Add math and learn-to-ask langchain demos #2

Kanaricc · 2026-01-11T03:24:58Z

No description provided.

…to LangChain

gemini-code-assist · 2026-01-11T03:25:13Z

Summary of Changes

Hello @Kanaricc, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly enhances the project by integrating new Langchain-based workflows. It introduces a sophisticated 'learn-to-ask' medical assistant demo, complete with reward functions for evaluating question quality, and a dedicated math agent capable of solving problems. These additions broaden the framework's capabilities in handling complex agentic tasks and demonstrate practical applications of large language models, supported by necessary dependency updates and minor code cleanup.

Highlights

New Langchain Demos: Introduced two new demonstration workflows utilizing the Langchain library: one for a 'learn-to-ask' medical assistant and another for a math-solving agent.
Dependency Update: Added langchain>=1.2.3 as a development dependency in pyproject.toml to support the new demos.
Compatibility Improvement: Expanded verl version compatibility in ajet/utils/launch_utils.py to include development versions (specifically '0.5.0.dev').

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This pull request introduces two new demonstration scripts utilizing langchain for a math agent and a 'learn-to-ask' agent. The changes primarily involve adding these new files and including langchain as a dependency. My review is focused on the new scripts, where I've identified opportunities for improvement. These include cleaning up unused imports, correcting a workflow name that seems to have been copied from another example, and addressing a more significant issue with the use of threading primitives in an asynchronous context. I have provided specific suggestions for these points to enhance code quality and correctness.

gemini-code-assist · 2026-01-11T03:26:16Z

tutorial/example_learn2ask/learn2ask_langchain.py

+_reward_semaphore = threading.Semaphore(16)
+
+async def reward_fn_with_semaphore(*args, **kwargs):
+
+    get_sem_ok = False
+    while not get_sem_ok:
+        get_sem_ok = _reward_semaphore.acquire(blocking=False)
+        if not get_sem_ok:
+            await asyncio.sleep(1)
+
+    try:
+        fn_result = await reward_fn(*args, **kwargs)
+    finally:
+        _reward_semaphore.release()
+
+    return fn_result


The current implementation uses threading.Semaphore with a polling loop (while not get_sem_ok...) in an asyncio context. This is inefficient as it involves busy-waiting. A more idiomatic and efficient approach for asynchronous code is to use asyncio.Semaphore, which integrates with the async with statement for cleaner and non-blocking synchronization.

Suggested change

_reward_semaphore = threading.Semaphore(16)

async def reward_fn_with_semaphore(*args, **kwargs):

get_sem_ok = False

while not get_sem_ok:

get_sem_ok = _reward_semaphore.acquire(blocking=False)

if not get_sem_ok:

await asyncio.sleep(1)

try:

fn_result = await reward_fn(*args, **kwargs)

finally:

_reward_semaphore.release()

return fn_result

_reward_semaphore = asyncio.Semaphore(16)

async def reward_fn_with_semaphore(*args, **kwargs):

async with _reward_semaphore:

return await reward_fn(*args, **kwargs)

gemini-code-assist · 2026-01-11T03:26:16Z

tutorial/example_learn2ask/learn2ask_langchain.py

+import asyncio
+import threading
+
+from agentscope.message import Msg


The Msg class is imported from agentscope.message but is not used within this file. It's good practice to remove unused imports to keep the code clean and maintainable.

gemini-code-assist · 2026-01-11T03:26:16Z

tutorial/example_learn2ask/learn2ask_langchain.py

+
+
+class ExampleLearn2Ask(Workflow):
+    name: str = "math_agent_workflow"


The workflow name math_agent_workflow seems to be a copy-paste from another example. To improve clarity and avoid potential confusion, it should be renamed to something more descriptive of its actual function, like learn2ask_langchain_workflow.

Suggested change

name: str = "math_agent_workflow"

name: str = "learn2ask_langchain_workflow"

gemini-code-assist · 2026-01-11T03:26:16Z

tutorial/example_math_agent/math_agent_langchain.py

+from loguru import logger
+from ajet import AjetTuner, Workflow, WorkflowOutput, WorkflowTask
+from openai.types.chat.chat_completion import ChatCompletion
+from openai.types.chat import ChatCompletionMessageToolCall
+from textwrap import dedent
+
+import json
+import asyncio
+import requests
+from langchain.agents import create_agent


This file contains several unused imports, including loguru, openai.types, json, asyncio, and requests. Removing these will improve code readability and maintainability. It's also a good practice to group and order imports according to PEP 8.

Suggested change

from loguru import logger

from ajet import AjetTuner, Workflow, WorkflowOutput, WorkflowTask

from openai.types.chat.chat_completion import ChatCompletion

from openai.types.chat import ChatCompletionMessageToolCall

from textwrap import dedent

import json

import asyncio

import requests

from langchain.agents import create_agent

from textwrap import dedent

from langchain.agents import create_agent

from ajet import AjetTuner, Workflow, WorkflowOutput, WorkflowTask

Kanaricc added 5 commits January 9, 2026 18:27

fix: bugs

d05496a

feat: introduce naive langchain examples

35025aa

fix: message format in agent langchain implementation

8b5f090

feat: new implementation migrating learn2ask example from AgentScope …

bb8edfe

…to LangChain

fix: restore original code

0347c80

Kanaricc requested a review from binary-husky January 11, 2026 03:25

gemini-code-assist bot reviewed Jan 11, 2026

View reviewed changes

Kanaricc merged commit 886d7c0 into main Jan 12, 2026
0 of 2 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add math and learn-to-ask langchain demos #2

Add math and learn-to-ask langchain demos #2

Uh oh!

Kanaricc commented Jan 11, 2026

Uh oh!

gemini-code-assist bot commented Jan 11, 2026

Uh oh!

gemini-code-assist bot left a comment

Uh oh!

gemini-code-assist bot Jan 11, 2026

Uh oh!

gemini-code-assist bot Jan 11, 2026

Uh oh!

gemini-code-assist bot Jan 11, 2026

Uh oh!

gemini-code-assist bot Jan 11, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants



		class ExampleLearn2Ask(Workflow):
		name: str = "math_agent_workflow"

	name: str = "math_agent_workflow"
	name: str = "learn2ask_langchain_workflow"

Add math and learn-to-ask langchain demos #2

Add math and learn-to-ask langchain demos #2

Uh oh!

Conversation

Kanaricc commented Jan 11, 2026

Uh oh!

gemini-code-assist bot commented Jan 11, 2026

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist bot Jan 11, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist bot Jan 11, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist bot Jan 11, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist bot Jan 11, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants