Custom losses, coherent embeddings

Nice property of TSNE, that is not exploited in most of implementations, is that it can be treated as a combination of two orthogonal components: loss function and optimization algorithm. For example one may visualize set on temporally varying vectors with a sequence on coherent embeddings by adding a loss term that penalizes unnecessary movement of each vector between those embeddings. Is it possible to have provide such flexibility to use additional constrains with UMAP?

The issue proposes separating UMAP loss functions from optimization and adding constraints for coherent embeddings, but it names no files, tests, or entry points. Start by locating the UMAP optimization and loss implementation; done would require a defined extensibility approach and validation for custom loss terms and constraints.

Written by the indexing model from the issue text.