Github Star Summarizer

github_star_summarizer.py

"""
Gets a list of a user's starred repos, reads descriptions for all the repos, categorizes, summarizes and quantifies the user's interests.
"""

import requests
import os
from openai import OpenAI

# Initialize OpenAI client with API key from environment variable
client = OpenAI(api_key=os.getenv('OPENAI_API_KEY'))

def get_starred_repositories(username, github_token=None):
    url = f"https://api.github.com/users/{username}/starred"
    headers = {'Authorization': f'token {github_token}'} if github_token else {}
    repos = []
    page = 1

    while True:
        response = requests.get(url, headers=headers, params={'page': page, 'per_page': 30})
        
        if response.status_code != 200:
            raise Exception(f"Error fetching starred repositories: {response.status_code}")
        
        page_repos = response.json()
        if not page_repos:
            break
        
        repos.extend(page_repos)
        page += 1

    return repos

def get_descriptions_and_urls_from_starred_repos(starred_repos):
    descriptions_and_urls = []
    for repo in starred_repos:
        description = repo.get('description', '')
        url = repo.get('html_url', '')
        if description:
            descriptions_and_urls.append((description, url))
    return descriptions_and_urls

def summarize_descriptions_with_gpt4(descriptions_and_urls):
    prompt = """
    Based on the following list of project descriptions and their URLs, summarize the user's interests in JSON format.
    The JSON should have categories as keys, and for each category, include the count of repositories and an array of repository URLs.
    Example schema:
    {
        "category1": {
            "count": count1,
            "urls": ["url1", "url2", ...]
        },
        "category2": {
            "count": count2,
            "urls": ["url3", "url4", ...]
        },
        ...
    }
    Descriptions and URLs:
    """
    for description, url in descriptions_and_urls:
        prompt += f"\nDescription: {description}\nURL: {url}\n"
    
    chat_completion = client.chat.completions.create(
        messages=[
            {"role": "user", "content": prompt}
        ],
        model="gpt-4o-mini",
        response_format={"type": "json_object"}
    )
    
    summary = chat_completion.choices[0].message.content.strip()
    return summary

def rolling_summarize(descriptions_and_urls, chunk_size=30):
    chunk_summaries = []
    for i in range(0, len(descriptions_and_urls), chunk_size):
        chunk = descriptions_and_urls[i:i + chunk_size]
        print(f"Summarizing chunk {i // chunk_size + 1}...")
        chunk_summary = summarize_descriptions_with_gpt4(chunk)
        chunk_summaries.append(chunk_summary)
        print(f"\n--- Summary for chunk {i // chunk_size + 1} ---\n")
        print(chunk_summary)
        print("\n----------------------------------------\n")
    
    # Summarize all chunk summaries
    final_summary = summarize_descriptions_with_gpt4([(summary, "") for summary in chunk_summaries])
    return final_summary

def main():
    username = input("Enter GitHub username: ")
    github_token = input("Enter GitHub token (optional, press Enter to skip): ").strip() or None
    
    print(f"Fetching starred repositories for user '{username}'...")
    starred_repos = get_starred_repositories(username, github_token)
    
    print(f"Retrieved {len(starred_repos)} starred repositories.")
    
    descriptions_and_urls = get_descriptions_and_urls_from_starred_repos(starred_repos)
    
    if not descriptions_and_urls:
        print("No descriptions found for the starred repositories.")
        return
    
    print("Summarizing the user's interests...")
    final_summary = rolling_summarize(descriptions_and_urls)
    
    print("\nFinal Summary of All Chunk Summaries:\n")
    print(final_summary)

if __name__ == "__main__":
    main()