contributor_stats.py

#!/usr/bin/env python3

import subprocess
import csv
from collections import defaultdict
from datetime import datetime

def get_contributor_commits():
    try:
        # Get all commits with commit date and author, sorted by oldest first
        git_log = subprocess.check_output(
            ['git', 'log', '--format=%cI###%aN', '--reverse', '--use-mailmap', '--mailmap'],
            universal_newlines=True
        )
    except subprocess.CalledProcessError:
        print("Error: Failed to get git log. Are you in a git repository?")
        return None

    # Dictionary to store monthly stats and author counts
    monthly_stats = defaultdict(lambda: defaultdict(int))
    all_time_author_commits = defaultdict(int)  # Track total commits per author
    current_month = None

    # Process each commit chronologically
    for line in git_log.strip().split('\n'):
        date_str, email = line.split('###')
        date = datetime.fromisoformat(date_str)
        month_key = f"{date.year}-{date.month:02d}"

        # When we move to a new month, calculate stats for all authors
        if month_key != current_month:
            if current_month is not None:
                # Calculate stats based on cumulative commits for all authors
                monthly_stats[current_month] = defaultdict(int)
                for author, commit_count in all_time_author_commits.items():
                    if commit_count == 1:
                        monthly_stats[current_month]['1'] += 1
                    elif 2 <= commit_count < 5:
                        monthly_stats[current_month]['2-5'] += 1
                    elif 5 <= commit_count < 20:
                        monthly_stats[current_month]['5+'] += 1
                    elif 20 <= commit_count < 100:
                        monthly_stats[current_month]['20+'] += 1
                    elif commit_count >= 100:
                        monthly_stats[current_month]['100+'] += 1

                print(f"Completed month {current_month}: "
                      f"1 commit: {monthly_stats[current_month]['1']}, "
                      f"2-5 commits: {monthly_stats[current_month]['2-5']}, "
                      f"5+ commits: {monthly_stats[current_month]['5+']}, "
                      f"20+ commits: {monthly_stats[current_month]['20+']}, "
                      f"100+ commits: {monthly_stats[current_month]['100+']}")
            current_month = month_key

        # Simply update the author's total commit count
        all_time_author_commits[email] += 1

    # Calculate stats for the final month
    if current_month:
        monthly_stats[current_month] = defaultdict(int)
        for author, commit_count in all_time_author_commits.items():
            if commit_count == 1:
                monthly_stats[current_month]['1'] += 1
            elif 2 <= commit_count < 5:
                monthly_stats[current_month]['2-5'] += 1
            elif 5 <= commit_count < 20:
                monthly_stats[current_month]['5+'] += 1
            elif 20 <= commit_count < 100:
                monthly_stats[current_month]['20+'] += 1
            elif commit_count >= 100:
                monthly_stats[current_month]['100+'] += 1

        print(f"Completed month {current_month}: "
              f"1 commit: {monthly_stats[current_month]['1']}, "
              f"2-5 commits: {monthly_stats[current_month]['2-5']}, "
              f"5+ commits: {monthly_stats[current_month]['5+']}, "
              f"20+ commits: {monthly_stats[current_month]['20+']}, "
              f"100+ commits: {monthly_stats[current_month]['100+']}")

    return monthly_stats

def write_csv(stats):
    # Sort months chronologically
    months = sorted(stats.keys())

    with open('contributor_stats.csv', 'w', newline='') as f:
        writer = csv.writer(f)
        # Write header
        writer.writerow(['Month', '1 commit', '2-5 commits', '5+ commits', '20+ commits', '100+ commits'])

        # Write data for each month
        for month in months:
            # Convert YYYY-MM to YYYY-MM-01 for Excel compatibility
            year, month_num = month.split('-')
            excel_date = f"{year}-{month_num}-01"
            writer.writerow([
                excel_date,
                stats[month]['1'],
                stats[month]['2-5'],
                stats[month]['5+'],
                stats[month]['20+'],
                stats[month]['100+']
            ])

def main():
    print("Analyzing git repository...")
    stats = get_contributor_commits()
    write_csv(stats)
    print("Statistics have been written to contributor_stats.csv")

if __name__ == '__main__':
    main()