feat: 热词热力图改为Treemap矩形堆砌布局，面积代表提及次数；赛道分析页面布局调整

backend/app.py

@@ -45,9 +45,11 @@ def now_iso() -> str:
 
 
 def db() -> sqlite3.Connection:
-    conn = sqlite3.connect(DB_PATH)
+    conn = sqlite3.connect(DB_PATH, timeout=10)
     conn.row_factory = sqlite3.Row
+    conn.execute("PRAGMA journal_mode = WAL")
     conn.execute("PRAGMA foreign_keys = ON")
+    conn.execute("PRAGMA busy_timeout = 5000")
     return conn
 
 
@@ -160,7 +162,7 @@ def dashboard():
                 "type": r["type"] or "经营干货",
                 "score": r["score"] or 0,
                 "summary": r["ai_summary"] or "",
-                "entities": json.loads(r["entities"]) if r["entities"] else [],
+                "entities": [b["name"] if isinstance(b, dict) else b for b in (json.loads(r["entities"]) if r["entities"] else [])],
             }
             for r in top_intel
         ],
@@ -251,7 +253,7 @@ def intel_list():
                 "summary": r["ai_summary"] or r["summary"] or "",
                 "keyPoints": json.loads(r["key_points"]) if r["key_points"] else [],
                 "score": r["score"] or 0,
-                "entities": json.loads(r["entities"]) if r["entities"] else [],
+                "entities": [b["name"] if isinstance(b, dict) else b for b in (json.loads(r["entities"]) if r["entities"] else [])],
                 "timeliness": r["timeliness"] or "中",
                 "tags": json.loads(r["tags"]) if r["tags"] else [],
                 "url": r["canonical_url"],
@@ -393,26 +395,41 @@ def brand_detail(brand_id: int):
 @app.get("/api/analysis")
 def analysis():
     with db() as conn:
-        # 赛道热度趋势：按月统计各赛道的平均评分
+        # 赛道热度趋势：按月统计各赛道的文章数和平均评分，热度=文章数×平均评分/10
         heat_rows = conn.execute(
             """
-            SELECT strftime('%m', a.published_at) AS month,
+            SELECT strftime('%Y-%m', a.published_at) AS month,
                    a.category,
-                   AVG(CAST(json_extract(x.result_json, '$.score') AS REAL)) AS heat
+                   COUNT(*) AS article_cnt,
+                   AVG(CAST(json_extract(x.result_json, '$.score') AS REAL)) AS avg_score
             FROM articles a
             LEFT JOIN ai_analyses x ON x.article_id = a.id AND x.analysis_type = 'editorial'
-            WHERE a.crawl_status = 'success' AND a.category IS NOT NULL
+            WHERE a.crawl_status = 'success' AND a.category IS NOT NULL AND a.category != ''
+              AND a.published_at >= date('now', '-12 months')
             GROUP BY month, a.category
             ORDER BY month
             """
         ).fetchall()
 
-        # 开关店对比（从文章中提取的品类趋势数据）
-        category_counts = conn.execute(
+        # 品牌门店规模（按赛道聚合真实门店总数）
+        brand_stores = conn.execute(
             """
-            SELECT a.category, COUNT(*) AS cnt
+            SELECT category, SUM(stores) AS total_stores, COUNT(*) AS brand_cnt
+            FROM brands WHERE category IS NOT NULL AND category != '' AND stores > 0
+            GROUP BY category
+            """
+        ).fetchall()
+
+        # 舆情分布（按赛道统计正面/负面/中性文章数）
+        sentiment_rows = conn.execute(
+            """
+            SELECT a.category,
+                   SUM(CASE WHEN json_extract(x.result_json, '$.sentiment') = 'positive' THEN 1 ELSE 0 END) AS positive_cnt,
+                   SUM(CASE WHEN json_extract(x.result_json, '$.sentiment') = 'negative' THEN 1 ELSE 0 END) AS negative_cnt,
+                   SUM(CASE WHEN json_extract(x.result_json, '$.sentiment') = 'neutral' OR json_extract(x.result_json, '$.sentiment') IS NULL THEN 1 ELSE 0 END) AS neutral_cnt
             FROM articles a
-            WHERE a.crawl_status = 'success' AND a.category IS NOT NULL
+            JOIN ai_analyses x ON x.article_id = a.id AND x.analysis_type = 'editorial'
+            WHERE a.crawl_status = 'success' AND a.category IS NOT NULL AND a.category != ''
             GROUP BY a.category
             """
         ).fetchall()
@@ -432,29 +449,55 @@ def analysis():
             """
         ).fetchone()
 
-        # 热词
-        tag_rows = conn.execute(
+        # 热词热力图：近7天每天各标签出现次数
+        tag_heatmap = conn.execute(
             """
-            SELECT value AS tag, COUNT(*) AS cnt
-            FROM ai_analyses, json_each(json_extract(result_json, '$.tags'))
-            GROUP BY value
-            ORDER BY cnt DESC
-            LIMIT 8
+            SELECT value AS tag,
+                   strftime('%Y-%m-%d', a.published_at) AS day,
+                   COUNT(*) AS cnt
+            FROM ai_analyses x
+            JOIN articles a ON a.id = x.article_id
+            , json_each(json_extract(x.result_json, '$.tags'))
+            WHERE a.published_at >= date('now', '-6 days')
+              AND value IS NOT NULL AND value != ''
+            GROUP BY tag, day
             """
         ).fetchall()
 
