Zachary Kurtz

Addresses my request in #210. 1. Introduces a _BaseLMNN/ LMNN class to operate on triplets, s.t. `d(triplets[i, 0],triplets[i, 1]) < d(triplets[i, 0], triplets[i, 2])` - the same setup as SCML....

[suggestion] update for latest conda/python

7

First I'd like to say thanks for this excellent tool... the application and feature sets fits my needs perfectly. However, to future proof your code you might want to consider...

I love the method and fantastic code base. I'm wondering if you've thought about an extension for pairwise constraints rather than regression. For instance in alignf, the kernel alignment depends...

Non-zero Gromov distance on identical graphs

1

Thanks for this excellent package, it's a very cool method. However, I've noticed what may be a technical bug where on *some* graphs, where I have found a non-zero Gromov...

Quick fix for #3

The tree_transport.ipynb is failing with ``` File ~/Repos/FGW/lib/graph.py:187, in Graph.all_matrix_attr(self, return_invd) 186 def all_matrix_attr(self,return_invd=False): --> 187 d=dict((k, v) for k, v in self.nx_graph.node.items()) 188 x=[] 189 invd={} AttributeError: 'Graph' object...

Can bgzip multi-threading be combined with random access reads?

2

Is there a general strategy for simultaneous random access to GZ blocks in an indexed gz file? For large files, i know the virtual positions and buffer sizes i want...

Would there be any interest in a Covariance module?

2

Similar to the [scikit-learn module](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.covariance). I'm primarily interested in implementing the Graphical Lasso, but empirical and shrinkage covariance estimators as well.

Things to fix: * Jobs are failing due to a missing dependency (rcmdcheck - shouldn't devtools be pulling this in? ) * Cache dependencies so builds don't take 20 minutes....