github-review-analysis.py

# Run this script inside a git directory and have Githubs gh command in the path.

import subprocess
import json
from datetime import datetime
import statistics

def run_command(command):
    result = subprocess.run(command, shell=True, text=True, capture_output=True)
    if result.returncode != 0:
        print(f'Error: {result.stderr}')
        exit(1)
    return json.loads(result.stdout)

pr_data = run_command('gh pr list --json number,title,createdAt,reviews -s closed -s merged')

def get_pull_request_stats(pr_data):
    pr_stats = {}

    for pull_request in pr_data:
        pr_id = pull_request['number']
        pr_title = pull_request['title']
        created_at = datetime.fromisoformat(pull_request['createdAt'])

        if pull_request.get('reviews', []):
            for review in pull_request['reviews']:
                review_state = review['state'].lower()
                if review_state in ['approved', 'changes_requested']:
                    reviewer = review['author']['login']
                    review_time = (datetime.fromisoformat(review['submittedAt']) - created_at).total_seconds() / 3600  # Convert to hours

                    if reviewer in pr_stats:
                        if not any([pr[0] == pr_id for pr in pr_stats[reviewer]]):
                            pr_stats[reviewer].append((pr_id, review_time, pr_title))
                    else:
                        pr_stats[reviewer] = [(pr_id, review_time, pr_title)]

    return pr_stats

pull_request_stats = get_pull_request_stats(pr_data)

for reviewer, reviews in pull_request_stats.items():
    avg_review_time = sum([review[1] for review in reviews]) / len(reviews)
    median_review_time = statistics.median([review[1] for review in reviews])

    print(f'Reviewer: {reviewer} | Num PRs Reviewed: {len(reviews)} | Avg. Review Time: {"{:.2f} hours".format(avg_review_time) if avg_review_time <= 24 else "{:.0f} days and {:.2f} hours".format(avg_review_time // 24, avg_review_time % 24)} | Median Review Time: {"{:.2f} hours".format(median_review_time) if median_review_time <= 24 else "{:.0f} days and {:.2f} hours".format(median_review_time // 24, median_review_time % 24)}')

    for pr_id, review_time, pr_title in reviews:
        out = f'  First Review: {"{:.2f} hours".format(review_time) if review_time <= 24 else "{:.0f} days and {:.2f} hours".format(review_time // 24, review_time % 24)}'
        rpad = 40 - len(out)
        print(out + ' ' * rpad + f' #{pr_id} {pr_title}' )