Closed kevinschaper closed 5 months ago
Tagging everyone here as reviewers to see how the modular ingests are getting hooked up to the kg build process right now. This is a really minimal solution, in particular because it doesn't have any kind of filtering support.
Longer term, we should either add the filtering support to cat-merge (maybe even using the kgx merge config format), or we should set up Koza to take biolink model class names to define both the input columns and the output columns, and have a sort of pass-through mode where the transform step just iterates over each row and the Koza filter code in the reader can be used.