Closed cthoyt closed 10 months ago
Attention: 25 lines
in your changes are missing coverage. Please review.
Files | Coverage Δ | |
---|---|---|
src/bioregistry/__init__.py | 100.00% <ø> (ø) |
|
src/bioregistry/record_accumulator.py | 91.41% <100.00%> (+0.27%) |
:arrow_up: |
src/bioregistry/uri_format.py | 91.66% <100.00%> (+5.00%) |
:arrow_up: |
src/bioregistry/resource_manager.py | 75.15% <66.66%> (-0.08%) |
:arrow_down: |
src/bioregistry/external/prefixcommons.py | 21.35% <14.28%> (+0.14%) |
:arrow_up: |
src/bioregistry/external/miriam.py | 31.34% <12.50%> (-1.45%) |
:arrow_down: |
src/bioregistry/external/n2t.py | 43.90% <22.22%> (-6.10%) |
:arrow_down: |
:loudspeaker: Thoughts on this report? Let us know!.
This PR gets rid of code that focuses on lists of
curies.Record
objects and instead works directly withcuries.Converter
objects.Along the way, this also identified issues with the data integrity on MIRIAM, N2T, and Prefix Commons with respect to the TAIR resources (
tair.gene
andtair.protein
) which all used non-specific, overlapping URLs. Therefore, these needed to get cleaned out before being import.Why do this? If we work directly with converters, we can make use of the CURIE prefix reconciliation tooling to more cleanly refactor the Bioregistry to Converter pipeline (which is causing issues when adding prefix casing variants in a related PR #969)