Gaussian splat update

[ericmehl]

This changes a couple of details of the ParticleField3DGaussianSplat loading.

• Using [] in the primitive variable name for radiance:sphericalHarmonicsCoefficients was causing problems in Gaffer because the [ character is used to start a character class in the string matcher rather than matching literally.
• We were loading HalfVectorData for opacities when the USD file provided them. To be consistent with the other Gaussian splats data types (which we always load as float) and to provide a consistent data type to users and renderer backends, we now convert opacities to FloatVectorData.

I'm not sure if the changelog entries are strictly necessary? I guess we need something to indicate a change, but I'd be happy to simplify the entries if we think this detail isn't helpful.

I'm also not seeing a more compact way of transferring the USD half opacities to FloatVectorData without an intermediate allocation of IECore::HalfVectorData. But if I'm missing something in the template magic that would allow that it would be nice to simplify this.

Checklist

• I have read the contribution guidelines.
• I have updated the documentation, if applicable.
• I have tested my change(s) in the test suite, and added new test cases where necessary.
• My code follows the Cortex project's prevailing coding style and conventions.

[johnhaddon]

If this fails, we make a PrimitiveVariable with null data, which I think will then trigger the warning about an invalid variable. But if the attribute isn't authored (which I assume would make Get() fail), we don't want to output a warning.

One simple approach to this might be to call readPrimitiveVariable() no matter the type, and then post-convert if we find halfs in there.

[ericmehl]

I did the readPrimitiveVariable() and post-convert in 14dad38

[ericmehl]

And a little more compact using variableData() in 51680c5

[ericmehl]

I fixed up the half -> float conversion, and tacked on a quick change to the wording of the changelog in 9365c41. Ready for a new look.

[murraystevenson]

Thanks Eric, a pair of minor notes inline. It would be good to get this ready for a release today so we can include the Arnold splats support in today's Gaffer release.

[ericmehl]

Thanks Murray and John, I touched up the code populating the float opacity vector and the changelog comment. I went ahead and squashed down as well to hopefully be ready to merge. I can get the Gaffer side of the splats updated as soon as the release is out.