Closed vc1492a closed 4 years ago
Identified the source of the ZeroDivisionError as AuthorRank scores being identical in some rare edge cases. When this situation occurs, authors are now given an identical score of 1.0. It is up to users of AuthorRank to interpret how to use this information in their analyses / products / visualizations.
I'll be merging this PR into the dev
branch.
There's a slight penalty in the level of coverage. It seems that the
ZeroDivisionError
handler is never invoked inutils.py
as part of the tests. Looking into it.