error in component_binning.py

[heleeno]

Hi,

I'm trying to run metabinner on coassembly data and it seems that I successfully created a coverage file and kmer file but I'm running into an error, see below. I have tried to remove the _ from "sklearn.cluster._kmeans" because I thought this may be the problem but I get the same error.
Do you maybe know what is going on?

Thank you, Heleen

Traceback (most recent call last):
File "./component_binning.py", line 23, in
from sklearn.cluster._kmeans import euclidean_distances, stable_cumsum, KMeans, check_random_state, row_norms, MiniBatchKMeans
ImportError: No module named _kmeans

real 0m0.888s
user 0m0.582s
sys 0m1.142s
Something went wrong with generating component binning results. Exiting.

[ziyewang]

Hi,

It seems that the version of scikit-learn in your environment is different from that we used. Could you please check the version of scikit-learn and make sure it equals 0.22.1?

Best,
Ziye

[heleeno]

Hi Ziye,

Thanks for your reply. I checked and the scikit-learn version in my environment is 0.22.1. Can you think of anything else it could be?

Best,
Heleen

[ziyewang]

Hi, Heleen,

It seems that only sklearn versions older then 0.22 cannot import sklearn.cluster._kmeans. Could you please add "import sklearn" and "print(sklearn.version)" before the "from sklearn.cluster._kmeans import euclidean_distances, stable_cumsum, KMeans, check_random_state, row_norms, MiniBatchKMeans" in the component_binning.py, and test whether the code successfully calls this sklearn version (0.22.1)?

Best,
Ziye