Fixed errors in soiling.hsu function and tests

[jsstein]

• Closes pvlib.soiling.hsu model implementation errors #970
• I am familiar with the contributing guidelines
• Tests added
• [n/a] Updates entries to docs/sphinx/source/api.rst 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 and Milestone are assigned to the Pull Request and linked Issue.

[cwhanse]

Looks like mostly formatting corrections, click through the stickler-ci details in the list of checks. Do you want me to get a copy of the reference paper and double check the default values?

[wholmgren]

E203 can be ignored in the test suite as far as I'm concerned.

[jsstein]

I think I have addressed most of the the stickler-ci details (except for the E203 ones which add readability). I pushed the changes.

[jsstein]

Heavy fingers in vi! From: Cliff Hansen <notifications@github.com> Reply-To: pvlib/pvlib-python <reply@reply.github.com> Date: Wednesday, June 10, 2020 at 8:38 PM To: pvlib/pvlib-python <pvlib-python@noreply.github.com> Cc: Joshua Stein <jsstein@sandia.gov>, Author <author@noreply.github.com> Subject: [EXTERNAL] Re: [pvlib/pvlib-python] Fixed errors in soiling.hsu function and tests (#977) @cwhanse commented on this pull request.

________________________________ In pvlib/tests/test_soiling.py<#977 (comment)>:

@@ -119,8 +110,8 @@ def test_hsu_defaults(rainfall_input, expected_output_1):

accumulation period. """ result = hsu( - rainfall=rainfall_input, cleaning_threshold=0.5, tilt=0.0, pm2_5=1.0, - pm10=2.0) + rainfall=rainfall_input, cleaning_threshold=0.5, tilt=0.0, i stray i here is causing the test failures — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#977 (review)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABJES5CVEXE4EOIDQVJJ2ZLRWA7SDANCNFSM4NZ3PLTA>.

[cwhanse]

Thanks @jsstein

[kandersolar]

A couple comments (sorry for the post-merge review)

[kandersolar]

Missed the PR number in the whatsnew entry

[cwhanse]

I merged too early, thanks for taking a careful look

[jsstein]

I just submitted PR980 that addresses @kanderso-nrel 's comment about arbitrary time intervals. Next I will submit a demonstration of the hsu function showing that it matches one of the figures in the original JPV paper.

[kandersolar]

Sorry if I'm thinking about this wrong, but I want to raise two points about this calculation:

• Is it correct to hardcode the 3600? Shouldn't it be based on the datetime index's interval? I'm not seeing any documentation requiring the input timeseries to be hourly.
• Regardless of whatever value that timescale coefficient takes, shouldn't it apply to both particulate sizes? Right now it only affects pm10.

[jsstein]

@kanderson-nrel: I agree that I should take another look at the hardcoding of 3600. About your second point, let me look more carefully at the Matlab version and see how it deals with it.