Mario Picciani

Also delete darglint from the dev dependencies then: https://github.com/cookiejar/cookietemple/blob/462ccf63bef17d9fdeb3884f2e94a88ba5703e13/cookietemple/create/templates/cli/cli_python/%7B%7B%20cookiecutter.project_slug_no_hyphen%20%7D%7D/pyproject.toml#L44

Additional information: Changing `legend_loc` doesn't have any effect here either.

Totally forgot about this one sorry :( **The problem** In `scatter_base`: https://github.com/theislab/scanpy/blob/040e61ff50836d4a6cdd7da7482dcb4ee50d05ae/scanpy/plotting/_utils.py#L736-L740 For non categorical variables, this code gets the current figure and adds a separate axis on which the...

> I would kindly ask you to also strongly consider implementing this. Having the edgeR and DEseq2 approaches being implemented here will also greatly boost the impact. I have no...

For other people who just want to get a solution / reason without debugging for 2 hours: This is an unfortunate combination of a breaking change and a misleading error...

Has been fixed here #130 but this is a breaking change since the entire API is reworked. I will leave this open until the next release and inform once we're...

Sorry I meant 1. 2 would probably be cleaner than 3 but other than that not really a different as in both casese input_data is a model attribute

@ilan-gold do we still need this? In my opinion, we can check all of this during model initialisation, that means during creation of an `InputDataGLM` object. Everything downstream is using...

I don't necessarily agree, because it doesn't only depend on the model but also on the specific "flavour". edgeR for example uses some fancy QR decomposition stuff to initialise the...

@davidsebfischer @LuckyMD This is quite an old issue. I'd like to close it since I've created an ipynb file for nb model fitting using simulated data which is currently in...