scikit-umfpack / scikit-umfpack

The umfpack scikit provides wrapper of UMFPACK sparse direct solver to SciPy.
BSD 3-Clause "New" or "Revised" License
45 stars 28 forks source link

umfpack not found in conda-forge; Where can I install it from in conda? #85

Open souvikmeister opened 1 year ago

souvikmeister commented 1 year ago

Hi, I am trying to re-install umfpack in a new set up. I see from this link that conda-forge is the channel to use. However, doing conda install -c conda-forge scikit-umfpack returns



  - scikit-umfpack

Current channels:

  - https://conda.anaconda.org/conda-forge/win-64
  - https://conda.anaconda.org/conda-forge/noarch
  - https://repo.anaconda.com/pkgs/main/win-64
  - https://repo.anaconda.com/pkgs/main/noarch
  - https://repo.anaconda.com/pkgs/r/win-64
  - https://repo.anaconda.com/pkgs/r/noarch
  - https://repo.anaconda.com/pkgs/msys2/win-64
  - https://repo.anaconda.com/pkgs/msys2/noarch
  - https://conda.anaconda.org/anaconda/win-64
  - https://conda.anaconda.org/anaconda/noarch```

Also, I did `conda search -c conda-forge *umf*` which returns
```Loading channels: - DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
/ DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/win-64/repodata.json HTTP/1.1" 304 0
- DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/noarch/repodata.json HTTP/1.1" 304 0
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/noarch/repodata.json HTTP/1.1" 304 0
\ DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/msys2/noarch/repodata.json HTTP/1.1" 304 0
/ DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/win-64/repodata.json HTTP/1.1" 304 0
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/msys2/win-64/repodata.json HTTP/1.1" 304 0
done
# Name                       Version           Build  Channel
django-enumfields              0.8.2          py27_0  conda-forge
django-enumfields              0.8.2          py34_0  conda-forge
django-enumfields              0.8.2          py35_0  conda-forge
django-enumfields              0.9.0          py27_0  conda-forge
django-enumfields              0.9.0          py35_0  conda-forge
django-enumfields              0.9.0          py36_0  conda-forge
django-enumfields             0.10.0          py27_0  conda-forge
django-enumfields             0.10.0       py27_1000  conda-forge
django-enumfields             0.10.0          py35_0  conda-forge
django-enumfields             0.10.0          py36_0  conda-forge
django-enumfields             0.10.0       py36_1000  conda-forge
django-enumfields             0.10.0       py37_1000  conda-forge
django-enumfields              1.0.0       py27_1000  conda-forge
django-enumfields              1.0.0       py36_1000  conda-forge
django-enumfields              1.0.0       py37_1000  conda-forge
django-enumfields              1.0.0       py38_1000  conda-forge
quantumflow                    1.2.1    pyhd8ed1ab_0  conda-forge
quantumflow                    1.3.0    pyhd8ed1ab_0  conda-forge
r-numform                      0.5.0   r36h6115d3f_0  pkgs/r
r-numform                      0.7.0   r42h142f84f_0  pkgs/r ```

confirming either the channel is not there or my conda-forge channel is not seeing it. Is there anything else I need to do to see this package? I confirmed by doing a search on jupyterlab that I am able to see that package in conda-forge. Otherwise, if I am supposed to add this from a different channel, then please advise.
MTIA.
rgommers commented 11 months ago

Works fine for me with conda-forge:

mamba search scikit-umfpack
Loading channels: done
# Name                       Version           Build  Channel
scikit-umfpack                 0.3.3 py310h4d09554_0  conda-forge
scikit-umfpack                 0.3.3 py310h6e7f00a_1  conda-forge
scikit-umfpack                 0.3.3 py310h8ddc3d6_1  conda-forge
scikit-umfpack                 0.3.3 py311h809b3aa_1  conda-forge
scikit-umfpack                 0.3.3 py311hefe60d2_1  conda-forge
scikit-umfpack                 0.3.3  py38h76083e6_1  conda-forge
scikit-umfpack                 0.3.3  py38h9e4b79f_1  conda-forge
scikit-umfpack                 0.3.3  py39h08fb805_1  conda-forge
scikit-umfpack                 0.3.3  py39h1608c9b_1  conda-forge

There isn't a package for Python 3.12 yet though: https://github.com/conda-forge/scikit-umfpack-feedstock/pull/36.

mhvwerts commented 2 months ago

Just as a reminder to those trying to install scikit-umfpack in Windows: the package appears to be only available for Linux and OSX in Conda-Forge: https://anaconda.org/conda-forge/scikit-umfpack

SuiteSparse, on which scikit-umfpack depends, is available for Windows in Conda-Forge: https://anaconda.org/conda-forge/suitesparse

rgommers commented 2 months ago

I'll note that Fortran on Windows is a pain, but hopefully it'll get a lot easier in the near future with the default compiler becoming flang-new so we have a full set of LLVM-based compilers on Windows.

mhvwerts commented 2 months ago

Thanks for this information. I'll stay tuned! Also @simulkade will be interested to hear this.

For now, we may try certain demanding calculations on a Linux system. Also, from a practical point of view, I am quite happy with the performance of Pypardiso on Windows for solving our current sparse matrix equations.