Infer thresholds for detect_clearsky

[ajonesr]

• Closes Extend detect_clearsky for different data periods #1808
• 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.

[cwhanse]

We should add a test that confirms _clearsky_get_threshold is called from detect_clearsky when infer_limitsis True. See the use of mocker.spy e.g. here.

[cwhanse]

@ajonesr please remove whatsnew/v0.10.0.rst from this PR. The test failures may be a problem with the new private function; somehow the alpha variable is getting set to NaN.

[ajonesr]

@ajonesr please remove whatsnew/v0.10.0.rst from this PR. The test failures may be a problem with the new private function; somehow the alpha variable is getting set to NaN.

Whoops, sorry about that. I think I've fixed it now?

[cwhanse]

@arjones still seeing the test error. You can run the tests locally pytest test_clearsky.py

[adriesse]

Could we have an issue with a brief high-level description linked to this?

[cwhanse]

Could we have an issue with a brief high-level description linked to this?

#1808

[cwhanse]

Thanks @ajonesr

[kandersolar]

Thanks @ajonesr! Well-structured code. A few suggestions below.

@cwhanse, perhaps this addition should be followed up with some change to pvanalytics.features.clearsky.reno.

[kandersolar]

LGTM!

[kandersolar]

Thanks @ajonesr and congrats on your first merged PR to pvlib!