In general, the parsers appear to my eye to contain too much code that could be abstracted away. I would hope for much less code in each one, making it far easier to create these parsers. Everything that is not expressly required by an individual parser should be removed. Any code that appears in more than one parser should be removed.
IMO (and this is perhaps unreasonable) the goal should be to make these parsers be limited to describing some files or sources, and driving some very generic functions for creating KGX/biolink data.
Some of this complaint is more reasonably covered in #72.
This issue is now: figure out how much of the parsers can be replaced with a single parser driven by some source specific metadata.
In general, the parsers appear to my eye to contain too much code that could be abstracted away. I would hope for much less code in each one, making it far easier to create these parsers. Everything that is not expressly required by an individual parser should be removed. Any code that appears in more than one parser should be removed.
IMO (and this is perhaps unreasonable) the goal should be to make these parsers be limited to describing some files or sources, and driving some very generic functions for creating KGX/biolink data.