scripts/dataset_tonic/setup_hf_dataset.py

#!/usr/bin/env python3
"""
Setup script for Hugging Face Dataset repository for Trackio experiments
"""

import os
import json
from datetime import datetime
from pathlib import Path
from datasets import Dataset
from huggingface_hub import HfApi, create_repo
import subprocess

def get_username_from_token(token: str) -> str:
    """Get username from HF token with fallback to CLI"""
    try:
        # Try API first
        api = HfApi(token=token)
        user_info = api.whoami()
        
        # Handle different possible response formats
        if isinstance(user_info, dict):
            # Try different possible keys for username
            username = (
                user_info.get('name') or 
                user_info.get('username') or 
                user_info.get('user') or 
                None
            )
        elif isinstance(user_info, str):
            # If whoami returns just the username as string
            username = user_info
        else:
            username = None
            
        if username:
            print(f"✅ Got username from API: {username}")
            return username
        else:
            print("⚠️  Could not get username from API, trying CLI...")
            return get_username_from_cli(token)
            
    except Exception as e:
        print(f"⚠️  API whoami failed: {e}")
        print("⚠️  Trying CLI fallback...")
        return get_username_from_cli(token)

def get_username_from_cli(token: str) -> str:
    """Fallback method to get username using CLI"""
    try:
        # Set HF token for CLI
        os.environ['HF_TOKEN'] = token
        
        # Get username using CLI
        result = subprocess.run(
            ["huggingface-cli", "whoami"],
            capture_output=True,
            text=True,
            timeout=30
        )
        
        if result.returncode == 0:
            username = result.stdout.strip()
            if username:
                print(f"✅ Got username from CLI: {username}")
                return username
            else:
                print("⚠️  CLI returned empty username")
                return None
        else:
            print(f"⚠️  CLI whoami failed: {result.stderr}")
            return None
            
    except Exception as e:
        print(f"⚠️  CLI fallback failed: {e}")
        return None

def setup_trackio_dataset():
    """Set up the Trackio experiments dataset on Hugging Face Hub"""
    
    # Configuration - get from environment variables with fallbacks
    hf_token = os.environ.get('HF_TOKEN')
    
    if not hf_token:
        print("❌ HF_TOKEN not found. Please set the HF_TOKEN environment variable.")
        print("You can get your token from: https://huggingface.co/settings/tokens")
        return False
    
    username = get_username_from_token(hf_token)
    if not username:
        print("❌ Could not determine username from token. Please check your token.")
        return False
    
    print(f"✅ Authenticated as: {username}")
    
    # Use username in dataset repository if not specified
    dataset_repo = os.environ.get('TRACKIO_DATASET_REPO', f'{username}/trackio-experiments')
    
    print(f"🚀 Setting up Trackio dataset: {dataset_repo}")
    print(f"🔧 Using dataset repository: {dataset_repo}")
    
