Pandas 2.1.0 does not support Python 3.8

[ajnelson-nist]

https://pandas.pydata.org/pandas-docs/version/2.1.0rc0/whatsnew/v2.1.0.html#increased-minimum-version-for-python

Edited to add, 2023-08-28: case_sparql_select makes use of Pandas for table formatting code. Pandas dropping Python 3.8 thus means case-utils currently can't satisfy a dependency unless using Python 3.9 or later.

[ajnelson-nist]

The resolutions are either:

1. Bump case-utils' minimal Python version to 3.9.
2. Pin pandas to be < 0.21.0.

[ajnelson-nist]

For extra context, Python 3.8's end of support is October, 2024, per:
https://www.python.org/downloads/

[ajnelson-nist]

For further context, the pandas Issue describing the minimal-version bump is here:
pandas-dev/pandas#52513

[ajnelson-nist]

Another option, though I'm not sure syntactically how to accomplish this, is to consider pandas an optional requirement, tying it to a new feature that provides case_sparql_select. (pandas was brought in to add several tabular output forms to that script.) While I'm aware of how to make features with their own dependent packages, I'm a documentation dive away from knowing how to pin the minimal supported Python version for just one feature.

So, I still suggest we have a raise-Python-floor vs. lower-Pandas-ceiling decision.

[ajnelson-nist]

Resolved with #128 .