Open hexylena opened 1 month ago
@hexylena What do you get when you try to install a package ? I tried it on main with the package multivelo, and it worked fine.
Yes, this should work? Which Galaxy and which Jupyter are you using?
It should also work for Rstudio.
It should also work for Rstudio.
Like here, for example: https://training.galaxyproject.org/training-material/topics/evolution/tutorials/mtb_phylogeny/tutorial.html#visualize-and-manipulate-the-tree
huh interesting that it works for you!
after conda installing python-levenshtein
it fails to become accessible;
Which Galaxy and which Jupyter are you using?.
eu of course, and jupylab
import Levenshtein
perhaps?
also fails. testing multivelo now
and resetting my API key because that screenshot leaked my notebook URL which means anyone could access it ðŸ˜
honestly can't tell if this is better
yeah ok multivelo imports. thanks for the tip @Delphine-L
Don't know what the problem with the IT context is, but in case python-levenshtein is really what you're trying to get to work: make sure you install from conda-forge (not bioconda) to get the latest version (0.25), and import Levenshtein
(with the captial L) is definitely how you'd import it.
No, it's a different package, squidpy
, I was trying to produce a smaller/easier repro and guess I stumbled on a package that also works poorly with conda.
ImportError: /opt/conda/lib/python3.8/site-packages/zmq/backend/cython/../../../../.././libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by /opt/conda/lib/python3.8/site-packages/pandas/_libs/window/aggregations.cpython-38-x86_64-linux-gnu.so)
needs to be solved some other way I guess :grimacing:
fwiw, I try to fix Squidpy but got stuck at https://github.com/conda-forge/xarray-dataclasses-feedstock/pull/2
Thank you for trying @bgruening :(
@bgruening reports that it should be fixed https://anaconda.org/conda-forge/squidpy/files which is fantastic news.
@bgruening @Delphine-L I was pointed to this tutorial recently https://training.galaxyproject.org/training-material/topics/galaxy-interface/tutorials/galaxy-intro-jupyter/tutorial.html which suggests you can conda install a package and have it available but I cannot reproduce this, and the two mentioned packages seem to be preinstalked in the notebook.
is there a way to conda install something and have it available? when I use
!conda info
I see no environment activated so I would suspect that nothing will be on the path correctlycc @mbaardwijk