Add a Complex32 datatype

[barrettp]

Many radio telescope A2D samplers are 8 bit, e.g., those used by the VLA. This means that the visibility data can be stored in half-float (16-bit) format without loss of precision. Use of half-floats will reduce the storage of such data by about half. It is recommended that half-floats be included as an xarray datatype.

[kettenis]

numpy/numpy#14753