The original version used remote taxon validation to give users a way to:
Avoid the file space needed for the largish local taxon database (1GB+)
Avoid having to build that database
However, the codebase now:
Supports database building with well scripted endpoints.
Has moved to clear timestamping of taxon DB versions (not so much of an issue for GBIF, which has stepped updates, but a big issue for NCBI which has rolling updates).
The file sizes are reasonably big, but not overwhelming for a local user (currently ~ 2GB for both NCBI and GBIF). There are also serious downsides to maintaining the remote validators:
Continual tweaking of the test database to ensure that the local and remote test data are in sync.
Significant runtime overhead for local testing and CI.
Code complexity in both the testing and main codebase.
Removing remote significantly improves maintenance for little cost and only provides support for extremely niche use cases.
The original version used remote taxon validation to give users a way to:
However, the codebase now:
The file sizes are reasonably big, but not overwhelming for a local user (currently ~ 2GB for both NCBI and GBIF). There are also serious downsides to maintaining the remote validators:
Removing remote significantly improves maintenance for little cost and only provides support for extremely niche use cases.