langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

astream bug #946

New Issue
Closed
opened 2026-02-20 17:42:29 -05:00 by yindo · 2 comments
yindo commented 2026-02-20 17:42:29 -05:00
Owner
Copy Link
Copy Source

Originally created by @xiaohuialg on GitHub (Aug 28, 2025).

Originally assigned to: @casparb on GitHub.

Checked other resources

  • This is a bug, not a usage question. For questions, please use the LangChain Forum (https://forum.langchain.com/).
  • I added a clear and detailed title that summarizes the issue.
  • I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example).
  • I included a self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue.

Example Code

agent_app = build_single_react_agent_app(
            tool_names=agent_configs.tool_names,
            mcp_servers=agent_configs.mcp_servers,
            knowledge_bases=agent_configs.knowledge_bases,
            agent_prompt=agent_configs.agent_prompt,
            think_switch=user_input.think_switch,
        )
        agent_app.get_graph().print_ascii()

        stream = agent_app.astream(
            input={"messages": messages},
            stream_mode=stream_mode,
            config=graph_config
        )

        try:
            async for output in stream:
                print(output)
                events = parser.filter_stream_output(output)
                if events:
                    for event in events:
                        yield f"data: {json.dumps(event, ensure_ascii=False)}\n\n"
                if (events and len(events) > 0 and
                        events[0].get('type') == 'finish_stop'):
                    return
        finally:
            try:
                await stream.aclose()
            except:
                pass

    return StreamingResponse(
        generate_json(),
        media_type="text/event-stream",
        headers={"Cache-Control": "no-cache"}
    )


def build_single_react_agent_app(
        tool_names: List[str],
        mcp_servers: Optional[Dict[str, Dict[str, Any]]] = None,
        knowledge_bases: List[str] = None,
        agent_prompt: str = "",
        think_switch: bool = False
):
    local_tools = get_tool()
    print("local_tools ", local_tools)
    mcp_tools = load_mcp_tools(mcp_servers)
    print("mcp_tools ", mcp_tools)
    tools_map = {**local_tools, **mcp_tools}
    print("tools_map ", tools_map)
    if not tool_names:
        tools = [tool for tool in tools_map.values() if not tool.name.startswith("workflow_")]
    else:
        workflow_tools_in_names = [name for name in tool_names if name.startswith("workflow_")]
        non_workflow_tools_in_names = [name for name in tool_names if not name.startswith("workflow_")]
        if non_workflow_tools_in_names:
            tools = [tools_map[name] for name in tool_names if name in tools_map]
        else:
            non_workflow_tools = [tool for tool in tools_map.values() if not tool.name.startswith("workflow_")]
            specified_workflow_tools = [tools_map[name] for name in workflow_tools_in_names if name in tools_map]
            tools = non_workflow_tools + specified_workflow_tools

    tools = [tool for tool in tools if not tool.name.startswith("alone_")]

    if knowledge_bases:
        db_session = next(get_db())
        crud = PostgresCRUD(db_url=None, model=RagKnowledgeBase, session=db_session)
        db_results = crud.get_by_names(knowledge_bases)
        if db_results:
            kb_info = {
                kb.name: kb.description
                for kb in db_results
            }
            kb_info_str = "\n".join(f"{name}:{desc}" for name, desc in kb_info.items())
            kb_keys = ", ".join(kb_info.keys())
            prompt_template = (
                "Use local knowledgebase from one or more of these:\n"
                "{kb_info_str}\n"
                "to get information. Only local data on this knowledge uses this tool. "
                "The 'database' should be one of the above [{kb_keys}]."
            )
            description = prompt_template.format(kb_info_str=kb_info_str, kb_keys=kb_keys)
            old_context = request_context.get()
            old_context["search_remote_knowledgebase_description"] = description
            request_context.set(old_context)
            tools.append(StructuredTool.from_function(
                func=sync_wrapper_for_dynamic_search,
                name="agent_search_knowledgebase",
                description=description,
                args_schema=DynamicInputSchema,
            ))

    agent = create_react_agent(
        model=gat_openai_model(think_switch),
        tools=tools,
        prompt=agent_prompt
    )

    return agent

Error Message and Stack Trace (if applicable)

Description

After using create_react_agent method, if you continue to use the astream method, if the MCP service tool parameter is empty, the call will get stuck and stop.

