Make first set of SDP processing functions ready for release

Based on an existing solution, a processing function library release should contain a (possibly simple) set of processing functions that is

• ready for usage (i.e. documented, with clear Python & C++ API )
• internally well factored (i.e. following clear design and language patterns and supporting robust testing and packaging via CI)

Such that it can be used as a template for future processing function releases. The approach should have been reviewed with relevant (developer) stakeholders.

OLD Notes for review:

populated with a skeleton structure for source folders, docs, and CI pipeline

Expect to need spike to review, update, and agree on structure following <TODO link!> straw man proposal.

(this will likely need 1FP - ~10+ people for a few hours each)

The remaining FP is intended to include

• Initial repo set up in gitlab under `sdp/ska-sdp-func` 
• Initial source structure (folders) defined with readme's in each to describe what goes into each folder.
• Initial doc skeleton (user and developer) structure set up with build steps (manual + CI pipeline)
• Initial build system set up for C/C++ package (including use of CUDA kernels) and Python bindings
• Initial skeleton CI pipeline developed (for build, test, lint, docs stages)

Where at all possible this should adopt the SKA standardised project structure or otherwise follow from existing projects rather than try to invent an entirely new approach.

This does have to be perfect and can be iterated on but needs to be in place to allow the development of this library to proceed with contributions from multiple developers/teams in parallel. 

 __