fix(backend): auto-purge polluted checkpoint when tool_calls lack ToolMessage

When a tool call crashes (e.g. httpx timeout in kb_search), the LangGraph
checkpoint retains an AIMessage with tool_calls but no corresponding
ToolMessage. Subsequent requests to the same conversation_id fail with:
  ValueError: Found AIMessages with tool_calls that do not have a
  corresponding ToolMessage

Now the except block in _stream_response detects this specific ValueError
by checking for "tool_calls" and "ToolMessage" in the error string, then
calls checkpointer.adelete_thread() to purge the corrupted thread state.
The frontend receives {"type":"error","content":"对话状态异常，已自动重置..."}
followed by {"type":"done"}, so the user can simply resend their message.

API confirmed: AsyncPostgresSaver.adelete_thread(thread_id) deletes from
checkpoints, checkpoint_blobs, and checkpoint_writes tables for the thread.

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

backend/app/api/chat.py

@@ -13,6 +13,7 @@ from litestar.response import Stream
 
 from app.graph.builder import get_chat_graph
 from app.schemas import ChatRequest
+from app.store.memory import get_checkpointer
 from app.store.postgres import Conversation, Message, async_session_factory
 from app.tools import resolve_tools
 from app.tools.search import set_search_model
@@ -118,8 +119,32 @@ async def _stream_response(request: ChatRequest) -> AsyncIterator[bytes]:
 
     except Exception as exc:
         logger.error("SSE stream error for conversation %s: %s", request.conversation_id, exc, exc_info=True)
+
+        error_msg = str(exc)
+        user_hint = "请求处理出现错误，请重试"
+
+        # Detect checkpoint pollution: a prior tool crash left an AIMessage
+        # with tool_calls but no corresponding ToolMessage. LangGraph refuses
+        # to continue the thread.  Purge the thread so the next request
+        # starts from a clean state.
+        if "tool_calls" in error_msg and "ToolMessage" in error_msg:
+            try:
+                checkpointer = await get_checkpointer()
+                await checkpointer.adelete_thread(request.conversation_id)
+                logger.warning(
+                    "Purged polluted checkpoint for thread %s",
+                    request.conversation_id,
+                )
+                user_hint = "对话状态异常，已自动重置，请重新发送消息"
+            except Exception:
+                logger.warning(
+                    "Failed to purge checkpoint for thread %s",
+                    request.conversation_id,
+                    exc_info=True,
+                )
+
         error_data = json.dumps(
-            {"type": "error", "content": "请求处理出现错误，请重试"},
+            {"type": "error", "content": user_hint},
             ensure_ascii=False,
         )
         yield f"data: {error_data}\n\n".encode("utf-8")