feat: 垃圾信息分类标签功能

新增垃圾信息细分类标签，在朴素贝叶斯二分类基础上对spam进行细分：
- 新增 spam_categorizer.py 分类模块（诈骗/骚扰/广告）
- SpamPredictionLog 和 ContentPost 模型添加 category 字段
- content_routes 和 spam_routes 接口返回分类标签

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/routes/spam_routes.py

@@ -3,6 +3,7 @@ from flask_jwt_extended import jwt_required
 
 from app.extensions import db
 from app.ml.naive_bayes_classifier import NaiveBayesSpamClassifier
+from app.ml.spam_categorizer import categorize_spam, get_category_label
 from app.models import DetectionConfig, SpamPredictionLog, SpamTrainingSample
 from app.utils.auth import admin_required, current_user
 from app.utils.response import fail, ok
@@ -58,10 +59,17 @@ def predict_one():
     threshold = _adjusted_threshold(user.credit_score or 100)
     blocked = float(result["spam_probability"]) >= threshold
 
+    # 分类标签：仅在判定为垃圾时进行细分
+    category = ""
+    category_label = ""
+    if blocked:
+        category, category_label = categorize_spam(result["text"])
+
     row = SpamPredictionLog(
         user_id=user.id,
         text=result["text"],
         prediction=result["prediction"],
+        category=category,
         spam_probability=result["spam_probability"],
         ham_probability=result["ham_probability"],
         confidence=result["confidence"],
@@ -71,7 +79,14 @@ def predict_one():
     db.session.add(row)
     db.session.commit()
 
-    return ok({**result, "log_id": row.id, "threshold": threshold, "blocked_by_threshold": blocked}, "识别成功")
+    return ok({
+        **result,
+        "log_id": row.id,
+        "threshold": threshold,
+        "blocked_by_threshold": blocked,
+        "category": category,
+        "category_label": category_label,
+    }, "识别成功")
 
 
 @spam_bp.post("/predict/batch")
@@ -98,12 +113,23 @@ def predict_batch():
         if len(content) < 2:
             continue
         result = clf.predict(content)
-        result["blocked_by_threshold"] = float(result["spam_probability"]) >= threshold
+        blocked = float(result["spam_probability"]) >= threshold
+        result["blocked_by_threshold"] = blocked
+
+        # 分类标签
+        category = ""
+        category_label = ""
+        if blocked:
+            category, category_label = categorize_spam(result["text"])
+        result["category"] = category
+        result["category_label"] = category_label
+
         rows.append(
             SpamPredictionLog(
                 user_id=user.id,
                 text=result["text"],
                 prediction=result["prediction"],
+                category=category,
                 spam_probability=result["spam_probability"],
                 ham_probability=result["ham_probability"],
                 confidence=result["confidence"],