r-trimbour / multixrank

MultiXrank - Universal Multilayer Exploration by Random Walk with Restart

MultiXrank is a Python package for the exploration of heterogeneous multilayer networks, with random walk with restart method. It permits prioritization of nodes between full heterogeneous networks, whatever their complexities. If you use MultiXrank in scientific works, please cite the following article:

Baptista, A., González, A., Baudot, A. Universal multilayer network exploration by random walk with restart. Commun Phys 5, 170 (2022)

https://doi.org/10.1038/s42005-022-00937-9

The scripts used in the paper and adaptable for your own applications of MultiXrank are available there: https://github.com/anthbapt/multixrank-tools

Commands for a quick installation:

conda create --name multixrank python=3.10 -y
python3 -m pip install multixrank

Commands for a quick working example in the python console:

import multixrank
multixrank.Example().write(path="airport")

This generates a working example based on the "airport" multiplex:

`-- airport
    |-- bipartite
    |   |-- 1_2.tsv
    |   |-- 1_3.tsv
    |   `-- 2_3.tsv
    |-- config_minimal.yml
    |-- multiplex
    |   |-- 1
    |   |   |-- FR26.tsv
    |   |   |-- FR3.tsv
    |   |   |-- FR3_2.tsv
    |   |   `-- FR7.tsv
    |   |-- 2
    |   |   |-- UK15.tsv
    |   |   |-- UK26.tsv
    |   |   `-- UK3.tsv
    |   `-- 3
    |       |-- G1.tsv
    |       |-- G24.tsv
    |       `-- G6.tsv
    `-- seeds.txt

The minimal configuration file 'config.yml' looks like this.

multiplex:
    1:
        layers:
            - multiplex/1/FR26.tsv
            - multiplex/1/FR3.tsv
            - multiplex/1/FR7.tsv
    2:
        layers:
            - multiplex/2/UK15.tsv
            - multiplex/2/UK26.tsv
            - multiplex/2/UK3.tsv
    3:
        layers:
            - multiplex/3/G1.tsv
            - multiplex/3/G24.tsv
            - multiplex/3/G6.tsv
bipartite:
    bipartite/1_2.tsv:
        source: 1
        target: 2
    bipartite/1_3.tsv:
        source: 1
        target: 3
    bipartite/2_3.tsv:
        source: 2
        target: 3
seed:
    seeds.txt

import multixrank
multixrank_obj = multixrank.Multixrank(config="airport/config_minimal.yml", wdir="airport")
ranking_df = multixrank_obj.random_walk_rank()
multixrank_obj.write_ranking(ranking_df, path="output_airport")
multixrank_obj.to_sif(ranking_df, path="output_airport/airport_seed7_top3.sif", top=3)

This runs the software and writes the results to the output_airport folder:

$ ls output_airport/
airport_seed7_top3.sif  multiplex_1.tsv  multiplex_2.tsv  multiplex_3.tsv

There is a ranking file for each multiplex:

$ head -n 4 output_airport/multiplex_1.tsv
multiplex   node    score
1   7   0.250002565842259
1   169 0.0025983048938841304
1   199 0.0018837852068513332

The MultiXrank documentation is hosted at ReadTheDocs.

MultiXrank is maintained by Anthony Baptista (anthony dot baptista at qmul dot ac dot uk)