Static analysis of python repo: Data flow between API Calls
Budget: $30 – $250 USD
I want to build a static analysis feature that takes a user's python code base and builds a data flow graph between LLM calls (e.g., OpenAI API calls). I used pyre (pysa) to do this and have a working approach where I put stub folders into the user's repo. However, I'd prefer not to add folders to the user's repo and just leave it untouched. I somehow wasn't able to figure this out.
It sounds trivial, and I'm probably just being an idiot, but I need to bootstrap a project quickly and would like to offload the task of coming up with a cleaner solution (i.e., that doesn't require modifying the user code base). I'm open to using other tools than pyre but I like pyre for their incremental update functionality.
I want to avoid you going down a rabbit hole and ultimately failing (like I did), so ideally, only apply if you are 100% confident that you can do it and did very similar things before / have extensive experience with pyre. I got similar things to work before and somehow couldn't figure this one out.
It sounds trivial, and I'm probably just being an idiot, but I need to bootstrap a project quickly and would like to offload the task of coming up with a cleaner solution (i.e., that doesn't require modifying the user code base). I'm open to using other tools than pyre but I like pyre for their incremental update functionality.
I want to avoid you going down a rabbit hole and ultimately failing (like I did), so ideally, only apply if you are 100% confident that you can do it and did very similar things before / have extensive experience with pyre. I got similar things to work before and somehow couldn't figure this one out.