feat: persist cleaned question insights

ai_knowledge_base_v2/apps/backend/app/models/__init__.py

@@ -4,6 +4,7 @@ from app.models.base import Base
 from app.models.chat import ChatMessage, ChatSession, TopicSession
 from app.models.entitlement import EntitlementPlan, UserEntitlement, UserEntitlementLog
 from app.models.growth import GrowthProfileRevision, PeriodicReport, ShareDraft, TeacherHelpCard, TopicSummary, UserGrowthProfile
+from app.models.insight import QuestionInsightCleanedQuestion
 from app.models.knowledge import (
     HumanAttentionHistory,
     HumanAttentionRecord,
@@ -54,6 +55,7 @@ __all__ = [
     "TopicSession",
     "TopicSummary",
     "Prompt",
+    "QuestionInsightCleanedQuestion",
     "Role",
     "SystemConfig",
     "ShareDraft",

ai_knowledge_base_v2/apps/backend/app/models/insight.py

@@ -0,0 +1,55 @@
+from __future__ import annotations
+
+from datetime import datetime
+
+from sqlalchemy import BigInteger, DateTime, ForeignKey, Integer, String, Text, UniqueConstraint, func
+from sqlalchemy.orm import Mapped, mapped_column
+
+from app.models.base import Base
+
+
+class QuestionInsightCleanedQuestion(Base):
+    """用户问题清洗后的持久化结果。
+
+    每个原始消息至少写入一条记录。没有有效问题的消息会写入 accepted=0
+    的占位记录，使增量清洗无需重复读取已经处理过的聊天原文。
+    """
+
+    __tablename__ = "sys_question_insight_cleaned_question"
+    __table_args__ = (
+        UniqueConstraint(
+            "message_id",
+            "cleaner_version",
+            "part_index",
+            name="uq_question_insight_message_version_part",
+        ),
+    )
+
+    id: Mapped[int] = mapped_column(
+        BigInteger().with_variant(Integer, "sqlite"),
+        primary_key=True,
+        autoincrement=True,
+    )
+    message_id: Mapped[int] = mapped_column(
+        ForeignKey("sys_chat_message.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    session_id: Mapped[int] = mapped_column(BigInteger, index=True, nullable=False)
+    user_id: Mapped[int] = mapped_column(BigInteger, index=True, nullable=False)
+    part_index: Mapped[int] = mapped_column(Integer, nullable=False)
+    cleaner_version: Mapped[str] = mapped_column(String(20), nullable=False)
+    source_hash: Mapped[str] = mapped_column(String(64), nullable=False)
+    cleaned_text: Mapped[str] = mapped_column(Text, nullable=False, default="")
+    normalized_text: Mapped[str] = mapped_column(Text, nullable=False, default="")
+    category: Mapped[str] = mapped_column(String(50), nullable=False, default="other")
+    tokens_json: Mapped[str] = mapped_column(Text, nullable=False, default="[]")
+    accepted: Mapped[int] = mapped_column(Integer, nullable=False, default=1)
+    filtered_reason: Mapped[str | None] = mapped_column(String(100), nullable=True)
+    source_created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False)
+    created_at: Mapped[datetime] = mapped_column(DateTime, server_default=func.now(), nullable=False)
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime,
+        server_default=func.now(),
+        onupdate=func.now(),
+        nullable=False,
+    )