Leland McInnes

@RichieHakim I think at this stage the best option is to look at [fast_hdbscan](https://github.com/TutteInstitute/fast_hdbscan) which is a newer multicore implementation. A [recent pull request](https://github.com/TutteInstitute/fast_hdbscan/pull/21) added support for semi-supervised clustering which...

I don't know any way of doing that with the current code, but I agree that it certainly makes a lot of sense. I'll add this as a feature request,...

Thanks for the explanation. Your approach is not so fr off from what I had in mind. I was actually thinking of making it a little more general and having...

Embedding constraint-wise that's certainly a little more complicated to implement directly. It certainly isn't impossible, but would require a different approach. I'll certainly keep it in mind however.

There were definitely some bugs in earlier versions of UMAP that could cause this to happen. Do you know what version you are using? Can you reproduce the problem on...

My best guess is that the mahalanobis distance is causing some issues somehow. Can you reproduce the result with, say, eucliudean distance?

If I change it now, it will break compatability with all earlier releases of sklearn. For this reason I think it will be better to accept some deprecation warnings and...

My argument is that ``n_neighbors`` makes sense for t-SNE, and many people have a much clearer intuitive grasp of what ``n_neighbors=15`` means than what ``perplexity=30`` means. One has a very...

I definitely understand your concerns, as I have had similar issues with, for example, the hdbscan libraries I've written. There are things I would like to tweak, but it does...

Bounded metrics as output metrics are often a bad idea. You would probably be better off doing something akin to the example of hyperbolic space in the docs: find a...