Add shared notes with tabbed pages editable by user and AI

Adds a note_pages table backing a tabbed notes panel where the user and
the model can both read and write. Only page titles (plus a short
preview) are injected into the system prompt on every turn; full page
content is kept out of context and fetched on demand via a read_note
tool, with write_note letting the model append to, replace, or create a
page.

Generalizes llm.resolve_web_search into resolve_tool_calls so the
tool-calling loop handles web_search and the new note tools through one
path. The chat stream emits a notes_updated SSE event when the model
edits a note so the frontend panel can refresh; local edits are tracked
as dirty and never clobbered by that refresh.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

backend/app/api/chat.py

@@ -12,6 +12,7 @@ from app.core.memory_kb import memory as mem
 from app.core import llm
 from app.core.memory_kb import knowledge
 from app.core.memory_kb import local_knowledge
+from app.core.notes import NOTE_TOOLS
 from app.core.search.websearch import WEB_SEARCH_TOOL
 from app.core.config import DEFAULT_MODEL
 from app.core.memory_kb.tagging import tag_message_by_id
@@ -210,9 +211,12 @@ async def stream_chat(req: ChatRequest, db: DBSession = Depends(get_db)):
         used_model = req.model or session.model
         full_response = []
         try:
-            resolved_context, searches = await llm.resolve_web_search(
-                context, model=used_model, session_id=session.id,
+            resolved_context, searches, notes_changed = await llm.resolve_tool_calls(
+                context, model=used_model, session_id=session.id, db=db,
             )
+            if notes_changed:
+                # AIがノートを書き換えたので、フロントのノートパネルを再読み込みさせる
+                yield sse({"type": "notes_updated"})
             if searches:
                 sources = [
                     {"title": r["title"], "url": r["url"]}
@@ -225,9 +229,10 @@ async def stream_chat(req: ChatRequest, db: DBSession = Depends(get_db)):
                         knowledge.save_knowledge(db, s["query"], summary, s["results"])
                         local_knowledge.save_search_result(s["query"], summary, s["results"])
 
-            search_tools = [WEB_SEARCH_TOOL] if len(resolved_context) > len(context) else None
+            # 履歴にtool_callsが含まれる場合、最終呼び出しにも同じツール定義を渡す必要がある
+            used_tools = [WEB_SEARCH_TOOL, *NOTE_TOOLS] if len(resolved_context) > len(context) else None
             async for chunk in llm.stream_chat(
-                resolved_context, model=used_model, session_id=session.id, tools=search_tools,
+                resolved_context, model=used_model, session_id=session.id, tools=used_tools,
             ):
                 full_response.append(chunk)
                 yield sse({"type": "chunk", "value": chunk})