refactor: extract event_loop_utils module and improve agent response handling by MkDev11 · Pull Request #999 · eigent-ai/eigent

MkDev11 · 2026-01-21T13:18:00Z

Summary

Refactors event loop utilities and agent streaming response handling for improved code organization and thread-safety.

Changes

New Module: `event_loop_utils.py`

Extracted set_main_event_loop() and _schedule_async_task() into a dedicated utility module
Provides thread-safe async task scheduling from worker threads using contextvars + global fallback

Refactored listen_chat_agent.py

Added reusable helper methods:
- _send_agent_deactivate() - Sends agent deactivation events
- _extract_tokens() - Extracts token count from responses
- _stream_chunks() / _astream_chunks() - Wraps streaming responses for consistent handling
Fixed thread-safety: Uses _schedule_async_task instead of asyncio.create_task for toolkit activation/deactivation calls
Fixed AsyncStreamingChatAgentResponse handling to properly iterate through chunks

Improved `single_agent_worker.py`

Better streaming response handling with last_chunk tracking
Properly captures usage info from streaming responses after iteration

Cleanup

Removed duplicate code

MkDev11 · 2026-01-21T13:36:57Z

@a7m-1st @fengju0213 could you please review the changes and let me know your feedback? thanks!

a7m-1st · 2026-01-21T18:07:26Z

Thanks for the PR @MkDev11 , I don't think I can manage to test this by today (UTC +3).
But have you taken a look at #767 ? kind of similar.

But UI wise can you attach a screenshot of which part you are exactly streaming? Just to get a faster picture !

The camel library throws 'AsyncChatCompletionStreamManager object has no attribute choices' when streaming is enabled for agents with tools attached. Changes: - Disable streaming for all astep calls in ListenChatAgent to avoid the error - Preserve model_config_dict when cloning agents (for future streaming support) - Add streaming infrastructure (ActionStreamingAgentOutputData) for when camel library support is available - Handle AsyncStreamingChatAgentResponse in single_agent_worker.py Note: Streaming text output for worker agents is blocked by upstream camel library limitation. Task decomposition streaming still works via decompose_text.

Shows real-time '[Tool] Shell Exec...', '[Tool] Write File...' etc. messages in the UI while tools are executing, giving users feedback on agent activity. Changes: - Add ActionStreamingAgentOutputData events in @listen_toolkit decorator - Add streaming events in _execute_tool and _aexecute_tool methods - Disable model streaming for worker agents (camel library limitation) - Replace emoji with [Tool] text for better compatibility Closes eigent-ai#87

MkDev11 · 2026-01-22T13:49:45Z

@a7m-1st I added some changes and updated the description. This PR adds some real-time feedback while tasks are running.

Before, you'd just see the task start and then wait until it finished. Now you'll see messages like [Tool] Shell Exec... pop up as the agent works, so you know what's actually happening behind the scenes.

I originally wanted to stream the model's actual thinking text (like ChatGPT does), but hit a wall - the camel library we use crashes when you try to stream with tools enabled. Since pretty much every agent uses tools, I pivoted to showing tool activity instead. Not quite what issue #87 asked for, but it's still a nice improvement.

MkDev11 · 2026-01-22T13:51:42Z

a7m-1st · 2026-01-22T16:36:52Z

I see awesome then Thanks @MkDev11, I need time so far as I am little occupied with other tasks. I will catch up with this PR once I got the chance.

MkDev11 · 2026-01-22T16:37:57Z

Cool!

…

On Thu, Jan 22, 2026 at 11:37 AM Ahmed Awelkair A ***@***.***> wrote: *a7m-1st* left a comment (eigent-ai/eigent#999) <#999 (comment)> I see awesome then Thanks @MkDev11 <https://github.com/MkDev11>, I need time so far as I am little occupied with other tasks. I will catch up with this PR once I got the chance. — Reply to this email directly, view it on GitHub <#999 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWOUTY6VPENCFY57A6CZXZT4ID4DVAVCNFSM6AAAAACSM4HTTSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOOBVGM4DQMRSGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

a7m-1st · 2026-01-22T16:46:03Z

Hi there @Wendong-Fan , I think this would be a cool feature maybe Douglas can confirm if the UI meets expectations?

fengju0213 · 2026-01-23T07:17:15Z

@MkDev11 thanks for your contribution，i noticed that you mentioned that the camel library throws AsyncChatCompletionStreamManager errors when streaming with tools，however, Camel normally supports tool calls in streaming mode. Could I please tell you under what circumstances you encounter the error?

MkDev11 · 2026-01-23T09:02:13Z

@MkDev11 thanks for your contribution，i noticed that you mentioned that the camel library throws AsyncChatCompletionStreamManager errors when streaming with tools，however, Camel normally supports tool calls in streaming mode. Could I please tell you under what circumstances you encounter the error?

