Python package for the analysis of complex networks

Is it a fork? Is it a re-implementation? I don't know! What I do know is that this project uses for the implementations the algorithms from the projects NetCenLib and NetworkX. This project also attempts to have the for NetworkX configurable speed-up options pre-configured.

This project is for the most part an extension to NetCenLib and therefore uses the same structure

The aim is to simplify the analysis step by returning an optionally sorted and/or with the computed values re-scaled pandas DataFrame.

Installation

pip install netcenframe

parameters

configuration

To configure the library, there is a netcenframe.centrality.configure() functionality where settings can be passed to for the duration of the program's runtime

• sort: default is True
  • sorts the returned dataframe from highest to lowest centrality value The default settings need to be applied with a call to netcenframe.centrality.configure()
• They aren't automatically applied because some people would like to use their own parallel settings.

Calls to compute centralities looks like the examples below, in which the graph measure betweenness centrality will be computed via the library NetworkX and return a pandas DataFrame which is already sorted. Before a dataframe is returned, the data is saved as a file in CSV format for later usage, for example in other programs.

Two ways of calling functionalities in the library

example usage with imports and configuration:

# imports for netcenframe to make it easier to use
from netcenframe.centrality import compute_centrality as compute
from netcenframe.centrality import configure
from netcenframe.taxonomies.Centrality import BETWEENNESS, CLOSENESS, DEGREE

configure()
sorted_df = compute(/$graph, BETWEENNESS)

example without imports

import netcenframe
# apply config:
netcenframe.centrality.configure()
# run centrality
sorted_df = netcenframe.centrality.compute_centrality(/$graph, netcenframe.taxonomies.Centrality.BETWEENNESS)