Raf Guns

The doc string is actually fairly clear: https://github.com/rafguns/linkpred/blob/master/linkpred/predictors/base.py#L48. As it says > This is useful to, for instance, make sure that we only predict new links that are not currently...

Thanks for the report. I am aware of this but since there isn't a separate issue about it, I will keep this one open. The reason is that linkpred only...

The universe parameter is an iterable (typically a list or set) of all possible links (i.e. all node pairs) in the graph. Because the number of node pairs increases exponentially...

Looking at this again. The main issue is that linkpred uses its own data structure (`Scoresheet`) to track prediction scores. This has at least two advantages: 1. Order of nodes...

Other SimRank implementations: - [In Perl](http://cpansearch.perl.org/src/SHOHEIK/Graph-Similarity-v0.0.5/lib/Graph/Similarity/SimRank.pm): uses Jeh & Widom algorithm - [In C++](https://gist.github.com/powerswitch/2628214): appears to use Antonellis algorithm - [In Python](http://www.jontedesco.net/2013/02/01/simrank-in-python/): uses Jeh & Widom algorithm but ironically yields...

Yet another implementation: https://github.com/amir-f/delta-simrank

What version of networkx is this? Please make sure you have version 2.1 or more recent. Can you see if that solves the problem? Current code in networkx seems more...

Are you sure about networkx? I am asking because this line in your stack trace ``` File "/home/danial/Envs/graph/lib/python3.6/site-packages/networkx/convert_matrix.py", line 369, in to_numpy_matrix M = np.zeros((nlen,nlen), dtype=dtype, order=order) + np.nan ```...

Thanks for the report. This looks like a bug. Can you provide a minimal test case (small graph) where the problem occurs?

Thanks. Good to know that it can be solved using `numpy.log` but I woud like to understand the issue before applying a fix. So if you can supply an example,...