Hey @fengju0213, good catch - my comment in the code was a bit misleading. It's not streaming + tools that's the issue, it's streaming + response_format (structured output).

Here's what happens:

model = ModelFactory.create(
    model_platform=ModelPlatformType.OPENAI,
    model_type=ModelType.GPT_4O_MINI,
    model_config_dict={'stream': True}
)
agent = ChatAgent(system_message='...', model=model)

# this breaks
response = await agent.astep('query', response_format=SomePydanticModel)
async for chunk in response:
    print(chunk)

You get:

AttributeError: 'AsyncChatCompletionStreamManager' object has no attribute 'choices'

The problem is in _handle_batch_response (chat_agent.py:3743) - it tries to do for choice in response.choices but OpenAI returns an AsyncChatCompletionStreamManager when you combine streaming with structured output, and that object doesn't have .choices.

So I just disable streaming when we need structured output:

worker_agent.model_backend.model_config_dict["stream"] = False
response = await worker_agent.astep(prompt, response_format=TaskResult)

Let me know what you think

The issue is streaming + response_format (structured output), not streaming + tools

…nt-output

Includes fix for AsyncChatCompletionStreamManager from PR #3743

a7m-1st · 2026-01-26T22:42:56Z

Hopfully we an resolve this in camel-ai/camel#3743

I believe we can simplify the configs now bcz it has been fixed @MkDev11 or are they unrelated? i.e. the manual setting of steam mode and on clone too

Remove specific mention of stream setting since the workaround for AsyncChatCompletionStreamManager is no longer needed.

MkDev11 · 2026-01-26T22:57:18Z

Hopfully we an resolve this in camel-ai/camel#3743

I believe we can simplify the configs now bcz it has been fixed @MkDev11 or are they unrelated? i.e. the manual setting of steam mode and on clone too

yep, I've simplified it! Removed the workarounds from both agent.py and single_agent_worker.py, and updated camel-ai to 0.2.85 which includes the fix. The stream mode settings and clone preservation are intentional config, not workarounds, so those stay.

MkDev11 · 2026-01-26T22:57:34Z

@a7m-1st please review the changes again

a7m-1st · 2026-01-27T13:51:03Z

All right sure, let me run it through the debugger to reconfirm the flow.
Aside than this, if @Wendong-Fan and @Douglasymlai find it a cool feature, I think it should be good 👍

…nt structure

MkDev11 · 2026-02-01T19:05:17Z

hello @Wendong-Fan I am really sorry for tagging you, could you please give me any update for me?

…ent output

…ature

…tep enum

a7m-1st · 2026-02-03T01:26:22Z

Hi there @MkDev11 , I have just pushed my enhancements, apologies for the huge delays.

simplified the set_main_event_loop & _schedule_async_task to not be redundant
isolate functions to _send_streaming_chunk, _send_agent_deactivate, _extract_tokens & _stream_chunks

…s and isolated functions - Remove streaming_agent_output SSE step and ActionStreamingAgentOutputData - Remove _send_streaming_chunk method and [Tool] display calls - Revert frontend files to upstream/main (no streaming UI changes) - Keep refactored utilities: event_loop_utils, _schedule_async_task, _send_agent_deactivate, _extract_tokens, _stream_chunks

Wendong-Fan

Thanks for the cleanup! The extraction of event loop utils and the streaming response improvements look good.

Found a couple of things:

_send_agent_deactivate uses asyncio.create_task() but this method gets called from streaming generators' finally blocks which can run in worker threads. Should use _schedule_async_task() instead (same as _execute_tool does).
In astep(), deactivation only fires when res is not None. On exceptions, res stays None so no deactivation is sent and the frontend thinks the agent is still running. The step() method handles this correctly by sending deactivation unconditionally for all non-streaming cases. Would be good to match that pattern here.

Wendong-Fan · 2026-02-03T13:57:53Z

added commit f7ab0f8 based on review

Resolved conflicts in: - app/agent/agent_model.py: Keep event_loop_utils import approach - app/agent/listen_chat_agent.py: Keep helper methods and cleaner formatting 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>

a7m-1st · 2026-02-03T20:11:37Z

Thanks @Wendong-Fan for your review, and thanks @MkDev11 for the PR

Wendong-Fan requested review from a7m-1st and fengju0213 and removed request for a7m-1st January 21, 2026 13:35

MkDev11 closed this Jan 22, 2026

MkDev11 force-pushed the feature/streaming-agent-output branch from d81f18c to 35f4ca2 Compare January 22, 2026 02:08

.

