Require TI must be length n_findex in core code

[paulf81]

Require TI must be n_findex in core code

After some discussions we agreed that a consistent to handle a few issues would be for the "core" code to require the turbulence intensities must be n_findex. This avoids ambiguities and cases where the code might perform unexpectedly. However, to simplify things for the user, the wind_data objects, WindRose, WindTIRose and TimeSeries will each now include methods for broadcasting a provided float value to match the given dimensions.

Further, the approach proposed in PR #821, where heterogeneity can be described with respect to wd, in addition along findex, will be adapted in this pull request to be methods of the respective WindData objects. This enables that FlorisInterface can go back to requiring het_map to be have 0th dimension n_findex long, simplifying that code.

Steps in this draft then:

• In flowfield: require that turbulence_intensities be length n_findex
• In FlorisInterface: remove helper code to catch unchanged TI and set automatically
• Update impacted tests to adjust to these changed requirements
• Ensure all examples are working with new stricter rules
• Enable broadcasting in the wind data objects
• Add tests of broadcasting
• Reorganize unpack function to TI appears before freq, consistent with inputs
• Implement heterogeneity via wind direction in the wind data objects
• Add tests of new heterogeneity functionality
• Close Enable heterogeneity to work with new 4D framework and wind_data objects #814
• Update the first example around these new rules

Related issue

PR #821

[paulf81]

hi @rafmudaf @misi9170 and @ejsimley , I think this is now just about ready for review, I kept the checklist above up to date as I went so should provide a good overview of what changed and why. I originally thought it smart to update examples in this pull request but now think it's maybe best to separate that piece out, so crossed it out. If you could have a look would appreciate!

[paulf81]

Ah shoot no, I need to update all the examples, sorry, one last thing

[paulf81]

Ok I think all examples are fixed and this is ready for review

[ejsimley]

Looks good to me, aside form some clarification in the docstrings about the heterogenous_inflow_config dimensions.

[ejsimley]

Made a few last docstring updates. I think it looks good now.

[misi9170]

I haven't pulled down and run, but this looks good overall to me

[paulf81]

Thanks for all the feedback gang, merging now!