Create dev doc outlining move of document ingestion and taxonomy processing into instructlab/instructlab

[bbrowning]

We need to determine exactly what code we want to move out of sdg and into the instructlab repo. What actual ingestion and processing code? What tests? What documentation?

What's the new set of inputs and outputs along with the method instructlab will call into sdg to kick off the data generation pipeline?

We need a dev doc that we can discuss and get feedback from the instructlab core maintainers team on the work to be done to ensure they're ready to accept the new scope.