DirichletDistribution and Infer are too similar. Infer ends up exposing many of the methods for DirichletDistribution anyway. It should be possible to have a single class, or alternatively, to make Infer a subclass of DirichletDistribution. Currently leaning towards a single class with cdef methods implementing most of what DirichletDistribution did, and cpdef methods for the Infer functionality.
DirichletDistributionandInferare too similar.Inferends up exposing many of the methods forDirichletDistributionanyway. It should be possible to have a single class, or alternatively, to makeInfera subclass ofDirichletDistribution. Currently leaning towards a single class withcdefmethods implementing most of what DirichletDistribution did, andcpdefmethods for theInferfunctionality.