Pathway sources should use INDRA-standard gene/protein IDs

[bgyori]

The pathway sources use a different ID scheme compared to other sources resulting in most (or maybe all) edges being dropped.

Wikipathway genes are represented as e.g., ncbigene:11303. These should ideally be normalized to HGNC for human genes and UniProt for non-human ones. Note that even if we expect these to be linked up to other nodes via xrefs, the prefix would still have to be EGID per the standard used in INDRA.

Reactome uses a mixture of other namespaces (e.g., chebi:28494, uniprot:A0A140T894) with a similar issue. UniProt IDs should be converted to HGNC for human proteins and to e.g., UP:A0A140T894 for non-human ones. chebi should probably be capitalized for it to work.