Leland McInnes
Leland McInnes
If ``allow_single_cluster`` is true then presuming then the root should not be removed as a candidate; if there is simply a single cluster that slowly decays (e.g. root. == leaf)...
You could reach into the code a little bit and write some routines that will do this. If you looke at the [HDBSCAN generic code](https://github.com/scikit-learn-contrib/hdbscan/blob/4c432505f4a92884a64a77159664f041a583fbec/hdbscan/hdbscan_.py#L78-L151) you can see the heart...
Sorry, I'm in the process of fixing these -- neither is serious, and shouldn't impact usage right now, but are things I would like to get resolved at some point....
Yes, same issue. I will try to get time to work on this eventually, but a number of other matters are more pressing for me right now.
Hopefully this 0.8.20 release will fix these lingering issues for now.
Thanks @klcooksey , it seems I was a little hasty. Is there any chance you can pull from master, build and test if that works?
That's okay, I'll see if I can roll a few more minor patches in and get a release out. In practice you can ignore this test failure -- it is...
@klcooksey: Thanks! I'm glad to be able to help. If you have any further issues (including usage questions) don't hesitate to file an issue here. I'm not always super-fast to...
At least one thing may be to get the data you want to cluster into a C contiguous numpy array before handing it to HDBSCAN. Still, at 12 million rows...
Sorry for taking so long to get to this. It looks like a useful addition. Any chance you could add a test to the test suite to check that it...