validate_project.py

#!/usr/bin/env python3
"""
Pre-deployment validation script
Checks that the project is ready for HF deployment
"""

import os
import sys
from pathlib import Path

def check_file_exists(filename, description):
    """Check if a file exists and return status"""
    if os.path.exists(filename):
        size = os.path.getsize(filename)
        print(f"✅ {filename} ({size:,} bytes) - {description}")
        return True
    else:
        print(f"❌ Missing: {filename} - {description}")
        return False

def check_file_content(filename, required_content):
    """Check if file contains required content"""
    try:
        with open(filename, 'r', encoding='utf-8') as f:
            content = f.read()
        
        missing = []
        for item in required_content:
            if item not in content:
                missing.append(item)
        
        if missing:
            print(f"⚠️  {filename} missing: {', '.join(missing)}")
            return False
        else:
            print(f"✅ {filename} content validated")
            return True
    except Exception as e:
        print(f"❌ Error reading {filename}: {e}")
        return False

def validate_project():
    """Run all validation checks"""
    print("Pre-Deployment Validation")
    print("=" * 40)
    
    all_good = True
    
    # Required files
    print("\n📁 File Structure Check:")
    required_files = [
        ("docs/README.md", "HF Space configuration and docs"),
        ("src/app.py", "Main Gradio application"),
        ("src/mlx_campp.py", "MLX model implementation"), 
        ("src/conversion_utils.py", "Conversion utilities"),
        ("requirements.txt", "Dependencies"),
        ("tests/test_converter.py", "Test suite"),
        (".gitignore", "Git ignore patterns"),
        ("docs/DEPLOYMENT_GUIDE.md", "Deployment instructions")
    ]
    
    for filename, desc in required_files:
        if not check_file_exists(filename, desc):
            all_good = False
    
    # Check README.md has HF Space configuration
    print("\n🎛️  HF Space Configuration Check:")
    readme_required = [
        "title: CAM++ MLX Converter",
        "emoji: 🎤",
        "sdk: gradio",
        "app_file: app.py"
    ]
    
    if not check_file_content("docs/README.md", readme_required):
        all_good = False
    
    # Check app.py has required components
    print("\n⚙️  App.py Component Check:")
    app_required = [
        "import gradio as gr",
        "from mlx_campp import CAMPPModel",
        "from conversion_utils import ConversionUtils",
        "def convert_model",
        "interface.launch()"
    ]
    
    if not check_file_content("src/app.py", app_required):
        all_good = False
    
    # Check requirements.txt has essential packages
    print("\n📦 Dependencies Check:")
    req_required = [
        "mlx>=0.19.0",
        "gradio>=4.44.0", 
        "huggingface_hub",
        "torch",
        "numpy"
    ]
    
    if not check_file_content("requirements.txt", req_required):
        all_good = False
    
    # Check for common issues
    print("\nCommon Issues Check:")
    
    # Check if any Python files have import errors
    try:
        # Try importing without actually running MLX (which might not be installed)
        import ast
        
        for py_file in ["src/app.py", "src/mlx_campp.py", "src/conversion_utils.py"]:
            try:
                with open(py_file, 'r') as f:
                    ast.parse(f.read())
                print(f"✅ {py_file} syntax is valid")
            except SyntaxError as e:
                print(f"❌ Syntax error in {py_file}: {e}")
                all_good = False
    except Exception as e:
        print(f"⚠️  Could not validate Python syntax: {e}")
    
    # File size checks
    print("\n📏 File Size Check:")
    for filename in ["src/app.py", "src/mlx_campp.py", "src/conversion_utils.py"]:
        if os.path.exists(filename):
            size = os.path.getsize(filename)
            if size < 1000:  # Less than 1KB is suspicious
                print(f"⚠️  {filename} seems too small ({size} bytes)")
            elif size > 100000:  # More than 100KB is unusual for these files
                print(f"⚠️  {filename} seems very large ({size} bytes)")
            else:
                print(f"✅ {filename} size looks good ({size:,} bytes)")
    
    # Check for secrets/tokens in files
    print("\n🔐 Security Check:")
    security_ok = True
    sensitive_patterns = ["hf_", "token", "password", "secret", "key"]
    
    for py_file in ["src/app.py", "src/mlx_campp.py", "src/conversion_utils.py"]:
        try:
            with open(py_file, 'r') as f:
                content = f.read().lower()
            
            for pattern in sensitive_patterns:
                if pattern in content and "example" not in content:
                    # Check if it's actually a hardcoded secret
                    lines = content.split('\n')
                    for i, line in enumerate(lines):
                        if pattern in line and '=' in line and len(line.split('=')[1].strip().strip('"\'')) > 10:
                            print(f"⚠️  Possible hardcoded secret in {py_file}:{i+1}")
                            security_ok = False
                            break
        except Exception:
            pass
    
    if security_ok:
        print("✅ No hardcoded secrets detected")
    
    # Final assessment
    print("\n" + "=" * 40)
    if all_good:
        print("🎉 PROJECT IS READY FOR DEPLOYMENT!")
        print("\nNext steps:")
        print("1. Run: chmod +x deploy.sh && ./deploy.sh")
        print("2. Or follow DEPLOYMENT_GUIDE.md for manual deployment")
        print("3. Monitor your HF Space after deployment")
        return True
    else:
        print("❌ PROJECT NEEDS FIXES BEFORE DEPLOYMENT")
        print("\nPlease fix the issues above before deploying.")
        return False

def show_deployment_preview():
    """Show what the deployed space will look like"""
    print("\n🌐 Deployment Preview:")
    print("Your HF Space will have:")
    print("• URL: https://your-username-campp-mlx-converter.hf.space")
    print("• Title: 🎤 CAM++ MLX Converter")
    print("• Interface: Clean Gradio form for model conversion")
    print("• Features: PyTorch→MLX conversion with quantization")
    print("• Output: Models uploaded to mlx-community automatically")

if __name__ == "__main__":
    print("CAM++ MLX Converter - Pre-Deployment Validation")
    print("=" * 50)
    
    is_valid = validate_project()
    
    if is_valid:
        show_deployment_preview()
        sys.exit(0)
    else:
        sys.exit(1)