from flask import Flask, render_template, request, jsonify
import requests
import base64
import markdown
from bs4 import BeautifulSoup
import os
import mimetypes
import json
app = Flask(__name__)
GITHUB_API = "https://api.github.com/repos/"
def get_repo_contents(repo_url):
"""Extract contents from GitHub repo URL"""
try:
parts = repo_url.rstrip('/').split('/')
owner, repo = parts[-2], parts[-1]
api_url = f"{GITHUB_API}{owner}/{repo}/contents"
response = requests.get(api_url)
response.raise_for_status()
return owner, repo, response.json()
except Exception as e:
return None, None, str(e)
def process_file_content(file_info, owner, repo):
"""Process individual file content"""
content = ""
file_path = file_info['path']
if file_info['type'] == 'file':
file_url = f"{GITHUB_API}{owner}/{repo}/contents/{file_path}"
file_response = requests.get(file_url)
file_data = file_response.json()
if 'content' in file_data:
file_extension = file_path.split('.')[-1] if '.' in file_path else ''
mime_type, _ = mimetypes.guess_type(file_path)
is_text = (mime_type and mime_type.startswith('text')) or file_extension in ['py', 'md', 'txt', 'js', 'html', 'css', 'json']
if is_text:
try:
decoded_content = base64.b64decode(file_data['content']).decode('utf-8')
# Special handling for JSON files
if file_extension == 'json':
try:
json_data = json.loads(decoded_content)
formatted_json = json.dumps(json_data, indent=2)
content = f"### File: {file_path}\n```json\n{formatted_json}\n```\n\n"
except json.JSONDecodeError:
content = f"### File: {file_path}\n```json\n{decoded_content}\n```\n[Note: Invalid JSON format]\n\n"
else:
content = f"### File: {file_path}\n```{(file_extension if file_extension else 'text')}\n{decoded_content}\n```\n\n"
except UnicodeDecodeError:
content = f"### File: {file_path}\n[Text content could not be decoded - possibly corrupted or non-UTF-8 text]\n\n"
else:
content = f"### File: {file_path}\n[Binary file - {file_data['size']} bytes]\n\n"
return content
def create_markdown_document(repo_url):
"""Create markdown document from repo contents"""
owner, repo, contents = get_repo_contents(repo_url)
if isinstance(contents, str):
return f"Error: {contents}"
markdown_content = f"# Repository: {owner}/{repo}\n\n"
markdown_content += "Below are the contents of all files in the repository:\n\n"
for item in contents:
markdown_content += process_file_content(item, owner, repo)
return markdown_content
@app.route('/')
def index():
return render_template('index.html')
@app.route('/process', methods=['POST'])
def process_repo():
repo_url = request.json.get('repo_url')
if not repo_url:
return jsonify({'error': 'Please provide a repository URL'}), 400
markdown_content = create_markdown_document(repo_url)
html_content = markdown.markdown(markdown_content)
return jsonify({
'markdown': markdown_content,
'html': html_content
})
html_template = """
GitHub Repo to Markdown
GitHub Repository to Markdown Converter
Enter a GitHub repository URL (e.g., https://github.com/username/repository)
Markdown Output:
Preview:
"""
if not os.path.exists('templates'):
os.makedirs('templates')
with open('templates/index.html', 'w') as f:
f.write(html_template)
if __name__ == '__main__':
app.run(host="0.0.0.0", port=7860, debug=True)