app.py

import gradio as gr
import requests
from datetime import datetime, timezone, timedelta
import os
import google.generativeai as genai

def get_all_commits_in_range(owner, repo_name, start_date, end_date, token):
    """
    Fetch all unique commits across all branches within a date range.
    Uses a more efficient approach by fetching commits once and filtering by date.
    """
    headers = {"Authorization": f"token {token}"}
    base_url = f"https://api.github.com/repos/{owner}/{repo_name}"
    
    # Get all branches
    try:
        branches_resp = requests.get(f"{base_url}/branches", headers=headers)
        branches_resp.raise_for_status()
        branches = [b['name'] for b in branches_resp.json()]
    except requests.exceptions.RequestException as e:
        return None, f"Error fetching branches: {str(e)}"
    
    # Define date range
    since = f"{start_date}T00:00:00Z"
    until = f"{end_date}T23:59:59Z"
    
    all_commits = {}
    
    # Fetch commits from all branches
    for branch in branches:
        try:
            params = {"sha": branch, "since": since, "until": until, "per_page": 100}
            commits_resp = requests.get(f"{base_url}/commits", headers=headers, params=params)
            commits_resp.raise_for_status()
            commits = commits_resp.json()
            
            # Add unique commits (keyed by SHA to avoid duplicates)
            for commit in commits:
                sha = commit["sha"]
                if sha not in all_commits:
                    all_commits[sha] = commit
        except requests.exceptions.RequestException as e:
            continue  # Skip branches that cause errors
    
    # Sort by commit date
    sorted_commits = sorted(
        all_commits.values(),
        key=lambda c: datetime.strptime(c["commit"]["author"]["date"], "%Y-%m-%dT%H:%M:%SZ")
    )
    
    return sorted_commits, None

def format_commits(commits, timezone_offset_hours=5, timezone_offset_minutes=45):
    """
    Format commits with timezone conversion and readable output.
    """
    formatted = []
    
    for commit in commits:
        sha = commit["sha"][:8]  # Short SHA for readability
        message = commit["commit"]["message"].split('\n')[0]  # First line only
        author = commit["commit"]["author"]["name"]
        utc_time_str = commit["commit"]["author"]["date"]
        
        # Parse and convert timezone
        utc_time = datetime.strptime(utc_time_str, "%Y-%m-%dT%H:%M:%SZ")
        utc_time = utc_time.replace(tzinfo=timezone.utc)
        
        # Apply timezone offset
        local_offset = timedelta(hours=timezone_offset_hours, minutes=timezone_offset_minutes)
        local_time = utc_time.astimezone(timezone(local_offset))
        
        formatted_line = (
            f"{sha} | {author:20s} | {local_time.strftime('%Y-%m-%d %H:%M:%S')} | {message}"
        )
        formatted.append(formatted_line)
    
    return "\n".join(formatted)

def summarize_with_gemini(repo_name, owner, commit_lines, start_date, end_date, api_key):
    """
    Generate a comprehensive summary using Gemini with an improved prompt.
    """
    genai.configure(api_key=api_key)
    model = genai.GenerativeModel("gemini-2.0-flash-exp")
    
    date_range = f"{start_date}" if start_date == end_date else f"{start_date} to {end_date}"
    
    prompt = f"""You are a technical project analyst reviewing Git commit history.

Repository: {owner}/{repo_name}
Date Range: {date_range}
Total Commits: {len(commit_lines.split(chr(10)))}

Commit History:
{commit_lines}

Please provide a comprehensive analysis with the following sections:

1. **Executive Summary**: A 2-3 sentence overview of the work accomplished during this period.

2. **Key Developments**: Highlight the most significant changes, features, or fixes (3-5 bullet points).

3. **Development Timeline**: Chronological narrative of how the work progressed throughout the day/period.

4. **Author Contributions**: For each author, provide:
   - Total number of commits
   - Time range of their work
   - Main focus areas and specific accomplishments
   - Notable commits with their purpose

5. **Technical Insights**: Identify patterns such as:
   - Types of work (features, bugs, refactoring, documentation)
   - Areas of the codebase that received attention
   - Any potential concerns or areas needing follow-up

6. **Impact Assessment**: Evaluate the overall progress and its significance to the project.

