fit_pvsyst_sandia uses a mutable dict keyword argument

[wholmgren]

the ivtools subpackage has code like this:

# utility.py

constants = {'E0': 1000.0, 'T0': 25.0, 'k': 1.38066e-23, 'q': 1.60218e-19}


# sdm.py
from pvlib.ivtools.utility import constants

def fit_pvsyst_sandia(ivcurves, specs, const=constants, maxiter=5, eps1=1.e-3):

The major thing that concerns me is the use of a mutable default. It's generally better practice to use a pattern like const=None in the signature and then if const is None: const = constants

Other things that I noticed while reporting this:

• We're starting to consistently follow the recommended practice of formatting module level "constants" in upper case. So constants would become CONSTANTS.

• utility.py is often called utils.py in other packages. We also have a top level tools.py which could be changed too.

thoughts on the above @cwhanse?

[cwhanse]

I'm OK with both suggestions:

constants -> CONSTANTS, and maybe move out of ivtools to someplace more discoverable?

ivtools\utility.py can be ivtools\utils.py. I don't see the contents as relevant outside of ivtools, at present.

[wholmgren]

constants -> CONSTANTS, and maybe move out of ivtools to someplace more discoverable?

Sounds like the other half of #483. Can we agree on a naming scheme quickly enough to get it into 0.8.0 (#825 is relevant to that aspect). If not, I'll propose that we keep this dictionary where it is for now so that we don't have to worry too much about breaking user code if we change it.

k and q can be pulled from scipy.constants once #1035 is merged.

[mikofski]

I've merged #1035