖
+COPY backend/pyproject.toml backend/uv.lock ./
+RUN --mount=type=cache,target=/root/.cache/uv \
+    pip install --no-cache-dir uv && \
+    uv sync --frozen --no-dev --index https://pypi.tuna.tsinghua.edu.cn/simple/
+
+# 复制后端代码
+COPY backend/app ./app
+COPY backend/main.py ./
+
+# ---------- 阶段3：最终运行镜像（仅Python+Uvicorn，托管前端静态） ----------
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# 复制构建好的后端虚拟环境
+COPY --from=backend-builder /backend/.venv /app/.venv
+COPY --from=backend-builder /backend/app /app/app
+COPY --from=backend-builder /backend/main.py /app/main.py
+
+# 复制前端编译好的静态产物，放到后端能访问的目录
+# 这里我们把静态文件放到 /app/static 目录
+COPY --from=frontend-builder /frontend/dist /app/app/static
+
+# 把venv加入PATH
+ENV PATH="/app/.venv/bin:$PATH"
+
+# 暴露Uvicorn端口
+EXPOSE 8000
+
+# 直接启动Uvicorn，由Uvicorn配合后端框架托管静态文件
+CMD ["python3", "main.py"]

frontend/index.html

@@ -4,7 +4,7 @@
     <meta charset="UTF-8">
     <link rel="icon" href="/favicon.svg">
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>InsightRadar - 全网热点监控中枢</title>
+    <title>聚势智见 - 基于语义聚类与大模型的热点资讯聚合平台</title>
     <!-- Font Awesome 图标库 -->
     <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.1/css/all.min.css">
   </head>

frontend/src/assets/main.css

@@ -1,4 +1,4 @@
-@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&family=Noto+Sans+SC:wght@400;500;600;700&display=swap');
+@import url(./font.css);
 
 /* =========================================
    1. 现代 SaaS 风格高级主题变量

frontend/src/components/UnifiedEventCard.vue

@@ -111,6 +111,14 @@ function getRankingChartOptions(history: number[], platformColor: string) {
       height: 56,
       sparkline: { enabled: true },
       animations: { enabled: true, easing: 'easeinout' as const, speed: 400 },
+      events: {
+        mounted: (chartContext: any) => {
+          chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+        },
+        updated: (chartContext: any) => {
+          chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+        }
+      }
     },
     stroke: { curve: 'smooth' as const, width: 2 },
     fill: {

frontend/src/layouts/DashboardLayout.vue

@@ -57,7 +57,7 @@ function toggleSidebar() {
         <!-- Logo -->
         <div class="sidebar-logo">
           <BrandLogo />
-          <span class="logo-text">InsightRadar<span class="logo-dot">.AI</span></span>
+          <span class="logo-text">聚势智见<span class="logo-dot">.AI</span></span>
         </div>
 
         <!-- 导航菜单 -->

frontend/src/views/AboutView.vue

@@ -1,7 +1,7 @@
 <!-- 关于页（占位） -->
 <template>
   <div class="about">
-    <h1>关于 InsightRadar</h1>
+    <h1>关于 聚势智见</h1>
   </div>
 </template>
 

frontend/src/views/DashboardView.vue

@@ -182,6 +182,14 @@ function getRankingChartOptions(history: number[], platformColor: string) {
       height: 56,
       sparkline: { enabled: true },
       animations: { enabled: true, easing: 'easeinout' as const, speed: 400 },
+      events: {
+        mounted: (chartContext: any) => {
+          chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+        },
+        updated: (chartContext: any) => {
+          chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+        }
+      }
     },
     stroke: { curve: 'smooth' as const, width: 2 },
     fill: {

frontend/src/views/HomeView.vue

@@ -31,7 +31,7 @@ async function handleLogout() {
         <div class="nav-brand">
           <div class="logo">
             <BrandLogo />
-            InsightRadar
+            聚势智见
           </div>
         </div>
         <div class="nav-actions">

frontend/src/views/LoginView.vue

@@ -150,7 +150,7 @@ onUnmounted(() => {
       <div class="brand-content">
         <div class="logo">
           <BrandLogo />
-          InsightRadar
+          聚势智见
         </div>
         <h1 class="brand-title">洞察全网热点<br />让信息更聚焦</h1>
         <p class="brand-desc">
@@ -192,7 +192,7 @@ onUnmounted(() => {
       <div class="form-container">
         <div class="form-header">
           <h2>欢迎回来</h2>
-          <p>登录后继续查看 InsightRadar 实时动态</p>
+          <p>登录后继续查看 聚势智见 实时动态</p>
         </div>
 
         <div class="login-mode-tabs">

frontend/src/views/RegisterView.vue

@@ -131,7 +131,7 @@ onUnmounted(() => {
       <div class="brand-content">
         <div class="logo">
           <BrandLogo />
-          InsightRadar
+          聚势智见
         </div>
         <h1 class="brand-title">开启智能<br />分析之旅。</h1>
         <p class="brand-desc">

frontend/src/views/SearchView.vue

@@ -72,6 +72,12 @@ const chartOptions = ref<ApexOptions>({
     },
     // 点击图表数据点：切换选中时间，再次点击则取消筛选
     events: {
+      mounted: (chartContext: any) => {
+        chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+      },
+      updated: (chartContext: any) => {
+        chartContext.el?.querySelector('.apexcharts-svg > title')?.remove()
+      },
       markerClick: function(event: unknown, chartContext: unknown, { dataPointIndex }: never) {
         if (searchResult.value && searchResult.value.timeline[dataPointIndex]) {
           const clickedTime = searchResult.value.timeline[dataPointIndex].time_label
@@ -585,7 +591,12 @@ async function handleSearch() {
 
 .chart-container {
   margin-top: 16px;
-  margin-left: -10px; /* 视觉上抵消 apexcharts 的默认左侧留白。 */
+  margin-left: -10px; 
+}
+
+.chart-container :deep(svg),
+.chart-container :deep(canvas) {
+  outline: none;
 }
 
 .events-section {
@@ -595,7 +606,6 @@ async function handleSearch() {
 .events-grid {
   display: flex;
   flex-direction: column;
-  /* 与 DashboardView 保持一致，列表按纵向堆叠展示。 */
 }
 
 .loading-state {

frontend/src/views/TopicsView.vue

@@ -156,7 +156,7 @@ onMounted(async () => {
             v-model="newKeyword"
             type="text"
             class="keyword-input"
-            placeholder="输入关键词，如「直升机」「科比」「佐巴扬」..."
+            placeholder="输入关键词，如「篮球」「科比」「科技」..."
             maxlength="100"
             @keydown="onInputKeydown"
           />