Github metrics aggregation

GitHub Metrics Aggregation

Aggregate raw GitHub PR and issue data into clean, structured metrics suitable for reports and dashboards.

Metrics Structure

The final report follows this structure:

{
  "pr": {
    "total": <int>,           # All PRs created during period
    "merged": <int>,          # PRs with state == "MERGED"
    "closed": <int>,          # PRs with state == "CLOSED" (and not merged)
    "avg_merge_days": <float>,# Average days from creation to merge
    "top_contributor": <str>  # login of contributor with most PRs
  },
  "issue": {
    "total": <int>,           # All issues created during period
    "bug": <int>,             # Issues with "bug" in any label
    "resolved_bugs": <int>    # Bug reports that were closed (any time)
  }
}

PR Metrics

Total PRs

Simply count all PR objects returned from the GitHub API query for the date range.

total_prs = len(prs)

Merged vs. Closed

• Merged: Count PRs where state == "MERGED"
• Closed (not merged): Count PRs where state == "CLOSED" and not already counted as merged

merged_prs = [pr for pr in prs if pr['state'] == 'MERGED']
closed_prs = [pr for pr in prs if pr['state'] == 'CLOSED']

metrics['merged'] = len(merged_prs)
metrics['closed'] = len(closed_prs)

Average Merge Time

See the time-to-merge-analysis skill for the algorithm. Include only merged PRs.

Top Contributor

Identify the GitHub user (author.login) who opened the most PRs:

from collections import Counter

authors = [pr['author']['login'] for pr in prs if pr.get('author')]
author_counts = Counter(authors)
top_contributor = author_counts.most_common(1)[0][0]  # returns login string

Edge case: If no PRs exist, return empty string or "N/A".

Issue Metrics

Total Issues

Count all issues returned for the date range.

total_issues = len(issues)

Bug Reports

Use the bug-report-identification skill to count issues with "bug" in any label name.

Resolved Bugs

Count issues where:

1. At least one label contains "bug" (substring match, case-insensitive)
2. state == "CLOSED"

Note: "Resolved bugs" counts bugs that were created during the period and are now closed (as of the query date), not necessarily closed during the period. Adjust this if the requirement is "bugs closed during the period" instead.

def is_bug(issue):
    return any('bug' in label['name'].lower() for label in issue.get('labels', []))

resolved_bugs = sum(1 for issue in issues if is_bug(issue) and issue['state'] == 'CLOSED')

Complete Aggregation Function (Python)

import json
from datetime import datetime
from collections import Counter
from typing import List, Dict

def aggregate_metrics(prs: List[Dict], issues: List[Dict]) -> Dict:
    """
    Aggregate GitHub data into structured metrics.

    Args:
        prs: List of PR objects from GitHub API
        issues: List of issue objects from GitHub API

    Returns:
        Metrics dict matching the report structure
    """
    # PR Metrics
    merged_prs = [pr for pr in prs if pr['state'] == 'MERGED']
    closed_prs = [pr for pr in prs if pr['state'] == 'CLOSED']

    # Average merge time
    if merged_prs:
        total_days = 0
        for pr in merged_prs:
            created = datetime.fromisoformat(pr['createdAt'].replace('Z', '+00:00'))
            merged = datetime.fromisoformat(pr['mergedAt'].replace('Z', '+00:00'))
            days = (merged - created).total_seconds() / 86400
            total_days += days
        avg_merge_days = round(total_days / len(merged_prs), 1)
    else:
        avg_merge_days = 0.0

    # Top contributor
    authors = [pr['author']['login'] for pr in prs if pr.get('author')]
    if authors:
        author_counts = Counter(authors)
        top_contributor = author_counts.most_common(1)[0][0]
    else:
        top_contributor = ""

    # Issue Metrics
    def is_bug(issue):
        return any('bug' in label['name'].lower() for label in issue.get('labels', []))

    bug_count = sum(1 for issue in issues if is_bug(issue))
    resolved_bugs = sum(1 for issue in issues if is_bug(issue) and issue['state'] == 'CLOSED')

    return {
        "pr": {
            "total": len(prs),
            "merged": len(merged_prs),
            "closed": len(closed_prs),
            "avg_merge_days": avg_merge_days,
            "top_contributor": top_contributor
        },
        "issue": {
            "total": len(issues),
            "bug": bug_count,
            "resolved_bugs": resolved_bugs
        }
    }

Output to File

Write the aggregated metrics to JSON:

with open('report.json', 'w') as f:
    json.dump(metrics, f, indent=2)

Validation Checklist

Before finalizing the report:

• ✓ pr.total ≥ pr.merged + pr.closed
• ✓ pr.avg_merge_days is rounded to 1 decimal place
• ✓ pr.top_contributor is non-empty (or handle empty case gracefully)
• ✓ issue.bug ≤ issue.total
• ✓ issue.resolved_bugs ≤ issue.bug
• ✓ All numeric values are correct type (int or float as specified)
• ✓ All string values are non-null (use empty string if needed)