Closed bmwoodruff closed 2 weeks ago
I'm currently generating examples for the last almost 100 functions that I haven't fun the generator on, minus the np
namespace. that one has 300 functions, and I'll probably batch it tonight before I go to bed. Then I'll push the rest of the logs up. We can focus on reviewing the output, and finding a good way to automate that. I think it's completely fair to separate the AI generation portion from the processing output portion of the project.
The tracking list logs are useful to store in the repo, as they make it easy to grab information in other places. They can be generated from list_callables.py, but I felt like leaving them in the repo was useful.
The logs from running example-generator.py take 2-10 minutes each to generate, so storing them in the repo makes perfect sense.
I'm almost done running the example-generator on every function in numpy except those in the base np space.