Data files need to store the versions of the schema and data packages on which they depend. git.get_repo_metadata() obtains these versions from local git repos being used to create a data file.
Expand this to obtain version info from pip-installed instances of a package being used to create a data file. the package must contain a unique identifier that can be associated with a repo reference, such as a tag.
Possible approach
Prepare the package
- Finalize a package release
- commit it
- tag it
- write the tag into the release
- commit that
- push to git repo
- install on PyPI
Get version from the package
- read tag
- get package repo URL from pip show 'Home-page' line
- clone the repo
- use "git show-ref -s tag" to get the package version