b3b7ed6

MkDev11 reopened this Jan 22, 2026

MkDev11 added 2 commits January 22, 2026 13:50

MkDev11 changed the title ~~feat: implement streaming output for agent responses~~ feat: add streaming tool activity messages during task execution Jan 22, 2026

Merge branch 'main' into feature/streaming-agent-output

d1e0e99

a7m-1st assigned MkDev11 Jan 22, 2026

a7m-1st added the Review Required label Jan 22, 2026

MkDev11 added 2 commits January 22, 2026 11:50

Merge branch 'main' into feature/streaming-agent-output

70b2357

Merge branch 'main' into feature/streaming-agent-output

15d5d7e

Wendong-Fan requested review from 4pmtong and Douglasymlai January 23, 2026 01:04

Merge branch 'main' into feature/streaming-agent-output

12a1a5c

fix: correct comment about streaming error cause

309883d

The issue is streaming + response_format (structured output), not streaming + tools

MkDev11 mentioned this pull request Jan 23, 2026

fix: detect AsyncChatCompletionStreamManager by class name camel-ai/camel#3743

Merged

6 tasks

Merge remote-tracking branch 'origin/main' into feature/streaming-age…

e7b2269

…nt-output

chore: update camel-ai to 0.2.85

6759353

Includes fix for AsyncChatCompletionStreamManager from PR #3743

refactor: update clone preservation comment

32316d2

Remove specific mention of stream setting since the workaround for AsyncChatCompletionStreamManager is no longer needed.

a7m-1st mentioned this pull request Jan 27, 2026

feat: add RAG toolkit for knowledge base queries #1003

Merged

4 tasks

Merge upstream/main - port streaming output feature to refactored age…

523ba97

…nt structure

MkDev11 force-pushed the feature/streaming-agent-output branch from 800cebd to 523ba97 Compare January 30, 2026 22:08

MkDev11 and others added 6 commits February 1, 2026 20:11

Merge upstream/main - resolve frontend conflicts and add streaming ag…

78d7962

…ent output

fix: PEP8 line too long in task.py

5c65669

style: yapf format task.py

927244f

fix: resolve conflict in node.tsx, preserve streaming agent output fe…

1087a17

…ature

fix: resolve merge conflicts with upstream/main, update to use AgentS…

7fdfadf

…tep enum

chore: isolate set_main_event_loop & _schedule_async_task

4dfef33

MkDev11 changed the title ~~feat: add streaming tool activity messages during task execution~~ refactor: extract event_loop_utils module and improve agent response handling Feb 3, 2026

MkDev11 added 3 commits February 3, 2026 03:07

fix: add mock for _schedule_async_task in agent model test

66c3df6

style: fix lint issues with ruff

c5ddcb4

style: fix import order with isort

257597f

Wendong-Fan reviewed Feb 3, 2026

View reviewed changes

update based on review comment

f7ab0f8

Wendong-Fan and others added 2 commits February 3, 2026 22:01

using _schedule_async_task

77f96d9

Wendong-Fan added this to the Sprint 13 milestone Feb 3, 2026

Wendong-Fan merged commit 7c287fe into eigent-ai:main Feb 3, 2026
5 of 6 checks passed

Conversation

MkDev11 commented Jan 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

New Module: event_loop_utils.py

Refactored listen_chat_agent.py

Improved single_agent_worker.py

Cleanup

Uh oh!

MkDev11 commented Jan 21, 2026

Uh oh!

a7m-1st commented Jan 21, 2026

Uh oh!

MkDev11 commented Jan 22, 2026

Uh oh!

MkDev11 commented Jan 22, 2026

Uh oh!

a7m-1st commented Jan 22, 2026

Uh oh!

MkDev11 commented Jan 22, 2026 via email

Uh oh!

a7m-1st commented Jan 22, 2026

Uh oh!

fengju0213 commented Jan 23, 2026

Uh oh!

MkDev11 commented Jan 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

a7m-1st commented Jan 26, 2026

Uh oh!

MkDev11 commented Jan 26, 2026

Uh oh!

MkDev11 commented Jan 26, 2026

Uh oh!

a7m-1st commented Jan 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

MkDev11 commented Feb 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

a7m-1st commented Feb 3, 2026

Uh oh!

Wendong-Fan left a comment

Choose a reason for hiding this comment

Uh oh!

Wendong-Fan commented Feb 3, 2026

Uh oh!

Uh oh!

a7m-1st commented Feb 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

MkDev11 commented Jan 21, 2026 •

edited

Loading

New Module: `event_loop_utils.py`

Improved `single_agent_worker.py`

MkDev11 commented Jan 23, 2026 •

edited

Loading

a7m-1st commented Jan 27, 2026 •

edited

Loading

MkDev11 commented Feb 1, 2026 •

edited

Loading