Personal git repo contribution statistics based on git log

personal_git_stat.py

import os
import re
import math
import subprocess
import matplotlib.pyplot as plt
from collections import defaultdict
import numpy as np
import logging

logging.basicConfig(level=logging.INFO)


def execute_command(cmd):
    """Execute shell command and return its output."""
    try:
        return subprocess.getoutput(cmd).strip()
    except Exception as e:
        logging.error(f"Failed to execute command: {cmd}, Error: {e}")
        return ''


def process_time_args(since, until):
    """Generate time arguments for git commands."""
    time_args = ""
    if since:
        time_args += f" --since='{since}'"
    if until:
        time_args += f" --until='{until}'"
    return time_args


def analyze_repo(repo, author, line_stats, daily_commits, since=None, until=None):
    os.chdir(repo)
    
    time_args = ""
    if since:
        time_args += f" --since='{since}'"
    if until:
        time_args += f" --until='{until}'"
    
    cmd_line_stats = (f"git log --author={author}{time_args} --pretty=tformat: --numstat | "
                      "awk '{ add += $1; subs += $2; loc += $1 - $2 } END { printf \"%s,%s,%s\", add, subs, loc }'")
    output = execute_command(cmd_line_stats).split(',')
    
    if '' not in output and len(output) == 3:  
        added, removed, total = map(int, output)
        line_stats[repo] = {'added': added, 'removed': removed, 'total': total}
    else:
        print(f"No commits found in the time range for repo {repo}")

    cmd_daily_commits = f"git log --author={author} --pretty='%ad' --date=format:'%Y-%m-%d'"
    output = execute_command(cmd_daily_commits).split('\n')
    for date in output:
        daily_commits[date] += 1

    os.chdir('..')


def add_text_labels(ax, values, bar_positions, bar_width, vertical_offset=0.1):
    """Add text labels to bars in a bar chart."""
    for index, value in enumerate(values):
        text_position_x = bar_positions[index] + bar_width / 2 - 0.15
        
        if value > 0:
            log_value = math.log10(value)
        else:
            log_value = 0
        
        text_position_y = 10 ** (log_value + vertical_offset)
        ax.text(text_position_x, text_position_y, str(value), ha='center')


def plot_line_stats(author, stats):
    """Plot a bar chart for line statistics."""
    repos = list(stats.keys())
    added = [stats[repo]['added'] for repo in repos]
    removed = [stats[repo]['removed'] for repo in repos]
    total = [stats[repo]['total'] for repo in repos]

    sum_added = sum(added)
    sum_removed = sum(removed)
    sum_total = sum(total)

    barWidth = 0.3
    r1 = np.arange(len(repos))
    r2 = [x + barWidth for x in r1]
    r3 = [x + barWidth for x in r2]
    
    fig, ax = plt.subplots(figsize=(15, 7))
    plt.subplots_adjust(right=0.8)

    ax.bar(r1, added, color='g', width=barWidth, label='added', log=True)
    ax.bar(r2, removed, color='r', width=barWidth, label='removed', log=True)
    ax.bar(r3, total, color='b', width=barWidth, label='total', log=True)

    add_text_labels(ax, added, r1, barWidth)
    add_text_labels(ax, removed, r2, barWidth)
    add_text_labels(ax, total, r3, barWidth)

    plt.xlabel('Repositories', fontsize=15)
    plt.ylabel('Line counts (Log scale)', fontsize=15)
    plt.title(f'Line Stats by {author}', fontsize=15)
    plt.xticks([r + barWidth for r in range(len(repos))], repos, rotation=45, ha="right", rotation_mode="anchor")
    
    legend_text = f"Total Added: {sum_added}\nTotal of Removed: {sum_removed}\nTotal of Total: {sum_total}"
    ax.text(1.05, 1, legend_text, transform=ax.transAxes, fontsize=12, verticalalignment='top', horizontalalignment='left')
    
    plt.legend()
    plt.show()


def plot_daily_commits(author, daily_commit_count):
    dates, commit_counts = zip(*sorted(daily_commit_count.items()))  # Sort based on dates
    total_commits = sum(commit_counts)
    
    fig, ax = plt.subplots()
    plt.subplots_adjust(right=0.8)  # Make space for the external legend
    ax.bar(dates, commit_counts, color='b')
    
    ax.set(xlabel='Dates', ylabel='Commit Counts',
           title=f'Daily Commit Counts by {author}')
    ax.grid()
    
    # Adding a total commit count legend in the top right corner
    legend_text = f"Total Commits: {total_commits}"
    ax.text(1.05, 1, legend_text, transform=ax.transAxes, fontsize=12, verticalalignment='top', horizontalalignment='left')
    
    plt.xticks(rotation=45, ha="right", rotation_mode="anchor")  # Rotate labels for better visibility
    plt.show()


def get_daily_line_stats_for_repo(repo, author, daily_line_stats, since=None, until=None):
    os.chdir(repo)
    
    time_args = ""
    if since:
        time_args += f" --since='{since}'"
    if until:
        time_args += f" --until='{until}'"
    
    cmd = f"git log --author={author}{time_args} --pretty='%ad' --date=format:'%Y-%m-%d' --numstat"
    output = subprocess.getoutput(cmd).split('\n')
    
    current_date = ""
    date_pattern = re.compile(r"\d{4}-\d{2}-\d{2}")  
    
    for line in output:
        if date_pattern.match(line):  
            current_date = line
        elif line.count('\t') == 2:  
            added, removed, _ = map(str.strip, line.split('\t'))
            
            if added.replace('-', '').isdigit() and removed.replace('-', '').isdigit():
                daily_line_stats[current_date]['added'] += int(added)
                daily_line_stats[current_date]['removed'] += int(removed)
                daily_line_stats[current_date]['total'] += int(added) - int(removed)
    
    os.chdir("..")

def plot_line_stats_per_day(author, daily_line_stats):
    dates = sorted(daily_line_stats.keys())
    added = [daily_line_stats[date]['added'] for date in dates]
    removed = [daily_line_stats[date]['removed'] for date in dates]
    total = [daily_line_stats[date]['total'] for date in dates]
    
    plt.figure(figsize=(15, 7))
    plt.stackplot(dates, added, removed, total, labels=['Added', 'Removed', 'Total'])
    plt.xlabel('Dates')
    plt.ylabel('Line counts (Log scale)')
    plt.yscale('log')  # Set the y-axis to a logarithmic scale
    plt.title(f'Line Stats Per Day Across All Repositories by {author}')
    plt.legend()
    plt.xticks(rotation=45, ha="right", rotation_mode="anchor")  # Rotate labels for better visibility
    plt.show()


def main():
    # [IMPORTANT] Change the config below before executing the script
    repos = ['hello-world']    # Modify this line to specify the local repo path you want to analyze
    author = "ShaneTsui"       # Modify this line to specify a different author
    since_date = '2023-09-01'  # Optional: specify the start date
    until_date = '2023-12-31'  # Optional: specify the end date
    
    line_stats = {}
    daily_commits = defaultdict(int)
    daily_line_stats = defaultdict(lambda: {'added': 0, 'removed': 0, 'total': 0})
    
    for repo in repos:
        if os.path.isdir(repo):
            analyze_repo(repo, author, line_stats, daily_commits, since=since_date, until=until_date)
            get_daily_line_stats_for_repo(repo, author, daily_line_stats, since=since_date, until=until_date)

    # Plotting
    plot_line_stats(author, line_stats)
    plot_daily_commits(author, daily_commits)
    plot_line_stats_per_day(author, daily_line_stats)

if __name__ == "__main__":
    main()