Add web search knowledge base, message tagging, and word-memory tip topics

Introduces websearch.py (Tavily) + knowledge.py to ground chat responses
in search results, tagging.py to tag user and assistant messages via LLM
and pick related topics for the Discord useless-tips job, and
word_memory.py to track recent conversation nouns (surfaced in the new
WordPool.vue sidebar) used to pick tip topics close to a related tag.

backend/app/db/models.py

@@ -27,6 +27,18 @@ class Message(Base):
     created_at: Mapped[datetime] = mapped_column(DateTime, default=lambda: datetime.now(timezone.utc))
 
     session: Mapped["Session"] = relationship(back_populates="messages")
+    tags: Mapped[list["MessageTag"]] = relationship(back_populates="message", cascade="all, delete-orphan")
+
+
+class MessageTag(Base):
+    __tablename__ = "message_tags"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    message_id: Mapped[int] = mapped_column(ForeignKey("messages.id", ondelete="CASCADE"))
+    tag: Mapped[str] = mapped_column(String)
+    created_at: Mapped[datetime] = mapped_column(DateTime, default=lambda: datetime.now(timezone.utc))
+
+    message: Mapped["Message"] = relationship(back_populates="tags")
 
 
 class Memory(Base):
@@ -39,6 +51,41 @@ class Memory(Base):
     created_at: Mapped[datetime] = mapped_column(DateTime, default=lambda: datetime.now(timezone.utc))
 
     session: Mapped["Session"] = relationship(back_populates="memories")
+    tags: Mapped[list["MemoryTag"]] = relationship(back_populates="memory", cascade="all, delete-orphan")
+
+
+class MemoryTag(Base):
+    __tablename__ = "memory_tags"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    memory_id: Mapped[int] = mapped_column(ForeignKey("memories.id", ondelete="CASCADE"))
+    tag: Mapped[str] = mapped_column(String)
+    created_at: Mapped[datetime] = mapped_column(DateTime, default=lambda: datetime.now(timezone.utc))
+
+    memory: Mapped["Memory"] = relationship(back_populates="tags")
+
+
+class KnowledgeEntry(Base):
+    """Web検索で得た知識。セッションを横断して蓄積し、以後の会話のシステムプロンプトに注入する。"""
+    __tablename__ = "knowledge_entries"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    query: Mapped[str] = mapped_column(String)
+    summary: Mapped[str] = mapped_column(Text)
+    created_at: Mapped[datetime] = mapped_column(DateTime, default=lambda: datetime.now(timezone.utc))
+
+    sources: Mapped[list["KnowledgeSource"]] = relationship(back_populates="entry", cascade="all, delete-orphan")
+
+
+class KnowledgeSource(Base):
+    __tablename__ = "knowledge_sources"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    entry_id: Mapped[int] = mapped_column(ForeignKey("knowledge_entries.id", ondelete="CASCADE"))
+    title: Mapped[str] = mapped_column(String)
+    url: Mapped[str] = mapped_column(String)
+
+    entry: Mapped["KnowledgeEntry"] = relationship(back_populates="sources")
 
 
 class ApiLog(Base):