Parametric aligned umap?

[IanQS]

I'm interested in first training a AlignedUmap on some time-series dataset. I then want to use the trained AlignedUmap to run inference on some new points to see if the new points lie on the surface learned by the Umap. In a standard UMAP, I can just use the ParametricUmap to learn the encoder and decoder, but is this possible with AlignedUmap? I checked the source code and didn't find anything about passing in an encoder or decoder.

Thank you!

[lmcinnes]

AlignedUMAP and ParametricUMAP don't currently fit together, so I don't quite think that will be possible I'm afraid. Sorry.