Fix: Default SRID is bypassed when using floating point coordinates.

[aballet]

Description

This is a very simple one-character fix. Since v0.15.0, the new SPLIT_WKT_PATTERN regex did not account for floating point coordinates. The match() method returns None, and the default_srid is never applied in format_geom_type().

How do you suggest testing for this? Current functional tests do not have any floating point, so I could add one there? Or make them all floats instead of integers?

Checklist

This pull request is:

• A documentation / typographical error fix
  • Good to go, no issue or tests are needed
• A short code fix
  • Please include: Fixes: #<issue number> in the description if it solves an existing issue
    (which must include a complete example of the issue).
  • Please include tests that fail with the main branch and pass with the provided fix.
• A new feature implementation or update an existing feature
  • Please include: Fixes: #<issue number> in the description if it solves an existing issue
    (which must include a complete example of the feature).
  • Please include tests that cover every lines of the new/updated feature.
  • Please update the documentation to describe the new/updated feature.

[adrien-berchet]

Hi @aballet
You're absolutely right, thank you very much for spotting this issue and sorry about this stupid miss...
I think it makes sense to update the tests to include floating point numbers, especially in test_functional.py::TestInsertionCore::test_insert_all_geom_types. You can add a pytest.mark.parametrize to test with only integers or with floating points.

[adrien-berchet]

I updated the test in aballet#1 so you can merge and then I'll merge this PR.

[aballet]

Ah, I see that I forgot to escape the dot myself, thank you!
I merged your code to include the tests.

[adrien-berchet]

Thanks for this work @aballet !
I'm going to push a new release.

[adrien-berchet]

The new release is available: https://pypi.org/project/GeoAlchemy2/0.15.1/
I hope it will solve your issue @aballet
Anyway, thanks again for this PR @aballet !