git_upload.py

#!/usr/bin/env python3
"""
Git-based Hugging Face upload for Bengali AI model
Repository: megharudushi/Sheikh
"""

import os
import subprocess

def git_upload_bengali_ai():
    """Upload using Git commands"""
    
    print("🚀 Git-based Hugging Face Upload")
    print("=" * 40)
    print("Repository: megharudushi/Sheikh")
    
    # Check if model directory exists
    if not os.path.exists("ready_bengali_ai"):
        print("❌ Error: ready_bengali_ai directory not found!")
        return False
    
    # Show files to upload
    files = os.listdir("ready_bengali_ai")
    print(f"\n📁 Files to upload ({len(files)} total):")
    for file in sorted(files):
        size = os.path.getsize(f"ready_bengali_ai/{file}") / (1024*1024)
        print(f"  📄 {file} ({size:.1f}MB)")
    
    print("\n" + "="*50)
    print("🔧 Git-based upload requires:")
    print("1. git-xet installed")
    print("2. SSH key configured")
    print("3. Repository access permissions")
    print("="*50)
    
    try:
        # Step 1: Install git-xet (if not installed)
        print("\n📦 Checking git-xet installation...")
        try:
            result = subprocess.run(["git", "xet", "--version"], capture_output=True, text=True)
            if result.returncode == 0:
                print("✅ git-xet already installed")
                print(f"Version: {result.stdout.strip()}")
            else:
                print("⚠️ git-xet not found, attempting install...")
                subprocess.run(["pip", "install", "git-xet"], check=True)
                print("✅ git-xet installed")
        except Exception as e:
            print(f"❌ git-xet install failed: {e}")
            print("Please install manually: https://hf.co/docs/hub/git-xet")
        
        # Step 2: Clone repository
        print("\n📥 Cloning repository...")
        repo_url = "git@hf.co:megharudushi/Sheikh"
        
        try:
            # Remove existing directory if it exists
            if os.path.exists("Sheikh"):
                subprocess.run(["rm", "-rf", "Sheikh"], check=True)
            
            # Clone repository
            result = subprocess.run(["git", "clone", repo_url], capture_output=True, text=True)
            if result.returncode == 0:
                print("✅ Repository cloned successfully")
            else:
                print(f"❌ Clone failed: {result.stderr}")
                print("This might mean:")
                print("- Repository doesn't exist yet")
                print("- SSH key not configured")
                print("- No access permissions")
                return False
                
        except Exception as e:
            print(f"❌ Repository clone error: {e}")
            return False
        
        # Step 3: Copy model files
        print("\n📤 Copying model files...")
        try:
            # Copy all files from ready_bengali_ai to Sheikh directory
            for file in os.listdir("ready_bengali_ai"):
                src = f"ready_bengali_ai/{file}"
                dst = f"Sheikh/{file}"
                subprocess.run(["cp", src, dst], check=True)
            print(f"✅ Copied {len(files)} files")
        except Exception as e:
            print(f"❌ File copy failed: {e}")
            return False
        
        # Step 4: Git commit and push
        print("\n📤 Committing and pushing...")
        try:
            os.chdir("Sheikh")
            
            # Git add all files
            subprocess.run(["git", "add", "."], check=True)
            
            # Git commit
            commit_message = "Add complete Bengali AI model - 355M parameters with tokenizer"
            subprocess.run(["git", "commit", "-m", commit_message], check=True)
            
            # Git push
            print("📤 Pushing to Hugging Face...")
            subprocess.run(["git", "push"], check=True)
            
            print("\n🎉 SUCCESS! Model uploaded via Git!")
            print("🌐 Repository: https://huggingface.co/megharudushi/Sheikh")
            
            return True
            
        except subprocess.CalledProcessError as e:
            print(f"❌ Git operation failed: {e}")
            print("\n🔧 Manual steps needed:")
            print("1. Configure SSH key: https://huggingface.co/settings/keys")
            print("2. Ensure repository access")
            print("3. Run: git push")
            return False
            
    except Exception as e:
        print(f"❌ Upload failed: {e}")
        return False

def create_manual_instructions():
    """Create manual upload instructions"""
    
    instructions = """# Manual Git Upload Instructions

## Prerequisites
1. **Install git-xet**: https://hf.co/docs/hub/git-xet
2. **Configure SSH key**: https://huggingface.co/settings/keys
3. **Repository access**: Ensure you have write access to megharudushi/Sheikh

## Manual Commands
```bash
# 1. Clone repository
git clone git@hf.co:megharudushi/Sheikh

# 2. Copy model files
cp ready_bengali_ai/* Sheikh/

# 3. Commit and push
cd Sheikh
git add .
git commit -m "Add Bengali AI model - 355M parameters"
git push
```

## Files to Upload
- model.bin (1.4GB) - Model weights
- tokenizer.json (3.4MB) - Tokenizer config
- vocab.json (780KB) - Vocabulary
- config.json (13KB) - Model config
- README.md - Model documentation
- And 6 other configuration files

## After Upload
Your model will be available at:
https://huggingface.co/megharudushi/Sheikh
"""
    
    with open("MANUAL_GIT_UPLOAD.md", "w", encoding="utf-8") as f:
        f.write(instructions)
    
    print("📄 Created manual instructions: MANUAL_GIT_UPLOAD.md")

if __name__ == "__main__":
    print("🇧🇩 BANGLI AI - GIT UPLOAD")
    print("=" * 35)
    
    # Create manual instructions
    create_manual_instructions()
    
    # Attempt git upload
    success = git_upload_bengali_ai()
    
    if success:
        print("\n🎊 Git upload successful!")
        print("Your Bengali AI is now live on Hugging Face!")
    else:
        print("\n⚠️ Git upload requires manual setup.")
        print("Please follow the manual instructions in MANUAL_GIT_UPLOAD.md")
        print("Or use the Python API approach instead.")