Fix reverse mapping issue and undesirable column reordering in get_psm3

[AdamRJensen]

• Closes pvlib.iotools.get_psm3 attribute "reverse mapping" produces invalid names #1629 and closes get_psm3() reorders the dataframe columns randomly #1647
• 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.

This PR adds a separate dictionary for variable mapping of the requested parameters (attributes).

I simply removed the set operation (as discussed in #1647), which had the purpose of dropping duplicate values in the list of requested parameters (attributes). If we want to retain that functionality, we can do:

attributes= list(dict.fromkeys(attributes))

[kandersolar]

This approach looks good to me. It would be good for the tests to include check that column ordering is not scrambled. Would it make sense to modify test_psm3.py::test_get_psm3_attribute_mapping to assert not just column presence but also ordering?

[AdamRJensen]

This approach looks good to me. It would be good for the tests to include check that column ordering is not scrambled. Would it make sense to modify test_psm3.py::test_get_psm3_attribute_mapping to assert not just column presence but also ordering?

Yes that should be an easy and simple approach.

Edit: it has now been implemented.

[kandersolar]

Both bugfixes warrant mention in the whatsnew IMHO

[kandersolar]

Thanks @AdamRJensen!