Pin the numpy dependency for koverage env

[abhi18av]

Hi @Vini2 ,

Congratulations on the great tool!

I am trying to integrate reneo into everest-nf and came across an issue during runtime

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

It seems that the problem might be the fact that pandas relies on a minimal version of numpy but not the maximum and since there are some breaking changes in numpy 2.0.0 the error propogates further up.

Reference: https://stackoverflow.com/a/78641304/6482586

CC @agudeloromero

[abhi18av]

Actually, I've just faced another issue with koverage - perhaps @beardymcjohnface might be able to provide a more stable fix for the koverage release.

The error I have now come across is related to jupyter

    raise DPClientError(f"{type(b.data).__name__} not supported for {self.__class__.__name__}")
datapane.client.exceptions.DPClientError: DataFrame not supported for XMLBuilder
Please run with `dp.enable_logging()`, restart your Jupyter kernel/Python instance, and/or visit https://www.github.com/datapane/datapane

[Vini2]

Hi @abhi18av,

Thanks for your interest in Reneo and thanks for raising this issue and putting a PR. I'll pin the numpy version in reneo.yaml as well and publish a new release.

@beardymcjohnface can you please take a look at this issue?

Thanks!

[beardymcjohnface]

@Vini2 if you want a quick fix and you don't need the koverage HTML report, then I would add --no-report to the koverage step in rule koverage_genomes. Unfortunately, generating the report uses a few libraries that can be flakey. I just ran it fine on a fresh install on our VM but the error has popped up before.

[Vini2]

Thanks @beardymcjohnface! I'll add the --no-report flag for now to publish a new release.

[abhi18av]

Thank you both for the quick progress here!