fix: resolve 10 code review issues

High priority:
- Fix concurrent race condition for view_count/like_count (atomic update)
- Add route request ID tracking to prevent race conditions
- Filter get_joke by status=approved (no pending content leak)
- Add error feedback for like button

Performance:
- Optimize random joke query (avoid full table sort)
- Limit page_size max to 100 (DoS prevention)

Medium:
- Add localStorage quota error handling
- Handle empty AI response gracefully
- Fix generate content title extraction

Low:
- Add rejected_jokes to stats API
- Update dashboard to show rejected count

api/app/models/setting.py

@@ -0,0 +1,23 @@
+from sqlalchemy import Boolean, Column, DateTime, Float, Integer, String, Text, func
+
+from app.database import Base
+
+
+class AiSetting(Base):
+    __tablename__ = "ai_settings"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    provider = Column(String(50), nullable=False, default="nvidia")
+    api_base = Column(String(500), nullable=False, default="https://integrate.api.nvidia.com/v1")
+    api_key = Column(String(500), nullable=False, default="")
+    model_name = Column(String(100), nullable=False, default="nvidia/llama-3.1-nemotron-70b-instruct")
+    temperature = Column(Float, nullable=False, default=0.7)
+    max_tokens = Column(Integer, nullable=False, default=2048)
+
+    crawl_enabled = Column(Boolean, nullable=False, default=False)
+    crawl_keywords = Column(Text, nullable=True, default="")
+    max_pages_per_run = Column(Integer, nullable=False, default=3)
+
+    is_active = Column(Boolean, nullable=False, default=True)
+    created_at = Column(DateTime, default=func.now())
+    updated_at = Column(DateTime, default=func.now(), onupdate=func.now())