Calculate the cyclomatic complexity of a DOT file

cyclomatic_complexity.py

#!/usr/bin/env python3

"""
Calculate the cyclomatic complexity of a graph in a DOT file.

Author: Adrian Herrera
"""


from argparse import ArgumentParser

import networkx as nx
from networkx.drawing.nx_pydot import read_dot


def parse_args():
    """Parse the command-line arguments."""
    parser = ArgumentParser(description='Calculate the cyclomatic complexity '
                                        'of a DOT file')
    parser.add_argument('-o', '--output', required=True, help='Output CSV')
    parser.add_argument('dot', nargs='+', metavar='DOT',
                        help='Path to DOT file')
    return parser.parse_args()


def main():
    """The main function."""
    args = parse_args()
    dot_paths = args.dot
    csv_rows = []

    for dot_path in dot_paths:
        # Read the graph
        graph = read_dot(dot_path)

        # Find the entry node
        entry_node = None
        for node, in_degree in graph.in_degree():
            if in_degree == 0:
                entry_node = node
                break

        if not entry_node:
            raise Exception('Failed to find entry node in `%s`' % dot_path)

        # Find the exit node(s) and connect them to the entry node (making the
        # graph strongly connected)
        for node, out_degree in graph.out_degree():
            if out_degree == 0:
                graph.add_edge(node, entry_node)

        num_edges = graph.number_of_edges()
        num_nodes = graph.number_of_nodes()
        num_scc = nx.components.number_strongly_connected_components(graph)

        cyclomatic_complexity = num_edges - num_nodes + num_scc
        csv_rows.append(dict(path=dot_path, cyclomatic=cyclomatic_complexity))

    # Write CSV
    with open(args.output, 'w') as outf:
        writer = CsvDictWriter(outf, fieldnames=('path', 'cyclomatic'))
        writer.writeheader()
        writer.writerows(csv_rows)


if __name__ == '__main__':
    main()