visualize git commits stats

git_vis.py

#requirements: click, pandas, matplotlib, GitPython

from matplotlib import pyplot as plt
import pandas as pd
import json
from datetime import datetime
import git
import click

def parse_repo(repo_path: str):
    g = git.Repo(repo_path)
    data = [
        {
            "hash": c.hexsha,
            "author": c.author,
            "date": c.authored_datetime.ctime()
        }
        for c in g.iter_commits()
    ]
    df = pd.DataFrame(data)
    df.index = df["hash"]
    df.drop("hash", 1, inplace=True)

    df.author = df.author.astype("string")

    df.date = pd.to_datetime(df.date)
    df["weekday"] = df.date.dt.weekday

    return df


@click.command()
@click.argument("repo_path", type=click.Path(file_okay=False, exists=True), default=".")
@click.option("-o", "--out", "out_file", type=click.Path(dir_okay=False), required=True)
@click.option('--visualize', is_flag=True)
def visualize(repo_path: str, out_file: str, visualize: bool):
    df = parse_repo(repo_path)

    fig, axes = plt.subplots(2, 1, figsize=(16, 16))
    day_counts = df.weekday.value_counts()
    axes[0].barh(["mo", "tu", "we", "th", "fr", "sa", "su"], day_counts)

    auth_counts = df.author.value_counts()
    axes[1].barh(auth_counts.index, auth_counts)

    plt.savefig(out_file)

    if visualize:
        plt.show()


if __name__ == "__main__":
    visualize()