Orientation sampling fails - outdated spherecluster pip version fails because of sklearn deprecation
Describe the bug
When sampling from a von Mises Fisher Distribution (fishdist.py) in GemPy the, code breaks since sklearn.externals.joblib was removed from sklearn in version 0.23(.1) (currently the lastest version). The package spherecluster (master updated 2 years ago) is still dependent on this package.
To Reproduce Run a model with scikit-learn==0.23 and try to sample from a vMF distribution.
Expected behavior Sampling from the vMF distribution should not result in an error. It seems like the development branch of spherecluster does not have this dependency anymore and was updated 'only' 6 months ago. So this could work. What worked for me now is downgrading sklearn. In the long term, we could try to implement our own solution expanding the previous work of @elimh to remove the dependency on spherecluster. Could be interesting for @Japhiolite as well in combination with #449.
Desktop (please complete the following information):
- OS: Win 10
- GemPy Version
- 2.1.1
Additional context Add any other context about the problem here.