Updating the biolink predicate being used. This is a bit tricky b/c the current predicate in 2.2.7 is actually biolink:co-occurs_in_literature_with which fails to trapi validate. The predicate being used here is a future one (2.2.8) but it's not going to cause any trouble.
Omnicorp was returning both a num_publications and a publications: [] attribute. The empty publications was causing trouble in weighting.
Getting the normalization values was completely wrong.
biolink:co-occurs_in_literature_with
which fails to trapi validate. The predicate being used here is a future one (2.2.8) but it's not going to cause any trouble.