@ialarmedalien writes (regarding the extraction of credit metadata from the NMDC database):
I was just thinking of standard unit tests for the parser part — save some API responses and then run them through the code that parses the credit metadata to ensure that what is returned is what you’d expect to be returned. It doesn’t need to be anything fancy or end2end.
The JDP, NMDC(, etc) APIs are going to be evolving in the coming months, but I agree that we should have some coverage of the logic that extracts credit metadata.