Checking all installs

[bioforkgithub]

python3 -c 'import sklearn' NOT OK
Missing python library "sklearn"

I have installed the scikit-learn package but still it shows this error on running the test_install.pl

[bioforkgithub]

I tried both the packages mamba as well as the git install

[fpusan]

What's the output when running python3 -c 'import sklearn' inside the activated conda environment?

[bioforkgithub]

1. micromamba list gives following:

scikit-learn 0.24.2 py36hc89565f_1 conda-forge
scipy 1.5.3 py36h81d768a_1 conda-forge
sed 4.8 he412f7d_0 conda-forge
setuptools 58.0.4 py36h5fab9bb_2 conda-forge
six 1.16.0 pyhd3eb1b0_1
sqlite 3.40.1 h5082296_0
squeezemeta 1.6.2post1 py36pl5262r36h7b7c402_0 fpusan

2. python3 -c 'import sklearn' gives following:

(SqueezeMeta) [a40540@hod localscratch]$ python3 -c 'import sklearn'
Traceback (most recent call last):
File "", line 1, in
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/sklearn/init.py", line 82, in
from .base import clone
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/sklearn/base.py", line 17, in
from .utils import _IS_32BIT
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/sklearn/utils/init.py", line 23, in
from .class_weight import compute_class_weight, compute_sample_weight
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/sklearn/utils/class_weight.py", line 7, in
from .validation import _deprecate_positional_args
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/sklearn/utils/validation.py", line 22, in
import joblib
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/joblib/init.py", line 129, in
from .parallel import Parallel
File "/gpfs/gpfs0/home/home2/a40540/micromamba/envs/SqueezeMeta/lib/python3.6/site-packages/joblib/parallel.py", line 23, in
from contextlib import nullcontext
ImportError: cannot import name 'nullcontext'

Even with the error I tried running the pipeline with test data. It ran without error but does it affect the credibility of result???????

[fpusan]

Not sure which package uses scikit, your results are probable fine if there was no error message.
Anyways I am working on a Foix for this
You can try mamba create -n SqueezeMeta -c conda-forge -c bioconda -c fpusan -c anaconda squeezemeta-dev=1.6.2post2 --no-channel-priority to install the de versiin with the Fox. It should pass the tests in test_install.py. I will do more testing myself and hopefully release the patch next week

[bioforkgithub]

Ok then I will do that next week and try again on the test data. Rightnow I am running it on my data. Waiting for any errors.
Will let you know for any error then.

[fpusan]

Yes, I have fixed this in the dev version, install with  mamba create -n SqueezeMeta -c conda-forge -c bioconda -c anaconda -c fpusan  squeezemeta-dev=1.6.2post2 --no-channel-priority And then configure with configure_nodb.pl /path/to/db I will make an official release later this week after I finish testing, but everything seems to be in order so far... Sobre 11/07/2023 15:32:22, RIJP ***@***.***> escribió: Hi, I've got same message from test_install.pl. I think version of scikit-learn is not compatible for python 3.6. When installing squeezemeta with mamba, scikit-learn version 0.24.2 is installed. However this version is not compatible for python 3.6. I tried downgrading scikit-learn to 0.20.2 which is compatible for python 3.6, but failed due to dependency problems. Any solution for this? — Reply to this email directly, view it on GitHub [#705 (comment)], or unsubscribe [https://github.com/notifications/unsubscribe-auth/ACAHO7X7HI22KLD65K6SE53XPVIWNANCNFSM6AAAAAAZ55I4PQ]. You are receiving this because you commented.Message ID: ***@***.***> [33aa4013-75ba-461c-8481-6101d19df998]

[RIJP-SU]

Sorry I should have cheked your dev version befor posting. It worked fine. Thanks.

*I accidentally deleted follwoing message. Applogize for making mess.

Hi, I've got same message from test_install.pl. I think version of scikit-learn is not compatible for python 3.6. When installing squeezemeta with mamba, scikit-learn version 0.24.2 is installed. However this version is not compatible for python 3.6. I tried downgrading scikit-learn to 0.20.2 which is compatible for python 3.6, but failed due to dependency problems. Any solution for this?

[fpusan]

No, it's fine. Dev versions are not normally meant for downloading, sometimes we have some non-working experimental stuff in there

[fpusan]

Should be fixed in ba9a595 and the new release v1.6.2post2

[bioforkgithub]

The execution is complete but i did not get any error messages for scikit. Seems everything is fine. This is from the previous install,not the updated one you mentioned. Will try the next one as well. If any errors, i will keep you posted

[fpusan]

Even if you got no error messages, CONCOCT binning (if requested) may have failed. In that case the pipeline would have progressed using the results from the other binning methods