Welcome to HiDeF’s documentation!

HiDeF [1] aims to reimagine hierarchical data clustering. The name HiDeF stands for “Hierarchical community Decoding Framework”. HiDeF integrates graph-based community detection and the idea of “persistent homology” in order to determine robust clustering patterns in complex data at multiple scales. Given the inputs in data points in graph or matrix formats, HiDeF returns a list of multiscale clusters with measurement of their robustness, as well as a directed acyclic graph (DAG) to represent the organization of these clusters.

Installation

Local installation of python package

From source:

python setup.py install

Installation via pip

pip install hidef

Cytoscape

HiDeF is separately distributed via the CDAPS framework [2] in Cytoscape.

Note

We try to maintain timely synchronization of the HiDeF versions across the Python package and Cytoscape. However, it may be possible to have small difference in results across the platforms due to the Cytoscape version is behind the latest version of the Python package.

What’s new

Version 1.1.4:

  • Updated setup.py and added requirements.txt to specify minimum versions of packages HiDef depends on.

  • Added Makefile which includes shortcuts to build and deploy software

Version 1.1.3:

  • Community detection with multiple resolutions now run in parallel with python multiprocessing module

  • The default algorithm changed to Leiden as it is faster than Louvain

  • Now support multiplex community detection

Tutorials

API

References

Footnotes