scikit-tda / ripser.py

A Lean Persistent Homology Library for Python
http://ripser.scikit-tda.org
Other
276 stars 61 forks source link

Jupyter notebook crash on input of large sparse distance matrix #133

Open xgao32 opened 3 years ago

xgao32 commented 3 years ago

Hello, I have a distance matrix represented in COO format (I,J, value) that can be accessed here. There are 149K vertices and almost 1.7 million edges, the file is just 47 MB.

When I try to run Ripser on the whole distance matrix or a subset of it, my Jupyter notebook crashes.

image

Here is a minimal example to replicate the issue. The same holds when I run this on Google Colab with 12Gb of RAM and I suspect it may be related to #121 and this on the Ripser repo. Getting this to work would be vital to my research and I pray that someone can point me in the right direction .

ulupo commented 2 years ago

Hi! #132 might help you, but this has not been integrated yet into a PyPI release by the looks of it. Perhaps you can try installing the package from sources?