visualize author contributions to one or more git repos

git_stats.py

#!/usr/bin/env python3

"""Describe author contributions for one or more git repositories by date

Output is a csv with columns (repo, author, timestamp, churn) where
'churn' is the sum of lines added and lines removed.

Requires pandas and plotnine

"""

import sys
import argparse
import subprocess
from pathlib import Path
import re
import csv

import pandas as pd
from plotnine import (ggplot, geom_bar, aes, facet_wrap, theme,
                      element_text, scale_y_log10)
import plotnine as p9


def parse_log(text, exclude=None):
    repo, author, ts, churn = None, None, None, 0

    if exclude:
        exclude = re.compile(exclude)

    for line in text.splitlines():
        if line.startswith('#'):
            if repo:
                yield (repo, author, ts, churn)
            (repo, author, ts), churn = line[1:].split('|'), 0
        elif line.strip():
            gain, loss, filename = line.split(None, 2)
            if exclude and exclude.search(filename):
                continue
            try:
                churn += int(gain) + int(loss)
            except ValueError:
                # eg, binary files report lines changes using '-'
                churn += 1

    yield (repo, author, ts, churn)


def main(arguments):

    parser = argparse.ArgumentParser(
        description=__doc__,
        formatter_class=argparse.RawDescriptionHelpFormatter)
    parser.add_argument('repos', help="Path to one or more repositories", nargs='+')
    parser.add_argument('-x', '--exclude', help='regular expression excluding filenames')
    parser.add_argument('-a', '--author-map', type=argparse.FileType(),
                        help="""headerless csv file with columns
                        (label, author-email) providing instructions
                        for consolidating or (if the first column is
                        left empty) excluding authors. """)
    parser.add_argument('-o', '--outfile', help="CSV output file")
    parser.add_argument('-p', '--plotfile', help="plot output file")

    args = parser.parse_args(arguments)

    if args.author_map:
        author_map = {v: k for k, v in csv.reader(args.author_map) if k}

    data = []
    for path in args.repos:
        pth = Path(path)
        repo = pth.name
        cmd = ['git', '-C', path, '--no-pager', 'log',
               '--numstat', 'master', f'--format=format:#{repo}|%ce|%ci']

        job = subprocess.run(cmd, capture_output=True, text=True)
        data += list(parse_log(job.stdout, exclude=args.exclude))

    df = pd.DataFrame(data, columns=['repo', 'author', 'timestamp', 'churn'])
    if args.outfile:
        df.to_csv(args.outfile, index=False)

    if author_map:
        df = df[df['author'].isin(author_map)]
        df['author'] = df['author'].apply(lambda a: author_map[a])

    df['date'] = pd.to_datetime(df['timestamp']).apply(lambda ts: ts.strftime('%Y-%m'))
    df.drop(['timestamp'], axis=1, inplace=True)
    grouped = df.groupby(['repo', 'author', 'date'])
    tab = grouped.sum().reset_index()

    plt = (ggplot(tab, aes('date', 'churn', fill='author')) +
           geom_bar(position='stack', stat='identity') +
           facet_wrap('~repo', ncol=1) +
           scale_y_log10() +
           p9.theme_538() +
           theme(
               axis_text_x=element_text(rotation=90),
               axis_text_y=element_text(size=0),
               axis_title_y=element_text(size=0),
               legend_position='top'))

    if args.plofile:
        plt.save(args.plotfile)


if __name__ == '__main__':
    sys.exit(main(sys.argv[1:]))