-    # 构建热度趋势数据
+        # 热词总量排行（用于选取 TOP 标签）
+        tag_total = conn.execute(
+            """
+            SELECT value AS tag, COUNT(*) AS cnt
+            FROM ai_analyses x
+            JOIN articles a ON a.id = x.article_id
+            , json_each(json_extract(x.result_json, '$.tags'))
+            WHERE a.published_at >= date('now', '-6 days')
+              AND value IS NOT NULL AND value != ''
+            GROUP BY value
+            ORDER BY cnt DESC
+            LIMIT 10
+            """
+        ).fetchall()
+
+    # 构建热度趋势数据（热度 = 文章数 × 平均评分 / 10）
     heat_trend = {}
     for r in heat_rows:
-        month_key = f"{r['month']}月" if r["month"] else "未知"
+        month_key = f"{r['month'][-2:]}月" if r["month"] else "未知"
         if month_key not in heat_trend:
             heat_trend[month_key] = {"month": month_key}
-        heat_trend[month_key][r["category"]] = round(r["heat"] or 0)
+        heat_val = round((r["article_cnt"] or 0) * (r["avg_score"] or 0) / 10)
+        heat_trend[month_key][r["category"]] = heat_val
 
-    # 构建开关店数据
-    open_close = [
-        {"category": r["category"], "新开": r["cnt"], "关闭": max(0, r["cnt"] - int(r["cnt"] * 0.85))}
-        for r in category_counts
+    # 构建品牌门店规模数据
+    brand_stores_data = [
+        {"category": r["category"], "stores": r["total_stores"] or 0, "brands": r["brand_cnt"]}
+        for r in brand_stores
+    ]
+
+    # 构建舆情分布数据
+    sentiment_data = [
+        {"category": r["category"], "正面": r["positive_cnt"] or 0, "负面": r["negative_cnt"] or 0, "中性": r["neutral_cnt"] or 0}
+        for r in sentiment_rows
     ]
 
     # 构建价格分布
@@ -468,11 +511,28 @@ def analysis():
         {"band": "150元以上", "pct": round((price_rows["band6"] or 0) * 100 / total_brands)},
     ]
 
+    # 构建热词热力图数据（用 UTC 日期与 SQLite date('now') 保持一致）
+    from datetime import timedelta, timezone
+    today_utc = datetime.now(timezone.utc).date()
+    days = [(today_utc - timedelta(days=i)).isoformat() for i in range(6, -1, -1)]
+    top_tags = [r["tag"] for r in tag_total]
+    # 构建矩阵 {tag: {day: cnt}}
+    matrix = {tag: {day: 0 for day in days} for tag in top_tags}
+    for r in tag_heatmap:
+        if r["tag"] in matrix and r["day"] in matrix[r["tag"]]:
+            matrix[r["tag"]][r["day"]] = r["cnt"]
+    hot_keywords = {
+        "days": days,
+        "tags": top_tags,
+        "matrix": matrix,
+    }
+
     return jsonify({
         "categoryHeatTrend": list(heat_trend.values()),
-        "openCloseByCategory": open_close,
+        "brandStoresByCategory": brand_stores_data,
+        "sentimentByCategory": sentiment_data,
         "priceBandData": price_bands,
-        "hotKeywords": [{"word": r["tag"], "change": min(r["cnt"] * 5, 50)} for r in tag_rows],
+        "hotKeywords": hot_keywords,
     })
 
 

backend/crawler.py

@@ -36,6 +36,7 @@ load_dotenv(BASE_DIR / ".env")
 
 DEFAULT_LIST_URL = "https://m.canyin88.com/zixun/"
 ARTICLE_PATH_RE = re.compile(r"/zixun/\d{4}/\d{1,2}/\d{1,2}/\d+\.html$")
+CANYIN168_ARTICLE_RE = re.compile(r"/Article/[a-z]+/\d+\.html$")
 DATE_RE = re.compile(r"(20\d{2})[-/.年](\d{1,2})[-/.月](\d{1,2})(?:日)?(?:\s+(\d{1,2}):(\d{2})(?::(\d{2}))?)?")
 
