This is a covertree library with some modifications to make it more suitable for real data. Currently it only implements the fast covertree, which is an extension of the original covertree (pdf). There are plans to enable support for full geometric multi-resolution analysis (GMRA, where the library get it's name from) and topological data analysis. Help is welcome! We'd love to collaborate on more cool tricks to do with covertrees or coding up the large backlog of planned features to support the current known tricks.

Data Access is handled through the pointcloud library. See here for pointcloud's documentation. This is meant to abstract many files and make them look like one, and due to this handles computations like adjacency matrices. The covertree implementation is inside the goko library, it's the bread and butter of the library. See here for it's documentation.

The pygoko library is a python & numpy partial wrap around goko. It can access the components of goko for gathering statistics on your trees. Once we settle on how this is implemented we will publish the documentation somewhere.