Format your response with clear headings and bullet points for readability."""

    try:
        response = model.generate_content(prompt)
        return response.text
    except Exception as e:
        return f"Error generating summary: {str(e)}"

def analyze_commits(repo_url, start_date, end_date, github_token, gemini_key):
    """
    Main function to orchestrate the commit analysis.
    """
    # Parse repository URL
    try:
        if "github.com/" in repo_url:
            parts = repo_url.rstrip('/').split('github.com/')[-1].split('/')
            owner = parts[0]
            repo_name = parts[1].replace('.git', '')
        else:
            # Assume format: owner/repo
            owner, repo_name = repo_url.split('/')
    except:
        return "❌ Invalid repository format. Use 'owner/repo' or full GitHub URL."
    
    # Validate tokens
    if not github_token or not gemini_key:
        return "❌ Please provide both GitHub and Gemini API tokens."
    
    # Fetch commits
    status_msg = f"🔍 Fetching commits from {owner}/{repo_name}...\n"
    commits, error = get_all_commits_in_range(owner, repo_name, start_date, end_date, github_token)
    
    if error:
        return f"{status_msg}❌ {error}"
    
    if not commits:
        return f"{status_msg}ℹ️ No commits found in the specified date range."
    
    status_msg += f"✅ Found {len(commits)} unique commits across all branches.\n\n"
    
    # Format commits
    formatted_commits = format_commits(commits)
    status_msg += "📝 Commit History:\n" + "="*80 + "\n"
    status_msg += formatted_commits + "\n\n"
    
    # Generate summary
    status_msg += "🤖 Generating AI summary...\n" + "="*80 + "\n"
    summary = summarize_with_gemini(repo_name, owner, formatted_commits, start_date, end_date, gemini_key)
    status_msg += summary
    
    return status_msg

# Gradio Interface
with gr.Blocks(title="Git Commit Story Analyzer", theme=gr.themes.Soft()) as app:
    gr.Markdown("""
    # 📚 Git Commit Story Analyzer
    
    Transform your Git commit history into comprehensive narratives using AI.
    This tool fetches commits from **all branches** and generates detailed insights.
    """)
    
    with gr.Row():
        with gr.Column(scale=2):
            repo_input = gr.Textbox(
                label="Repository",
                placeholder="owner/repo or https://github.com/owner/repo",
                info="Enter GitHub repository in format 'owner/repo' or paste the full URL"
            )
            
            with gr.Row():
                start_date_input = gr.Textbox(
                    label="Start Date",
                    value=datetime.now().strftime("%Y-%m-%d"),
                    placeholder="YYYY-MM-DD"
                )
                end_date_input = gr.Textbox(
                    label="End Date", 
                    value=datetime.now().strftime("%Y-%m-%d"),
                    placeholder="YYYY-MM-DD"
                )
            
            github_token_input = gr.Textbox(
                label="GitHub Personal Access Token",
                type="password",
                placeholder="ghp_xxxxxxxxxxxx",
                info="Required for API access. Generate at: Settings → Developer settings → Personal access tokens"
            )
            
            gemini_key_input = gr.Textbox(
                label="Gemini API Key",
                type="password", 
                placeholder="AIzaSy...",
                info="Get your API key from Google AI Studio: https://makersuite.google.com/app/apikey"
            )
            
            analyze_btn = gr.Button("🚀 Analyze Commits", variant="primary", size="lg")
        
        with gr.Column(scale=1):
            gr.Markdown("""
            ### 💡 Tips
            
            - **Single Day**: Set both dates to the same day
            - **Date Range**: Use different start and end dates
            - **All Branches**: Automatically analyzes commits from all branches
            - **Timezone**: Displays times in Nepal timezone (UTC+5:45)
            
            ### 🔒 Privacy
            
            Your tokens are only used for this session and are never stored.
            """)
    
    output = gr.Textbox(
        label="Analysis Results",
        lines=25,
        max_lines=50,
        show_copy_button=True
    )
    
    analyze_btn.click(
        fn=analyze_commits,
        inputs=[repo_input, start_date_input, end_date_input, github_token_input, gemini_key_input],
        outputs=output
    )
    
    gr.Markdown("""
    ---
    ### 📖 How to Use
    
    1. **Enter Repository**: Provide the GitHub repository (e.g., `torvalds/linux` or full URL)
    2. **Select Date Range**: Choose single day or date range for analysis
    3. **Add Tokens**: Provide your GitHub Personal Access Token and Gemini API Key
    4. **Analyze**: Click the button to generate your commit story
    
    **Note**: The tool fetches commits from all branches and removes duplicates automatically.
    """)

if __name__ == "__main__":
    app.launch()