noticed that Monarch was using prefixcommons via curies package which is no longer updated and does not have a prefix entry for "FB" which is the flybase preferred prefix. Updated that code to use merged.csv from this repo instead, and that change generated errors on the "action" prefix (as we had both "action" and "actions" that pointed to the same namespace, but only actions was declared as a canonical prefix. action was only declared as a namespace_alias).
this PR updates a few of the bioregistry-related dependencies and regenerates the extended prefixmaps. An artifact of the "refresh" target running through tox is that we need the pydantic 1.0 pin in the pyproject.toml to force tox to use pydantic 1.0. If there is a better way to get tox to install the exact same environment as its parent poetry environment (besides moving the refresh command from being executed with tox to being executed with poetry -- which might be the better approach), I am happy to try it instead of pinning. This was a confusing thing to debug.
https://github.com/biopragmatics/bioregistry/blob/d35cb98c54c15e2362ff0d92b99494284078718a/src/bioregistry/utils.py#L25 seems to be the spot where pydantic 1 is required.
https://github.com/monarch-initiative/monarch-app/issues/383
noticed that Monarch was using prefixcommons via curies package which is no longer updated and does not have a prefix entry for "FB" which is the flybase preferred prefix. Updated that code to use
merged.csvfrom this repo instead, and that change generated errors on the "action" prefix (as we had both "action" and "actions" that pointed to the same namespace, but onlyactionswas declared as acanonicalprefix.actionwas only declared as anamespace_alias).this PR updates a few of the bioregistry-related dependencies and regenerates the extended prefixmaps. An artifact of the "refresh" target running through tox is that we need the pydantic 1.0 pin in the pyproject.toml to force tox to use pydantic 1.0. If there is a better way to get tox to install the exact same environment as its parent poetry environment (besides moving the
refreshcommand from being executed with tox to being executed with poetry -- which might be the better approach), I am happy to try it instead of pinning. This was a confusing thing to debug. https://github.com/biopragmatics/bioregistry/blob/d35cb98c54c15e2362ff0d92b99494284078718a/src/bioregistry/utils.py#L25 seems to be the spot where pydantic 1 is required.