July 26, 2024
Imagine being able to compare your code base against hundreds of pages of architectural best practices - getting advice on where to invest effort and what to improve.
Thatβs exactly what I did with Gemini and the Google Cloud architecture framework!Β To use it, you just need to pass the arguments of the model you want to use, and the repo you want to analyse! Hereβs it is running on Cloud Run:
Now, of course I pointed it at my own repo and the results are fascinating. Let me quote you parts of it:
*Secret Management:** The code uses Secret Manager to store the GitHub PAT, demonstrating good practice for sensitive information management.
*Infrastructure as Code:** While the code doesn't explicitly demonstrate IaC, the use of cloudbuild.yaml for deployment hints at a potential for incorporating IaC tools like Terraform. This would further enhance automation and reproducibility.
*Implement Testing:** Add unit tests for individual functions in gemini_bits.py and github_to_text.py. Implement integration tests to verify interactions between these modules.
You get the point. Itβs actually really good!
The real magic lies in Gemini, its large context window, and what I pass to it.
It's like a supercharged architect who's read every page of the Google Cloud Architecture Framework. It analyses my code through the lenses of Operational Excellence, Security, Reliability, Cost Optimisation, and Performance Optimisation/
Intrigued?
Dive into my code and see how it's evolving: [Link to Code Repo]
Explore the Google Cloud Architecture Framework for yourself: [Link to Framework]
Itβs not just Gemini Code Assist, or Co-pilot. Itβs how you can interact with your code in really new ways.. Share your thoughts below! π¬