Automatic sort argument for SQLAlchemyInterface

[gbunkoczi]

When using SQLAlchemyConnectionField with a SQLAlchemyInterface to get a polymorphic connection, one has to set the sort argument to None to disable automatic sort enum generation. However, the type check in register_sort_enum is too strict, and being an instance of SQLAlchemyBase seems to be sufficient for the code to work.

Tests run with Python 3.8 and 3.12.

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (9c2bc84) 94.74% compared to head (f5e6d32) 94.74%.

:exclamation: Current head f5e6d32 differs from pull request most recent head 3ade1e6. Consider uploading reports for the commit 3ade1e6 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #400   +/-   ##
=======================================
  Coverage   94.74%   94.74%           
=======================================
  Files          10       10           
  Lines        1333     1333           
=======================================
  Hits         1263     1263           
  Misses         70       70           

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[gbunkoczi]

Apologies for the delay. I put some tests in place.

Somewhat tangentially, there seems to be more cases of "issubclass(obj_type, SQLAlchemyObjectType)" tests in the codebase, but since these were not triggered so far, I have not changed them. In addition, "- obj_type : SQLAlchemyObjectType" appears multiple times in function docstrings where " -obj_type: SQLALchemyBase" would be more appropriate, e.g. https://github.com/graphql-python/graphene-sqlalchemy/blob/9c2bc8468f4c88fee2b2f6fd2c0b8725fa9ccee3/graphene_sqlalchemy/enums.py#L94 https://github.com/graphql-python/graphene-sqlalchemy/blob/9c2bc8468f4c88fee2b2f6fd2c0b8725fa9ccee3/graphene_sqlalchemy/enums.py#L166

(these call registry.register_sort_enum, which has been switched to test SQLALchemyBase). I am happy to correct these as part of this PR if this would help.

[erikwrede]

@gbunkoczi Awesome! if you have time, please feel free to go ahead. Thanks for your effort! 😊

[gbunkoczi]

The checks have all been changed. I tried each call with a SQLAlchemyInterface instance, and all worked as expected. The docstrings were also updated when appropriate.