search

haesleinhuepf / human-eval-bia

Benchmarking Large Language Models for Bio-Image Analysis Code Generation
MIT License
20 stars 13 forks source link

Are OME-Zarr tests in scope #98

Open jluethi opened 2 months ago

jluethi commented 2 months ago

@haesleinhuepf given your call for contributing tests: Are additional small tests on extracting information from OME-Zarrs in scope for this repo? I'd be happy to add some to encourage LLMs to learn about NGFF ;)

I saw you test the very basic opening of Zarr files here: https://github.com/haesleinhuepf/human-eval-bia/blob/main/test_cases/open_zarr.ipynb

If similar tests for reading labels from an OME-Zarr, reading a tiny image and e.g. summing up its intensities etc. are of interest, I can come up with a tiny example OME-Zarr that can fit in the repo and working on some notebook tests.

haesleinhuepf commented 2 months ago

Absolutely! Please send a pull-request for such test cases! Consider writing small test cases testing one single functionality - and add a complex workflow involving multiple steps.

Great examples using ngff and zarr are very welcome, also because I couldn't write them 😃