Open arkavo-com opened 2 months ago
Thank you for the suggestion. It’s an interesting idea, and I’ll consider it for future development. However, I can’t make any promises at this stage. Thanks again for your input!
Hello @arkavo-com I've just updated to version 2.1.0. This version includes a new code review feature. Please take a look at the link. This feature uses a large number of tokens, so it should be used with caution. This feature is still experimental and has many aspects that need to be refined. I'll keep this issue open for now.
tested. works well with llama3.1! nice job.
I am wondering how best to integrate with my workflow... first thought... Post commit code review:
That's an interesting feature... I'm not sure if it's appropriate to do a code review after commit. It might be better to perform a code review after git add
is triggered.
Also, this feature seems to be different from the original purpose of aicommt2, so I’m considering creating a new program for it.
Agreed on both points.
The thought around after a commit was assuming this workflow:
The local commit represents a unit of work that is ready for a code review, but not necessarily the complete feature or fix.
thinking more... I think the end goal is "AI-Enhanced Git Workflow".
perhaps this is the new program.
BTW, I think the feature you added is efficient for this ticket.
an update... I use this code review feature all the time now.. thank you!
Hi. I'm currently researching for watch mode development. As you said, once the git commit event occurs, I'm going to request a code review. This feature is very experimental, but it seems useful and interesting.
Feature request
I would like to propose adding a code review feature to aicommit2. This feature would allow users to leverage the integrated AI models to automatically review code changes before commits are finalized. The review process could analyze code quality, adherence to best practices, and even suggest improvements or highlight potential issues.
Why?
Currently, aicommit2 excels at generating commit messages using multiple AI models. However, code reviews are an essential part of the software development process to ensure quality and consistency. Integrating AI-driven code reviews directly within the CLI tool would enhance the overall development workflow by providing immediate feedback on code quality, potentially reducing the number of bugs and improving the maintainability of the codebase.
Alternatives
Some alternative solutions include using separate AI tools for code review or relying on manual code reviews through pull requests. However, these approaches can be time-consuming and less integrated with the current aicommit2 workflow.
Additional context
This feature could be implemented as an optional step that runs before the commit message generation. Users could configure the AI models to focus on specific aspects of the code, such as security, performance, or style. The feature could also be integrated into Git hooks, allowing automatic reviews for each commit.