    # Initial experiment data
    initial_experiments = [
        {
            'experiment_id': 'exp_20250720_130853',
            'name': 'petite-elle-l-aime-3',
            'description': 'SmolLM3 fine-tuning experiment',
            'created_at': '2025-07-20T11:20:01.780908',
            'status': 'running',
            'metrics': json.dumps([
                {
                    'timestamp': '2025-07-20T11:20:01.780908',
                    'step': 25,
                    'metrics': {
                        'loss': 1.1659,
                        'grad_norm': 10.3125,
                        'learning_rate': 7e-08,
                        'num_tokens': 1642080.0,
                        'mean_token_accuracy': 0.75923578992486,
                        'epoch': 0.004851130919895701
                    }
                },
                {
                    'timestamp': '2025-07-20T11:26:39.042155',
                    'step': 50,
                    'metrics': {
                        'loss': 1.165,
                        'grad_norm': 10.75,
                        'learning_rate': 1.4291666666666667e-07,
                        'num_tokens': 3324682.0,
                        'mean_token_accuracy': 0.7577659255266189,
                        'epoch': 0.009702261839791402
                    }
                },
                {
                    'timestamp': '2025-07-20T11:33:16.203045',
                    'step': 75,
                    'metrics': {
                        'loss': 1.1639,
                        'grad_norm': 10.6875,
                        'learning_rate': 2.1583333333333334e-07,
                        'num_tokens': 4987941.0,
                        'mean_token_accuracy': 0.7581205774843692,
                        'epoch': 0.014553392759687101
                    }
                },
                {
                    'timestamp': '2025-07-20T11:39:53.453917',
                    'step': 100,
                    'metrics': {
                        'loss': 1.1528,
                        'grad_norm': 10.75,
                        'learning_rate': 2.8875e-07,
                        'num_tokens': 6630190.0,
                        'mean_token_accuracy': 0.7614579878747463,
                        'epoch': 0.019404523679582803
                    }
                }
            ]),
            'parameters': json.dumps({
                'model_name': 'HuggingFaceTB/SmolLM3-3B',
                'max_seq_length': 12288,
                'use_flash_attention': True,
                'use_gradient_checkpointing': False,
                'batch_size': 8,
                'gradient_accumulation_steps': 16,
                'learning_rate': 3.5e-06,
                'weight_decay': 0.01,
                'warmup_steps': 1200,
                'max_iters': 18000,
                'eval_interval': 1000,
                'log_interval': 25,
                'save_interval': 2000,
                'optimizer': 'adamw_torch',
                'beta1': 0.9,
                'beta2': 0.999,
                'eps': 1e-08,
                'scheduler': 'cosine',
                'min_lr': 3.5e-07,
                'fp16': False,
                'bf16': True,
                'ddp_backend': 'nccl',
                'ddp_find_unused_parameters': False,
                'save_steps': 2000,
                'eval_steps': 1000,
                'logging_steps': 25,
                'save_total_limit': 5,
                'eval_strategy': 'steps',
                'metric_for_best_model': 'eval_loss',
                'greater_is_better': False,
                'load_best_model_at_end': True,
                'data_dir': None,
                'train_file': None,
                'validation_file': None,
                'test_file': None,
                'use_chat_template': True,
                'chat_template_kwargs': {'add_generation_prompt': True, 'no_think_system_message': True},
                'enable_tracking': True,
                'trackio_url': 'https://tonic-test-trackio-test.hf.space',
                'trackio_token': None,
                'log_artifacts': True,
                'log_metrics': True,
                'log_config': True,
                'experiment_name': 'petite-elle-l-aime-3',
                'dataset_name': 'legmlai/openhermes-fr',
                'dataset_split': 'train',
                'input_field': 'prompt',
                'target_field': 'accepted_completion',
                'filter_bad_entries': True,
                'bad_entry_field': 'bad_entry',
                'packing': False,
                'max_prompt_length': 12288,
                'max_completion_length': 8192,
                'truncation': True,
                'dataloader_num_workers': 10,
                'dataloader_pin_memory': True,
                'dataloader_prefetch_factor': 3,
                'max_grad_norm': 1.0,
                'group_by_length': True
            }),
            'artifacts': json.dumps([]),
            'logs': json.dumps([]),
            'last_updated': datetime.now().isoformat()
        },
        {
            'experiment_id': 'exp_20250720_134319',
            'name': 'petite-elle-l-aime-3-1',
            'description': 'SmolLM3 fine-tuning experiment',
            'created_at': '2025-07-20T11:54:31.993219',
            'status': 'running',
            'metrics': json.dumps([
                {
                    'timestamp': '2025-07-20T11:54:31.993219',
                    'step': 25,
                    'metrics': {
                        'loss': 1.166,
                        'grad_norm': 10.375,
                        'learning_rate': 7e-08,
                        'num_tokens': 1642080.0,
                        'mean_token_accuracy': 0.7590958896279335,
                        'epoch': 0.004851130919895701
                    }
                },
                {
                    'timestamp': '2025-07-20T11:54:33.589487',
                    'step': 25,
                    'metrics': {
                        'gpu_0_memory_allocated': 17.202261447906494,
                        'gpu_0_memory_reserved': 75.474609375,
                        'gpu_0_utilization': 0,
                        'cpu_percent': 2.7,
                        'memory_percent': 10.1
                    }
                }
            ]),
            'parameters': json.dumps({
                'model_name': 'HuggingFaceTB/SmolLM3-3B',
                'max_seq_length': 12288,
                'use_flash_attention': True,
                'use_gradient_checkpointing': False,
                'batch_size': 8,
                'gradient_accumulation_steps': 16,
                'learning_rate': 3.5e-06,
                'weight_decay': 0.01,
                'warmup_steps': 1200,
                'max_iters': 18000,
                'eval_interval': 1000,
                'log_interval': 25,
                'save_interval': 2000,
                'optimizer': 'adamw_torch',
                'beta1': 0.9,
                'beta2': 0.999,
                'eps': 1e-08,
                'scheduler': 'cosine',
                'min_lr': 3.5e-07,
                'fp16': False,
                'bf16': True,
                'ddp_backend': 'nccl',
                'ddp_find_unused_parameters': False,
                'save_steps': 2000,
                'eval_steps': 1000,
                'logging_steps': 25,
                'save_total_limit': 5,
                'eval_strategy': 'steps',
                'metric_for_best_model': 'eval_loss',
                'greater_is_better': False,
                'load_best_model_at_end': True,
                'data_dir': None,
                'train_file': None,
                'validation_file': None,
                'test_file': None,
                'use_chat_template': True,
                'chat_template_kwargs': {'add_generation_prompt': True, 'no_think_system_message': True},
                'enable_tracking': True,
                'trackio_url': 'https://tonic-test-trackio-test.hf.space',
                'trackio_token': None,
                'log_artifacts': True,
                'log_metrics': True,
                'log_config': True,
                'experiment_name': 'petite-elle-l-aime-3-1',
                'dataset_name': 'legmlai/openhermes-fr',
                'dataset_split': 'train',
                'input_field': 'prompt',
                'target_field': 'accepted_completion',
                'filter_bad_entries': True,
                'bad_entry_field': 'bad_entry',
                'packing': False,
                'max_prompt_length': 12288,
                'max_completion_length': 8192,
                'truncation': True,
                'dataloader_num_workers': 10,
                'dataloader_pin_memory': True,
                'dataloader_prefetch_factor': 3,
                'max_grad_norm': 1.0,
                'group_by_length': True
            }),
            'artifacts': json.dumps([]),
            'logs': json.dumps([]),
            'last_updated': datetime.now().isoformat()
        }
    ]
    
