Open pgierz opened 2 years ago
hey @pgierz! thanks for your input! Installing psy-view without conda-forge is quite difficult. not because of psy-view itself, but rather because auf the dependencies (particularly pyqt
, cartopy
, numpy
and pandas
. They are extremely difficult to install if you want to do this yourself.
I do not yet know about the plans for conda-forge to support the M1 ARM Chips but I strongly assume that they are working on this? No idea.
I'm happy to look I to it if some points me at the packaging repo.
not sure what you mean exactly. The source for the python package is in this repository, the source for the conda-package is in https://github.com/conda-forge/psy-view-feedstock/
My apologies that I cannot be very helpful here
Hey @Chilipp:
The worst one to install "by hand" (so, via pip) for me has always been cartopy. I managed this once on accident and still don't know what magic I needed in the background. Another strong point for the "take better notes" argument ;-)
For conda publishing: there normally was a separate repo that was required to upload the project to conda forge, so that's exactly the link I needed to get to look at the current details, thank you! While I have only a little bit of experience in that arena, I will try to work out the packaging details and will send you a PR with M1-ARM support if I can figure it out.
Have a nice evening! Paul
awesome! thanks @pgierz and good luck!
Well that was surprisingly straightforward. I just needed to make a PR in the Conda-forge repo: https://github.com/conda-forge/conda-forge-pinning-feedstock/pull/2999
I think once it goes in that is all there is to do. I will test as soon as that happens and let you know :-)
Moin,
I was trying to follow the conda install demo, yet on new Apple computers the conda-forge channel does not provide dependencies. I'm happy to look into it if someone points me at the packaging repo.
All the best, PG