System module sphinx documentation

[kperrynrel]

Description

Thank you for your contribution! Please provide a brief description of the problem and the proposed solution or new feature (if not already fully described in a linked issue)

Checklist

The following items must be addressed before the code can be merged.
Please don't hesitate to ask for help if you are unsure of how to accomplish any of the items.
You are free to remove any checklist items that do not apply or add additional items that are
not on this list

• Closes #xxx
• Added new API functions to docs/api.rst
• Clearly documented all new API functions with PEP257 and numpydoc compliant docstrings
• Adds description and name entries in the appropriate "what's new" file
  in docs/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`).
• Non-API functions clearly documented with docstrings or comments as necessary
• Added tests to cover all new or modified code
• Pull request is nearly complete and ready for detailed review

[kandersolar]

Summarizing offline discussion with @kperrynrel: at least in the PVWatts fits, the inaccuracy stems from several sources:

1. Assuming static ambient temperature of 25 degrees
2. Not-great accuracy of the Ineichen model w/ fixed turbidities
3. Assuming isotropic diffuse irradiance in the transposition, and perhaps other minor parts of the performance model

We came up with an alternative approach using PSM3 data that addresses (1) and (2). Let's take up (3) separately in #147.

Other notes from a brief high-level look:

• Do we need both of the CSV files? Seems like only one of them gets used in the examples.
• The CSVs are pretty big at 2.3 MB each. The (compressed) dist files for v0.1.1 are only 85 kB so including these is a big jump. At least for the purposes of this example, I suspect there's not much value in having multiple years of data since a single year (or maybe even just 6 months) has all the relevant seasonality being fitted here. Should we pare those files down?
• I think the docs/examples/.pylint.d/infer-orientation-daily-peak1.stats file can/should be removed

[kperrynrel]

@kanderso-nrel I made the recommended updates. Let me know when you're good to move ahead on this, and I'll send it over to Cliff!

[kandersolar]

There's some ongoing offline discussion about whether the daily-peak function currently in pvanalytics actually implements the desired functionality -- more on that in the coming days.

Anyway a few comments below about the other two examples, otherwise LGTM! Not hitting "approve" yet since we're still undecided on the daily-peak point.

[kperrynrel]

@kanderso-nrel I'm for us delaying this PR until after PVPMC, as we are actively working on these functions in our new data validation. Maybe we just push through the orientation PR before August 22nd?

[kperrynrel]

@kanderso-nrel I cleaned things up for a new release. Let me know if this looks solid and I'll add Cliff as a reviewer. Thanks!

[kandersolar]

OK with me