Mistral Client

[marklysze]

This is a Mistral Client that provides a way to use the Mistral.AI API for an agent. It's available as 'MistralAIClient(similar toGeminiClient`).

This client provides some benefits over attempting to use the Mistral.AI API through the standard config as it can cater for some of the nuances of the Mistral.AI API such as not allowing the name key on a message or failing if a system role appears after an assistant role.

This client supports text generation and function/tool calling. Additionally, costs align with the Mistral.AI API costs.

This is my first client class, hopefully I've set everything up correctly. I based changes on the Gemini Client class.

Related issue number

Relates to PR #2748.

I haven't added any documentation at this stage, will do once it has been reviewed.

Checks

• I've included any doc changes needed for https://microsoft.github.io/autogen/. See https://microsoft.github.io/autogen/docs/Contribute#documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[gitguardian]

⚠️ GitGuardian has uncovered 4 secrets following the scan of your pull request.

Please consider investigating the findings and remediating the incidents. Failure to do so may lead to compromising the associated services or software components.

🔎 Detected hardcoded secrets in your pull request

GitGuardian id	GitGuardian status	Secret	Commit	Filename
10493810	Triggered	Generic Password	39b7c1e	notebook/agentchat_pgvector_RetrieveChat.ipynb	View secret
10493810	Triggered	Generic Password	39b7c1e	notebook/agentchat_pgvector_RetrieveChat.ipynb	View secret
10493810	Triggered	Generic Password	39b7c1e	notebook/agentchat_pgvector_RetrieveChat.ipynb	View secret
11616921	Triggered	Generic High Entropy Secret	39b7c1e	notebook/agentchat_agentops.ipynb	View secret

🛠 Guidelines to remediate hardcoded secrets

1. Understand the implications of revoking this secret by investigating where it is used in your code.
2. Replace and store your secrets safely. Learn here the best practices.
3. Revoke and rotate these secrets.
4. If possible, rewrite git history. Rewriting git history is not a trivial act. You might completely break other contributing developers' workflow and you risk accidentally deleting legitimate data.

To avoid such incidents in the future consider

• following these best practices for managing and storing secrets including API keys and other credentials
• install secret detection on pre-commit to catch secret before it leaves your machine and ease remediation.

---

^{🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.}

[Josephrp]

would be nice to make a toy notebook or app in examples and a short blog post about it , to accompany the next release : this is cool ! & thanks for the contribution !

[marklysze]

i think, instead of adding mistral into the autogen dependencies, let the user have it installed manually if working with non-openai models as mentioned in the https://microsoft.github.io/autogen/docs/topics/non-openai-models/about-using-nonopenai-models/

cc @ekzhu

Hey @Hk669, yes, you make a good point - my intention is to do that and have users choose it in the same way as the GeminiClient, whereby it will install Mistral's API library if you specify it:
pip install pyautogen[mistral]

So similar to what is shown here.

If I've not done that correctly, can you let me know.

[yiranwu0]

I did the following changes:

• I changed the mock in test_create_response_with_tool_call, basically we want to mock only the line mistral_response = client.chat(**mistral_params) in create function, so all other parts in your create function will still be run.
• Did some cleaning in mistral.py, just moving things around.
• I also remove the try catch around mistral_response = client.chat(**mistral_params), catch the error and then throw it is not necessary.

I have two minor questions, other than that I am good!

[marklysze]

I did the following changes:

• I changed the mock in test_create_response_with_tool_call, basically we want to mock only the line mistral_response = client.chat(**mistral_params) in create function, so all other parts in your create function will still be run.
• Did some cleaning in mistral.py, just moving things around.
• I also remove the try catch around mistral_response = client.chat(**mistral_params), catch the error and then throw it is not necessary.

I have two minor questions, other than that I am good!

Brilliant, thanks so much for your help on that! I've just been testing and it's working with the sample AutoGen workflows I tried before.

Let me check those two questions out now.

Sorry, @yiranwu0, would you be able to link me to the two questions you had? I can't find them.