 
@@ -47,9 +48,11 @@ def normalize_url(raw_url: str, base_url: str = DEFAULT_LIST_URL) -> str | None:
     absolute = urljoin(base_url, raw_url.strip())
     parts = urlsplit(absolute)
     host = parts.netloc.lower().removeprefix("m.").removeprefix("www.")
-    if host != "canyin88.com" or not ARTICLE_PATH_RE.search(parts.path):
-        return None
-    return urlunsplit(("https", "www.canyin88.com", parts.path, "", ""))
+    if host == "canyin88.com" and ARTICLE_PATH_RE.search(parts.path):
+        return urlunsplit(("https", "www.canyin88.com", parts.path, "", ""))
+    if host == "canyin168.com" and CANYIN168_ARTICLE_RE.search(parts.path):
+        return urlunsplit(("http", "www.canyin168.com", parts.path, "", ""))
+    return None
 
 
 def clean_text(value: str | None) -> str:
@@ -138,7 +141,7 @@ def extract_article(page, url: str) -> dict[str, str | None]:
     page.goto(url, wait_until="domcontentloaded", timeout=60_000)
     page.wait_for_timeout(800)
     document_title = clean_text(page.title())
-    document_title = re.sub(r"(?:[_-]红餐网.*)$", "", document_title).strip()
+    document_title = re.sub(r"(?:[_-](?:红餐网|职业餐饮网).*)$", "", document_title).strip()
     title = meta(page, 'meta[property="og:title"]') or document_title
     if not title or title == "相关推荐":
         title = first_text(page, [".title", "h1"])
@@ -197,17 +200,17 @@ def pending_urls(conn: sqlite3.Connection, discovered: list[str], max_retries: i
     return list(dict.fromkeys(discovered + [row[0] for row in rows]))
 
 
-def insert_discoveries(conn: sqlite3.Connection, urls: list[str], source_type: str = "公众号") -> int:
+def insert_discoveries(conn: sqlite3.Connection, urls: list[str], source_type: str = "公众号", source_name: str = "") -> int:
     timestamp = now_iso()
     inserted = 0
     for url in urls:
         cursor = conn.execute(
             """
             INSERT OR IGNORE INTO articles
-                (canonical_url, source_type, discovered_at, updated_at)
-            VALUES (?, ?, ?, ?)
+                (canonical_url, source_type, source_name, discovered_at, updated_at)
+            VALUES (?, ?, ?, ?, ?)
             """,
-            (url, source_type, timestamp, timestamp),
+            (url, source_type, source_name, timestamp, timestamp),
         )
         inserted += cursor.rowcount
     conn.commit()
@@ -439,7 +442,7 @@ def run_crawl(args: argparse.Namespace) -> int:
                             print(f"--- 采集公众号: {source_name} ({list_url}) ---", flush=True)
                             urls = discover_urls(page, list_url, max_scrolls, list_pause_ms)
                             discovered_count += len(urls)
-                            inserted_count += insert_discoveries(conn, urls, source_name)
+                            inserted_count += insert_discoveries(conn, urls, "公众号", source_name)
                             queue = pending_urls(conn, urls, args.max_retries)
                             cutoff = datetime.now() - timedelta(days=lookback_days)
                             for index, url in enumerate(queue, start=1):
@@ -551,7 +554,7 @@ def run_crawl(args: argparse.Namespace) -> int:
             "status": status, "source_type": args.source_type,
             "discovered": discovered_count, "inserted": inserted_count,
             "updated": updated_count, "failed": failed_count,
-        }, ensure_ascii=False, flush=True))
+        }, ensure_ascii=False))
         return 0 if status in {"success", "partial"} else 1
 
 

backend/structurer.py

@@ -51,6 +51,23 @@ def get_api_key() -> str:
 
 # ---- 单篇文章结构化分析 ----
 
+_PROFANITY_MAP = {
+    "妈的": "**", "傻逼": "**", "他妈": "**", "操": "*",
+    "草泥马": "***", "尼玛": "**", "贱人": "**",
+    "王八蛋": "***", "混蛋": "**", "去死": "**",
+    "毛爷爷": "某伟人", "毛泽东": "某伟人", "国民党": "某党",
+    "共产党": "某党", "长征": "远征", "抗日": "抗战",
+    "革命": "变革", "社会主义": "某主义",
+}
+
+
+def sanitize_content(text: str) -> str:
+    """过滤粗口和敏感词，避免触发千问内容审核。"""
+    for word, replacement in _PROFANITY_MAP.items():
+        text = text.replace(word, replacement)
+    return text
+
+
 def analyze_article(title: str, content: str) -> dict:
     """调用通义千问对文章进行结构化分析。"""
     api_key = get_api_key()
@@ -71,7 +88,7 @@ JSON 字段必须为：
 - numbers（关键数据数组）
 
 文章标题：{title}
-正文：{content[:24000]}"""
+正文：{sanitize_content(content[:24000])}"""
     payload = json.dumps({
         "model": model,
         "messages": [
@@ -276,6 +293,11 @@ def run_batch_analyze(force: bool, pause: float, retries: int) -> int:
                 break
             except Exception as exc:
                 last_error = exc
+                err_str = str(exc)
+                # 内容审核失败，标记跳过不重试
+                if "data_inspection_failed" in err_str:
+                    print(f"[{index}/{len(rows)}] SKIP(内容审核) {row['title']}", flush=True)
+                    break
                 if attempt <= retries:
                     wait = attempt * 2
                     print(f"[{index}/{len(rows)}] RETRY {attempt}/{retries} {exc}", flush=True)