Closed mhvwerts closed 8 months ago
Thanks. I'll try it out.
Apparently, somewhere along the way, the dash in pytest-notebook
has become an underscore in Conda-forge: pytest_notebook
. Confusingly, two conda packages actually co-exists now:
The installation instructions for conda presently lead to installation of the obsolete version.
It might be a good idea to change the underscore back to a dash in the conda package name, if possible, such that the pip and conda packages have the same name.
ah stupid conda lol
Yeh this is because it needs to link to https://pypi.io/packages/source/p/pytest_notebook/pytest_notebook-v0.8.1.tar.gz; you know that pip normalises -
\ _
, so pip install pytest_notebook
is exactly the same as pip install pytest-notebook
, but conda does not normalize, so you have to do e.g. https://github.com/conda-forge/importlib_resources-feedstock/blob/fee669df78940da030671ad650eac5e99bb34cf8/recipe/meta.yaml#L37
I didn't know this subtlety. I'll stick with the underscore for now!
yep thats technically what its on pypi as: https://pypi.org/project/pytest_notebook/
I have been able to install using conda (more precisely: mamba... hard to keep up for slow people like me).
conda install pytest_notebook
(mind the underscore)
And now I can further explore this very useful program/package/plug-in. Great!
It may be wise to update the installation instructions in the documentation to have the package name with the underscore.
Also, the link to the Conda badge (as used in README.md
) would better be changed to point to the pytest_notebook
package. I do not know where the link to that badge is defined. The line in README.md
just reads [![Conda][conda-badge]][conda-link]
Anyways, it works for me now. Thanks for your help!
Many thanks for
pytest-notebook
. When installing it using conda/Conda-forge and trying it out, I ran into an incompatibility issue withnbconvert
(#10), which has been solved since version 0.7.0.It turns out that the present version on Conda-forge is 0.6.1 (over 2 years old). For now, I installed using
pip
, but it would of course be preferable to have a more recent version also on Conda-forge.