Plot histogram of git commit times.

plot_git_commit_times.py

#!/usr/bin/env python
#
# Run: Plot histogram of git commit history times.
#  ./plot_git_commit_times.py \
#    -r git_repo_path \
#    -a git_author \
#    -o output_path

import os
import subprocess
import argparse
import pandas as pd
from plotnine import ggplot, aes, theme, themes, element_blank, element_line, \
                     element_text, geom_histogram, ggtitle, xlab, ylab, \
                     scale_x_continuous
from plotnine.options import figure_size


def my_theme(base_size=19, base_family='sans', rotate=True):
    ''' Plotnine plotting theme. '''
    thm = themes.theme_bw(base_size=base_size, base_family = base_family) + \
        theme(
            legend_background = element_blank(),
            legend_key = element_blank(),
            panel_background = element_blank(),
            panel_border = element_blank(),
            strip_background = element_blank(),
            plot_background = element_blank(),
            panel_grid = element_blank(),
            axis_line = element_line(colour = 'black', size = 1),
            axis_text_y = element_text(colour = 'black')
        )
    if rotate:
        thm += theme(axis_text_x=element_text(rotation=45, hjust=1))
    if figure_size is not None:
        thm += theme(figure_size=figure_size)
    return thm

def run_git_log(repo_path, author=None):
    '''  Run git log, with optional parameter for commit author. '''
    if author:
        command = ['git', '-C', repo_path, 'log', '--author', author]
    else:
        command = ['git', '-C', repo_path, 'log']
    git_log = subprocess.run(command, capture_output = True, text = True).stdout
    df = pd.DataFrame( [git_log.split(' ')[6] for git_log in list(map(lambda _ : _.strip(), git_log.split('\n'))) if git_log.startswith('Date')], columns=['Time'] )
    df['Hour'] = df['Time'].apply(lambda _: pd.to_datetime(_).hour + pd.to_datetime(_).minute / 60)
    return df

def plot_commit_times(repo_path, author=None):
    ''' Plot histogram ofgit commit times. '''
    repo_name = os.path.normpath(repo_path).split('/')[-1]
    military_time = list(range(0, 25, 4))
    normal_time = ['12:00 AM', '4:00 AM', '8:00 AM', '12:00 PM', '4:00 PM', '8:00 PM', '12:00 AM']
    if author:
        df = run_git_log(repo_path, author)
        title = repo_name + ' commit times by ' + author
    else:
        df = run_git_log(repo_path)
        title = repo_name + ' commit times'
    return (ggplot(df, aes(x='Hour')) +
                geom_histogram(bins=48) +
                ggtitle(title) +
                xlab('Time') +
                ylab('Frequency') +
                scale_x_continuous(breaks=military_time, labels=normal_time, limits=[0,24]) +
                my_theme(base_size=20, rotate=True))


def parse_arguments():
    '''Parse and return the command-line arguments. '''
    parser = argparse.ArgumentParser(description='Plot git histogram of git commits by time.')
    parser.add_argument('-r', '--repo_path', dest='repo_path', help='Git repo path.', required=False, default=os.getcwd())
    parser.add_argument('-a', '--author', dest='author', help='Optional: git commit author.', required=False)
    parser.add_argument('-o', '--out_path', dest='out_path', help='Optional: path to save plot.', required=False)
    args = parser.parse_args()
    return args


def main(args):
    commit_histogram = plot_commit_times(args.repo_path, args.author)
    if args.out_path:
        commit_histogram.save(args.out_path)
    else:
        print(commit_histogram)

if __name__ == '__main__':
    args = parse_arguments()
    main(args)