This PR represents a first step to reducing the scattering of external package
version numbers by mapping them to package variables like BASILISK_PANDAS_VERSION.
If this is successful, further systematization can be brought in with a function or object
that collects the versions suitable for a given release. Appropriate choices of external
package versions can be dependent on the miniconda version used.
External package versions for pandas, scipy, numpy etc. were the current release
versions as of 16 Jan 2023. Some external package specs are deliberately earlier
versions to demonstrate capacity to manage multiple versions of a given package.
One test that assesses ability to depend on python 2.7 was commented out to be moved
to longtests in a future PR.
This PR represents a first step to reducing the scattering of external package version numbers by mapping them to package variables like BASILISK_PANDAS_VERSION. If this is successful, further systematization can be brought in with a function or object that collects the versions suitable for a given release. Appropriate choices of external package versions can be dependent on the miniconda version used.
External package versions for pandas, scipy, numpy etc. were the current release versions as of 16 Jan 2023. Some external package specs are deliberately earlier versions to demonstrate capacity to manage multiple versions of a given package.
One test that assesses ability to depend on python 2.7 was commented out to be moved to longtests in a future PR.