networkrepository_analysis

Several analysis scripts that were used to analyze data from networkrepository.com for graph automorphisms

Install

No installation is needed but you must assure that all required packages are available on yout system. python setup.py --requires shows the packages. Most of them can be obtained using standard Python mechanisms for packages. pysaucy and pycggcrg need to be obtained separately, please follow their installation instructions.

All scripts were tested using Python 2.7, however, they should be compatible with Python 3 as well.

Usage

The purpose of those scripts is to download network/graph data, investigate this data with a focus on graph automorphisms, and analyze the results. The scripts must be executed in a given order:

1. networkrepository_download_datasets.py: Actually download data sets
2. networkrepository_download_stats.py: Download an overview of all data sets available for later analysis
3. networkrepository_run_analysis.py: Execute the analysis of the data sets
4. networkrepository_clean_dups.py: (Optional) Interactively clean the analysis results for duplicates in the data
5. networkrepository_compute_stats.py: Compute statistics on the obtained results

Each script file is executed calling python <script name>.py, most of them require additional information as command line parameters. For each script, the --help option is available.

References

Ball, Fabian, and Andreas Geyer-Schulz. 2018. “How Symmetric Are Real-World Graphs? A Large-Scale Study.” Symmetry 10 (1):17. https://doi.org/10.3390/sym10010029.