Fix port of old-precision-decimal to present-atol in tests

[echedey-ls]

• Closes comment Change accuracy of pvsyst_cell test function #2080 (comment)
• I am familiar with the contributing guidelines
• [ ] Tests added
• [ ] Updates entries in docs/sphinx/source/reference for API changes.
• [ ] Adds description and name entries in the appropriate "what's new" file in docs/sphinx/source/whatsnew for all changes. Includes link to the GitHub Issue with :issue:`num` or this Pull Request with :pull:`num`. Includes contributor name and/or GitHub username (link with :ghuser:`user`).
• [ ] New code is fully documented. Includes numpydoc compliant docstrings, examples, and comments where necessary.
• Pull request is nearly complete and ready for detailed review.
• Maintainer: Appropriate GitHub Labels (including remote-data) and Milestone are assigned to the Pull Request and linked Issue.

Some tests were using a tolerance higher than expected.

Feel free to find other occurrences with (assert_allclose)\((.*), ?(.*), ?([a-zA-Z0-9=]*)?(\d)\)

These are the left integer tolerances:

But I have played a bit with some of em and didn't think they are a misunderstood decimal position.

[echedey-ls]

This regex to show that we are talking either about assert_allclose or assert_frame_equal. Constrains finds to functions that start with assert_

(assert_.*)\((.*), ?(.*), ?([a-zA-Z0-9]*=)?(\d)\)$

EDIT:
To take into account multiline calls:

(assert_.*)\((.*),[ \n]*?(.*),[ \n]*?([a-zA-Z0-9]*=)?(\d)[ \n]*\)$

[cwhanse]

I'm OK with the assert_allclose statements that aren't changing. These statements are in tests of PVSystem methods, where assert_called_with is the primary check. The assert_allclose statements are only intended to confirm that the results of the method are somewhat as expected.

[kandersolar]

Thanks as always for the comprehensive codebase greps @echedey-ls!