integrate_thb_test.py

#!/usr/bin/env python3
"""
Integration example for using TalkingHeadBench fake test dataset with FairTalking project.

This script shows how to:
1. Load THB fake test dataset
2. Create a compatible dataset class
3. Integrate with existing FairTalking data module
"""

import os
import pandas as pd
from pathlib import Path

class THBFakeTestDataset:
    """Dataset class for TalkingHeadBench fake test videos."""
    
    def __init__(self, csv_file: str, transform=None, target_transform=None):
        """
        Initialize the dataset.
        
        Args:
            csv_file: Path to CSV file with video paths
            transform: Optional transform to be applied on video
            target_transform: Optional transform to be applied on label
        """
        self.data = pd.read_csv(csv_file)
        self.transform = transform
        self.target_transform = target_transform
        
    def __len__(self):
        """Return the number of samples in dataset."""
        return len(self.data)
    
    def __getitem__(self, idx):
        """Get a sample from the dataset."""
        row = self.data.iloc[idx]
        
        # Video path
        video_path = row['video_path']
        
        # Label (all fake videos have label 1)
        label = row['label']
        
        # Generator information
        generator = row['generator']
        
        # In a real implementation, you would load the video here
        # For demonstration, we'll just return the path and metadata
        sample = {
            'video_path': video_path,
            'label': label,
            'generator': generator,
            'filename': row['filename']
        }
        
        # Apply transforms if specified
        if self.transform:
            sample = self.transform(sample)
        if self.target_transform:
            sample['label'] = self.target_transform(sample['label'])
        
        return sample

def create_thb_test_config():
    """Create configuration for THB test dataset integration."""
    
    config = {
        'thb_root': '/apdcephfs_gy4/share_303628665/joywu/dataset/TalkingHeadBench',
        'fake_test_csv': '/apdcephfs_gy4/share_303628665/joywu/dataset/TalkingHeadBench/fake_test_dataset.csv',
        'batch_size': 8,
        'num_workers': 4,
        'video_params': {
            'num_frames': 16,
            'frame_size': 224,
            'audio_seconds': 2.56
        }
    }
    
    return config

def integrate_with_fairtalking():
    """Show how to integrate THB dataset with existing FairTalking code."""
    
    print("=" * 60)
    print("THB Fake Test Dataset Integration with FairTalking")
    print("=" * 60)
    
    # Configuration
    config = create_thb_test_config()
    
    # Check if CSV file exists
    if not os.path.exists(config['fake_test_csv']):
        print(f"Error: CSV file not found: {config['fake_test_csv']}")
        return
    
    # Load dataset
    dataset = THBFakeTestDataset(config['fake_test_csv'])
    
    print(f"Dataset loaded successfully!")
    print(f"  Total samples: {len(dataset)}")
    print(f"  CSV file: {config['fake_test_csv']}")
    
    # Show sample data
    print(f"\nSample data from dataset:")
    for i in range(3):
        sample = dataset[i]
        print(f"  Sample {i+1}:")
        print(f"    Video: {Path(sample['video_path']).name}")
        print(f"    Generator: {sample['generator']}")
        print(f"    Label: {sample['label']}")
    
    # Integration with existing FairTalking datamodule
    print(f"\nIntegration with FairTalking DataModule:")
    print("1. Modify your datamodule.py to support THB dataset")
    print("2. Add a new method for THB test dataset loading")
    print("3. Update configuration to use THB test set")
    
    # Example modification to datamodule
    print(f"\nExample datamodule modification:")
    print("""
# In your datamodule.py, add this method:

def thb_fake_test_loader(self):
    '''Create DataLoader for THB fake test dataset.'''
    from .thb_dataset import THBFakeTestDataset
    
    dataset = THBFakeTestDataset(
        csv_file=self.config.thb_fake_test_csv,
        transform=self.test_transform
    )
    
    return DataLoader(
        dataset,
        batch_size=self.config.batch_size,
        num_workers=self.config.num_workers,
        shuffle=False
    )
""")

def main():
    """Main function to demonstrate THB dataset integration."""
    
    # Check dataset statistics
    csv_path = '/apdcephfs_gy4/share_303628665/joywu/dataset/TalkingHeadBench/fake_test_dataset.csv'
    
    if os.path.exists(csv_path):
        df = pd.read_csv(csv_path)
        
        print("Dataset Statistics:")
        print(f"  Total videos: {len(df)}")
        print(f"  Generators: {df['generator'].unique().tolist()}")
        print(f"  Label distribution: {df['label'].value_counts().to_dict()}")
        
        # Show generator distribution
        print(f"\nGenerator distribution:")
        for generator, count in df['generator'].value_counts().items():
            print(f"  {generator}: {count} videos")
    
    # Show integration example
    integrate_with_fairtalking()
    
    print(f"\nNext steps:")
    print("1. Create a THBFakeTestDataset class in your project")
    print("2. Modify your datamodule to support THB dataset")
    print("3. Update your training/testing scripts to use THB test set")
    print("4. Remember: All videos are fake (label=1), so you'll need real videos for binary classification")

if __name__ == "__main__":
    main()