cluster_validity submodule in python missing

[lm-jkominek]

Hi, I've been using genieclust for some of my work recently, and absolutely love the efficiency of it - 10K rows and 10K columns, clustered in ~20 seconds, works fantastic. I wanted to try assessing the best number of clusters in my data, and the docs mention all the different indices to do that (Calinski-Harabasz, Silhouette etc), but for some reason I cannot run/see their respective functions in my Python install. In fact, it looks like the entire cluster_validity submodule is missing/hidden on my machine.

>>> import genieclust as gc
>>> dir(gc)
['GIc', 'Genie', '__builtins__', '__cached__', '__doc__', '__file__', '__loader__', '__name__', '__package__', '__path__', '__spec__', '__version__', 'compare_partitions', 'genie', 'inequity', 'internal', 'plots', 'tools']

I use Python 3.11.4, and tried genieclust v1.1.3, v1.1.4 as well as a direct github install of the latest dev version (v1.1.4.9001), with the same result. Any help or at least clarity on why that would be happening? Thanks a lot in advance!

[gagolews]

Ooops! I missed an import line in __init__.py! This is now corrected in the github version

If you want to stick with the PyPI version (I plan a new PyPI release in 1-2 months), import genieclust.cluster_validity will do, for now.

Thanks for the report!

[lm-jkominek]

Thanks a lot, appreciate the rapid response! I don't mind using the Github version for now, will just update to v1.1.5 when it's out.