search

DavidLapous / multipers

[NeurIPS2023,ICML2024] Multiparameter Persistence for Machine Learning
https://www-sop.inria.fr/members/David.Loiseaux/doc/multipers/index.html
MIT License
11 stars 1 forks source link

multipers has no attribute 'simplex_tree_multi' #11

Closed jmou19 closed 4 months ago

jmou19 commented 4 months ago

Hello,

I'm attempting to read through the _multipersintroduction notebook, and I am receiving an error when running the following block near line 27: 

# In practice, it is recommended to not use the sparse parameter, as there is not theoretical guarantees in this case.
(simplextree,), = mmp.PointCloud2SimplexTree(expand_dim=2,bandwidths=[0.2], num_collapses="full", sparse=.1, threshold=2, complex='rips').fit_transform([X])

The entire error is given below in form of a screenshot. image

I have version 1.2.2 of multipers installed. Please let me know if I can provide any further information about my set up to assist in diagnosing the problem.

jmou19 commented 4 months ago

I resolved the issue by running the recommended command at the top of the notebook:

 mp.simplex_tree_multi.SAFE_CONVERSION=True ## Enable this if you encounter crashes.
DavidLapous commented 4 months ago

Hmm this line shouldn't fix this issue, so there's still a bug somewhere, thanks for noticing it. I'm glad that you could fix this though :+1:

I'm currently rewriting the simplextree module, so I hope this will disappear with this 😇