Remove pvlib.forecast

[kandersolar]

• Closes forecast module errors: Multiple feature collections cannot be written as a CF dataset #1596
• Closes Forecasting page in User's guide #1417
• Closes Allow request headers and other arguments to pass through TDSCatalog #918
• 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.

Following the thumbs up on #1735 (comment), this PR removes pvlib.forecast. Removing pvlib.forecast, combined with #1654, means that several optional dependencies (siphon, netcdf4, cftime) are no longer used anywhere and can also be removed. This PR removes them here, but we should also remember to remove them from our conda-forge feedstock when we release v0.10.0.

[kandersolar]

Note that the remote-data tests are expected to fail here due to the fact that the modified workflow is not run on the PR containing the modification. See here for the longer explanation.

[AdamRJensen]

I did a light review and things seem to check out.

[kandersolar]

Hearing no objections, away it goes!

[wholmgren]

The end of an era! Thanks for cleaning up my mess!

[stefan-hoermann]

Hey, I am Stefan, a Student from Germany. Might I ask why you removed the feature ? I was going to use it for my masters thesis.

[kandersolar]

Hello @stefan-hoermann, see #1057 and #1426 for discussion about this decision.

I'm not sure if it will be helpful, but you might try searching the pvlib google group to see what other people are using instead of pvlib.forecast.

[AdamRJensen]

@stefan-hoermann Note, you can always install previous versions of pvlib, which has the forecasting module, although not all of the functionality still works.

[stefan-hoermann]

Thank you for your quick answers. Looking through the discussions with all of the bugs I might need an alternative. I am looking for a radiation prognosis tool, maybe the "Solar Forecast Arbiter" mentioned in #1057 is an alternative. Although I hope there may be a solution which directly accesses the database of institutions like e.g. the "Deutscher Wetter Dienst". At least I think thats how pvlib worked, or not?