System Info

chatchat-server/chatchat/cli.py start --api

Originally created by @xiaohuialg on GitHub (Aug 28, 2025). Originally assigned to: @casparb on GitHub. ### Checked other resources - [x] This is a bug, not a usage question. For questions, please use the LangChain Forum (https://forum.langchain.com/). - [x] I added a clear and detailed title that summarizes the issue. - [x] I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example). - [x] I included a self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue. ### Example Code ```python agent_app = build_single_react_agent_app( tool_names=agent_configs.tool_names, mcp_servers=agent_configs.mcp_servers, knowledge_bases=agent_configs.knowledge_bases, agent_prompt=agent_configs.agent_prompt, think_switch=user_input.think_switch, ) agent_app.get_graph().print_ascii() stream = agent_app.astream( input={"messages": messages}, stream_mode=stream_mode, config=graph_config ) try: async for output in stream: print(output) events = parser.filter_stream_output(output) if events: for event in events: yield f"data: {json.dumps(event, ensure_ascii=False)}\n\n" if (events and len(events) > 0 and events[0].get('type') == 'finish_stop'): return finally: try: await stream.aclose() except: pass return StreamingResponse( generate_json(), media_type="text/event-stream", headers={"Cache-Control": "no-cache"} ) def build_single_react_agent_app( tool_names: List[str], mcp_servers: Optional[Dict[str, Dict[str, Any]]] = None, knowledge_bases: List[str] = None, agent_prompt: str = "", think_switch: bool = False ): local_tools = get_tool() print("local_tools ", local_tools) mcp_tools = load_mcp_tools(mcp_servers) print("mcp_tools ", mcp_tools) tools_map = {**local_tools, **mcp_tools} print("tools_map ", tools_map) if not tool_names: tools = [tool for tool in tools_map.values() if not tool.name.startswith("workflow_")] else: workflow_tools_in_names = [name for name in tool_names if name.startswith("workflow_")] non_workflow_tools_in_names = [name for name in tool_names if not name.startswith("workflow_")] if non_workflow_tools_in_names: tools = [tools_map[name] for name in tool_names if name in tools_map] else: non_workflow_tools = [tool for tool in tools_map.values() if not tool.name.startswith("workflow_")] specified_workflow_tools = [tools_map[name] for name in workflow_tools_in_names if name in tools_map] tools = non_workflow_tools + specified_workflow_tools tools = [tool for tool in tools if not tool.name.startswith("alone_")] if knowledge_bases: db_session = next(get_db()) crud = PostgresCRUD(db_url=None, model=RagKnowledgeBase, session=db_session) db_results = crud.get_by_names(knowledge_bases) if db_results: kb_info = { kb.name: kb.description for kb in db_results } kb_info_str = "\n".join(f"{name}:{desc}" for name, desc in kb_info.items()) kb_keys = ", ".join(kb_info.keys()) prompt_template = ( "Use local knowledgebase from one or more of these:\n" "{kb_info_str}\n" "to get information. Only local data on this knowledge uses this tool. " "The 'database' should be one of the above [{kb_keys}]." ) description = prompt_template.format(kb_info_str=kb_info_str, kb_keys=kb_keys) old_context = request_context.get() old_context["search_remote_knowledgebase_description"] = description request_context.set(old_context) tools.append(StructuredTool.from_function( func=sync_wrapper_for_dynamic_search, name="agent_search_knowledgebase", description=description, args_schema=DynamicInputSchema, )) agent = create_react_agent( model=gat_openai_model(think_switch), tools=tools, prompt=agent_prompt ) return agent ``` ### Error Message and Stack Trace (if applicable) ```shell ``` ### Description After using create_react_agent method, if you continue to use the astream method, if the MCP service tool parameter is empty, the call will get stuck and stop. ### System Info chatchat-server/chatchat/cli.py start --api
yindo added the bugpending labels 2026-02-20 17:42:29 -05:00
yindo closed this issue 2026-02-20 17:42:29 -05:00
yindo commented 2026-02-20 17:42:29 -05:00
Author
Owner
Copy Link
Copy Source

@casparb commented on GitHub (Sep 12, 2025):

Hi @xiaohuialg, I am closing this issue because your code is not an MRE. An MRE is a "self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue.".

If you're still experiencing issues, please tag me in a reply that includes an MRE that I can run as-is to reproduce the error, and I would be happy to help you out. Thanks!

@casparb commented on GitHub (Sep 12, 2025): Hi @xiaohuialg, I am closing this issue because your code is not an MRE. An MRE is a _"**self-contained, minimal example** that demonstrates the issue **INCLUDING** all the relevant imports. The code run **AS IS** to reproduce the issue."_. If you're still experiencing issues, please tag me in a reply that includes an MRE that I can run as-is to reproduce the error, and I would be happy to help you out. Thanks!
yindo commented 2026-02-20 17:42:29 -05:00
Author
Owner
Copy Link
Copy Source

@xiaohuialg commented on GitHub (Sep 12, 2025):

我已经收到，谢谢您！

@xiaohuialg commented on GitHub (Sep 12, 2025): 我已经收到，谢谢您！
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label bug pending
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#946
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?