search

chklovski / CheckM2

Assessing the quality of metagenome-derived genome bins using machine learning
GNU General Public License v3.0
165 stars 19 forks source link

Unable to install scikit-learn==0.23.2 #49

Open sjaenick opened 1 year ago

sjaenick commented 1 year ago

Hi,

on Ubuntu 20.04 LTS, I'm unable to install CheckM2 because the outdated scikit-learn fails to compile. This is with Python 3.11.2, and conda or Docker currently isn't an option for us.

            creating build/temp.linux-x86_64-3.11/scipy/cluster
            INFO: compile options: '-I/tmp/pip-build-env-l2e7hsoe/overlay/lib/python3.11/site-packages/numpy/core/include -I/tmp/pip-build-env-l2e7hsoe/overlay/lib/python3.11/site-packages/numpy/core/include -Ibuild/src.linux-x86_64-3.11/numpy/distutils/include -I/vol/mgx-sw/include/python3.11 -c'
            extra options: '-msse -msse2 -msse3'
            INFO: gcc: scipy/cluster/_vq.c
            scipy/cluster/_vq.c:196:12: fatal error: longintrepr.h: No such file or directory
              196 |   #include "longintrepr.h"
                  |            ^~~~~~~~~~~~~~~
            compilation terminated.
chklovski commented 1 year ago

Unfortunately the scikit-learn version is a hard requirement due to incompatible version-changes to pickled data scalers checkm2 loads. I am looking into getting rid of that, but for the moment conda is the only alternative solution.