cleanup_episodes.py

import json
import os
import glob

def get_valid_episode_indexes():
    """Read the episodes.jsonl file and extract valid episode indexes."""
    valid_indexes = set()
    
    try:
        with open('meta/episodes.jsonl', 'r', encoding='utf-8') as f:
            for line_num, line in enumerate(f, 1):
                line = line.strip()
                if not line:  # Skip empty lines
                    continue
                try:
                    episode_data = json.loads(line)
                    if 'episode_index' in episode_data:
                        valid_indexes.add(episode_data['episode_index'])
                    else:
                        print(f"Warning: No 'episode_index' found on line {line_num}")
                except json.JSONDecodeError as e:
                    print(f"Warning: Could not parse JSON on line {line_num}: {e}")
                    print(f"Line content: '{line[:100]}...' (truncated)")
                    continue
    except FileNotFoundError:
        print("Error: meta/episodes.jsonl file not found!")
        return set()
    except Exception as e:
        print(f"Error reading episodes.jsonl: {e}")
        return set()
    
    return valid_indexes

def cleanup_video_files():
    """Remove video files that don't correspond to valid episode indexes."""
    valid_indexes = get_valid_episode_indexes()
    
    if not valid_indexes:
        print("No valid episode indexes found. Exiting.")
        return
    
    print(f"Found {len(valid_indexes)} valid episode indexes: {sorted(list(valid_indexes)[:10])}{'...' if len(valid_indexes) > 10 else ''}")
    
    # Find all video directories
    video_dirs = glob.glob('videos/chunk-*/observation.images.base/')
    
    if not video_dirs:
        print("No video directories found matching pattern 'videos/chunk-*/observation.images.base/'")
        return
    
    print(f"Found {len(video_dirs)} video directories")
    
    removed_files = []
    kept_files = []
    
    for video_dir in video_dirs:
        print(f"\nProcessing directory: {video_dir}")
        # Find all episode video files in this directory
        video_files = glob.glob(os.path.join(video_dir, 'episode_*.mp4'))
        
        for video_file in video_files:
            # Extract episode number from filename
            filename = os.path.basename(video_file)
            # Extract number from episode_XXXXXX.mp4 format
            episode_num_str = filename.replace('episode_', '').replace('.mp4', '')
            
            try:
                # Handle leading zeros by converting to int
                episode_num = int(episode_num_str.lstrip('0') or '0')
                
                if episode_num not in valid_indexes:
                    print(f"  Removing: {filename} (episode {episode_num} not in valid list)")
                    os.remove(video_file)
                    removed_files.append(video_file)
                else:
                    print(f"  Keeping: {filename} (episode {episode_num} is valid)")
                    kept_files.append(video_file)
                    
            except ValueError:
                print(f"  Warning: Could not parse episode number from {filename}")
    
    print(f"\n" + "="*50)
    print(f"CLEANUP SUMMARY:")
    print(f"Files removed: {len(removed_files)}")
    print(f"Files kept: {len(kept_files)}")
    
    if removed_files:
        print(f"\nRemoved files:")
        for file in removed_files:
            print(f"  - {file}")

if __name__ == "__main__":
    cleanup_video_files()