    try:
        # Initialize HF API
        api = HfApi(token=hf_token)
        
        # First, try to create the dataset repository
        print(f"Creating dataset repository: {dataset_repo}")
        try:
            create_repo(
                repo_id=dataset_repo,
                token=hf_token,
                repo_type="dataset",
                exist_ok=True,
                private=True  # Make it private for security
            )
            print(f"✅ Dataset repository created: {dataset_repo}")
        except Exception as e:
            print(f"⚠️  Repository creation failed (may already exist): {e}")
        
        # Create dataset
        dataset = Dataset.from_list(initial_experiments)
        
        # Get the project root directory (2 levels up from this script)
        project_root = Path(__file__).parent.parent.parent
        templates_dir = project_root / "templates" / "datasets"
        readme_path = templates_dir / "readme.md"
        
        # Read README content if it exists
        readme_content = None
        if readme_path.exists():
            with open(readme_path, 'r', encoding='utf-8') as f:
                readme_content = f.read()
            print(f"✅ Found README template: {readme_path}")
        
        # Push to HF Hub
        print("Pushing dataset to HF Hub...")
        dataset.push_to_hub(
            dataset_repo,
            token=hf_token,
            private=False  # Make it private for security
        )
        
        # Create README separately if available
        if readme_content:
            try:
                print("Uploading README.md...")
                api.upload_file(
                    path_or_fileobj=readme_content.encode('utf-8'),
                    path_in_repo="README.md",
                    repo_id=dataset_repo,
                    repo_type="dataset",
                    token=hf_token
                )
                print("📝 Uploaded README.md successfully")
            except Exception as e:
                print(f"⚠️  Could not upload README: {e}")
        
        print(f"✅ Successfully created dataset: {dataset_repo}")
        print(f"📊 Added {len(initial_experiments)} experiments")
        if readme_content:
            print("📝 Included README from templates")
        print("🔓 Dataset is public (accessible to everyone)")
        print(f"👤 Created by: {username}")
        print("\n🎯 Next steps:")
        print("1. Set HF_TOKEN in your Hugging Face Space environment")
        print("2. Deploy the updated app.py to your Space")
        print("3. The app will now load experiments from the dataset")
        
        return True
        
    except Exception as e:
        print(f"❌ Failed to create dataset: {e}")
        print("\nTroubleshooting:")
        print("1. Check that your HF token has write permissions")
        print("2. Verify the dataset repository name is available")
        print("3. Try creating the dataset manually on HF first")
        return False

if __name__ == "__main__":
    setup_trackio_dataset()