Feature/OpenAI financial market

[Tejasv-Singh]

Summary

Introduced a functional financial market simulation that uses a Real LLM (OpenAI) for agent decision-making.

Motive

To address the feedback that the previous "Mock LLM" example was insufficient for showcasing Mesa's capabilities. This update demonstrates a production-ready pattern for integrating real AI APIs (OpenAI) into Mesa agents, allowing for true natural language processing and sentiment analysis within the simulation.

Implementation

• Real AI Integration: Implemented FinancialLLM in model.py which interfaces with the OpenAI API.
• Security: Uses os.getenv("OPENAI_API_KEY") to safely load API keys from the environment to prevent accidental commits.
• Mesa 3.0 Compatibility: Ensuring the example uses modern Mesa patterns:
  • Replaced deprecated RandomActivation with self.agents.shuffle_do("step").
  • Corrected super().__init__() usage for Agent and Model classes.
• Dependencies: Added openai to requirements.txt.

Usage Examples

1. Set your API Key:

   export OPENAI_API_KEY="sk-..."
   

Run the simulation:

bash

python examples/financial_market/run.py

Additional Notes

Gracefully handles missing API keys by stopping the simulation with a clear error message instead of crashing.
Configured with gpt-3.5-turbo and low temperature for cost-effective and consistent testing.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[Tejasv-Singh]

This PR supersedes #52. Based on the feedback received, I have replaced the mock LLM with a fully functional OpenAI API client.

[Tejasv-Singh]

I noticed the CI checks are failing, but it appears to be a Codecov upload error (Token required - not valid tokenless upload) caused by the PR coming from a fork.

The actual tests seem to have generated the coverage report successfully before this upload step failed.

I have also pushed a small update to use a 'lazy import' for OpenAI, ensuring the example doesn't crash environments where the library isn't installed.

[colinfrisch]

Thanks for your work, I'll review it asap ! We are indeed having a problem with codecov currently and it's being handled externally to this PR, so you don't have to worry abt it

[colinfrisch]

It's a good start ! But you should spend a bit more time on looking at the examples and really looking at the docs if you want to contribute. There are a few things that you rebuild that already exist in the repo, as well as a few good practices explained in mesa and mesa-llm tutorials that you ignore here.

[colinfrisch]

File is a bit short, you can check out examples folder in the mesa-llm repo as what you can do for a readme.

[colinfrisch]

No need for a requirements file. Just indicate them directly in the readme

[colinfrisch]

Please separate model and agents in different files like done in mesa-llm examples.

[colinfrisch]

Making a new client variable defeats the whole purpose of the Built-In ModuleLLM class... do you think that you could use it instead ?