Radar and Spinner Lidar Sampling

[ndevelder]

Large commit for new sampling that mimics Spinner Lidar and Radar

[ndevelder]

Ok, so I'm moving this out of draft as soon as all the checks go through. I think I've addressed everything (albeit in the most minimal way possible). There are a few details to note:

• I removed ConeSampler since it is untested and not part of any work package...it was a waypoint on the path to getting the radar sampler in.
• The spinner and radar sampler unit tests only check the expected number of sampling locations, but could have a lot more work done to them. I'm hoping we can file this under future nice things to have. We've tested both samplers outside of the unit and reg test frameworks and have thought the results are reasonable.
• After thinking about it some more, I don't think there is a place for actuator testing here...the sampler "attaches" to one openfast instance and reads the orientation matrix coming from openfast...but has no actual interaction with actuator points. And it seems like we're still working on a simple/seamless integration of openfast models in the testing infrastructure? If there are any ideas on something to test related to this I'm happy to consider, but I think we need openfast testing integration first?

I'd also like to add a reg test for each sampler in a subsequent PR. I can open an issue and assign to myself if that seems worth it.

[mbkuhn]

There is a lot of code added in SamplingUtils, and these standalone functions seem like good candidates for unit testing. Don't need tests for every single one, but ideally if there is a basic problem with a known solution for a few of the complicated ones, it would help solidify this part of the code and rule out errors here if there is debugging required in the future.

[mbkuhn]

I think this PR is really close. 2 things come to mind: unit tests for SamplingUtils, which I mentioned above, and the preexisting sampling methodology. Because this PR reworked some of the functions in Sampling.cpp, is there any concern that the changes have affected the ordinary sampling there? Unit tests exist for some of that already, I'm just wanting to confirm that this aspect has been looked at.

[ndevelder]

Jumping back into this today! A couple points related to your comments:

• Happy to look at sampling utils and construct a test or two
• We've done testing and production runs with this PR for the RAAW project (multiple samplers in addition to the spinner) and Dan didn't see any problems. I do think messing with the buffer variable comes with some risk, but we haven't been able to break it yet.
• Other than that, it was mostly tidying, line-of-sight calculation for the radar, and allowing for data reduction in the sampler (with the default for most samplers that num points = num output points)

[mbkuhn]

Excellent. @ndevelder I'll approve and merge after some unit tests for Sampling Utils functions are added.

[ndevelder]

I got some of the simpler SamplingUtils tests in, but I don't have an easy way to get an independent result for the spherical cap quadrature stuff. That was effectively a direct copy from Nalu-Wind implemented by Robert K. I suppose I could try to put in something random and get a result, and just use that result to make sure it doesn't change in the future?

[mbkuhn]

Thanks for doing this, Nate! That is fine with me, if it's not too much of a lift. Just to safeguard the current version.

[mbkuhn]

Great work, Nate. Thanks for cooperating with my requests and improving the standards of the code.