Match pvsystem.i_from_v, v_from_i single diode parameters with singlediode order.

[reepoi]

• Closes Match pvsystem.i_from_v, v_from_i single diode parameters with singlediode order. #1718
• 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.

The argument order for i_from_v, v_from_i is changed to:

voltage/current, photocurrent, saturation_current, resistance_series, resistance_shunt, nNsVth

In tests/test_singlediode.py, the tests test_method_spr_e20_327, test_newton_fs_495 call i_from_v with the comment:

# the singlediode method doesn't actually get the MPP correct

singlediode does get the MPP correct now, so I updated the tests to compare all returned values from the pvs and out variables.

The test test_newton_fs_495 passes ivcurve_pnts but then does nothing with the returned i, v values.
Since the precision of the i, v values are tested below this test, I updated the test to pass ivcurve_pnts as None.

[cwhanse]

I'm +1 on this change.

It would be nice to deprecate. But I can't think of an easy way to do that for a change in the order of positional arguments. Because these functions are likely used by a small minority of pvlib users, I'd be OK with this breaking change without deprecation.

[kandersolar]

I think this change would make the next version 0.10.0 instead of 0.9.6. Doesn't matter, just pointing it out.

The only deprecation mechanism I can think of that could work here is to inspect the input and guess whether it is more likely to be a voltage or a resistance based on the range/value. We've done something similar in the past, but it's inelegant and unreliable. I'd be okay with no deprecation.

For documenting, in addition to the what'snew entry, a .. versionchanged:: docstring entry seems appropriate. Example:

pvlib-python/pvlib/iotools/psm3.py

Lines 71 to 74 in b0ac666

	.. versionchanged:: 0.9.0
	The function now returns a tuple where the first element is a dataframe
	and the second element is a dictionary containing metadata. Previous
	versions of this function had the return values switched.

[kandersolar]

A few minor comments but LGTM. Thanks @reepoi, it's great to see consistency improvements like this!

[reepoi]

I missed updating the argument order of pvlib.tracking.SingleAxisTracker.i_from_v. I will do that next.

[reepoi]

Nevermind on updating pvlib.tracking.SingleAxisTracker.i_from_v. SingleAxisTracker inherits from PVSystem, so the documentation for it is already updated.

[kandersolar]

I missed updating the argument order of pvlib.tracking.SingleAxisTracker.i_from_v. I will do that next.

No need to update SingleAxisTracker IMHO. It's deprecated and only still included for legacy code while people migrate to the better alternative, so there's no need to make any changes motivated by long-term benefits (like consistent parameter ordering, for example).

[cwhanse]

Feel free to rename this internal I here and elsewhere to mollify stickler. Not important.

[reepoi]

Ok, I renamed I to current. To make the rest of the code match, I changed I to current and V to voltage everywhere.

[reepoi]

Never mind, I changed it only at the stickler comments. I is also used as part of a variable name in many places.

[kandersolar]

Thanks again @reepoi!