test_github_scanner.py

import asyncio
import sys
import os

# Add current directory to sys.path to ensure we can import github_scanner
sys.path.append(os.path.dirname(os.path.abspath(__file__)))

import github_scanner

class MockWebSocket:
    async def send_json(self, data):
        print(f"[WebSocket Mock] {data}")

async def main():
    if len(sys.argv) < 2:
        print("Usage: python test_github_scanner.py <domain>")
        return

    domain = sys.argv[1]
    print(f"Testing GitHub Scanner against: {domain}")
    
    # Check if token is loaded
    headers = await github_scanner.get_github_headers()
    if "Authorization" in headers:
        print("✅ GITHUB_TOKEN is found and loaded.")
    else:
        print("⚠️  GITHUB_TOKEN NOT FOUND. You might hit rate limits.")

    mock_ws = MockWebSocket()
    
    # Run the main scan
    await github_scanner.scan_github(domain, mock_ws)
    
    print("\n--- Testing Scanner Logic with Dummy Data ---")
    dummy_content = "Here is a leak: https://my-secret-bucket.s3.amazonaws.com/database.dump"
    print(f"Scanning content: '{dummy_content}'")
    findings = github_scanner.scan_content(dummy_content, "test_file.txt", "http://github.com/test/repo")
    print(f"Findings: {findings}")
    
    if findings and findings[0]['type'] == 'bucket' and 'my-secret-bucket' in findings[0]['match']:
        print("✅ Regex verification PASSED: Bucket detected.")
    else:
        print("❌ Regex verification FAILED: Bucket NOT detected.")

if __name__ == "__main__":
    asyncio.run(main())