random_frame_extractor.py

#!/usr/bin/env python3
"""
Script to randomly extract k frames from WebP files in data/shots/ and save them as PNG files.

Usage:
    python random_frame_extractor.py <k> <output_dir>
    
Where:
    k: Number of frames to extract
    output_dir: Directory to save the extracted frames

Input:  data/shots/sample-XXX-Y.webp (randomly selected)
Output: <output_dir>/sample-XXX-Y-fZ.png (where Z is the frame number, e.g., f20 for frame 20)
"""

import os
import sys
import random
from pathlib import Path
from PIL import Image
import re
import argparse


def get_frame_count(webp_path):
    """
    Get the total number of frames in a WebP file.
    
    Args:
        webp_path (Path): Path to the WebP file
        
    Returns:
        int: Number of frames in the file
    """
    try:
        with Image.open(webp_path) as img:
            frame_count = 0
            try:
                while True:
                    img.seek(frame_count)
                    frame_count += 1
            except EOFError:
                pass
            return max(1, frame_count)  # At least 1 frame for static images
    except Exception as e:
        print(f"Error reading {webp_path}: {e}")
        return 0


def extract_random_frames(webp_path, output_dir, num_frames_to_extract):
    """
    Extract random frames from a WebP file and save them as PNG files.
    
    Args:
        webp_path (Path): Path to the input WebP file
        output_dir (Path): Directory to save the extracted frames
        num_frames_to_extract (int): Number of frames to extract from this file
        
    Returns:
        list: List of extracted frame filenames
    """
    try:
        # Parse filename to get sample and shot numbers
        filename = webp_path.stem  # e.g., "sample-123-2"
        match = re.match(r'sample-(\d+)-(\d+)', filename)
        if not match:
            print(f"Warning: Filename {filename} doesn't match expected pattern")
            return []
        
        sample_num = match.group(1)
        shot_num = match.group(2)
        
        # Get total frame count
        total_frames = get_frame_count(webp_path)
        if total_frames == 0:
            return []
        
        # Determine how many frames to extract (don't exceed available frames)
        frames_to_extract = min(num_frames_to_extract, total_frames)
        
        # Select random frame indices
        if total_frames == 1:
            selected_frames = [0]
        else:
            selected_frames = sorted(random.sample(range(total_frames), frames_to_extract))
        
        extracted_files = []
        
        with Image.open(webp_path) as img:
            for frame_idx in selected_frames:
                try:
                    # Seek to the specific frame
                    img.seek(frame_idx)
                    
                    # Convert to RGB if necessary
                    frame = img.convert('RGB')
                    
                    # Create output filename: sample-XXX-Y-fZ.png
                    output_filename = f"sample-{sample_num}-{shot_num}-f{frame_idx}.png"
                    output_path = output_dir / output_filename
                    
                    # Save the frame as PNG
                    frame.save(output_path)
                    extracted_files.append(output_filename)
                    print(f"Extracted frame {frame_idx} from {webp_path.name}: {output_filename}")
                    
                except Exception as e:
                    print(f"Error extracting frame {frame_idx} from {webp_path}: {e}")
        
        return extracted_files
        
    except Exception as e:
        print(f"Error processing {webp_path}: {e}")
        return []


def main():
    """Main function to randomly extract k frames from WebP files"""
    
    # Parse command line arguments
    parser = argparse.ArgumentParser(description='Randomly extract k frames from WebP files')
    parser.add_argument('k', type=int, help='Number of frames to extract')
    parser.add_argument('output_dir', type=str, help='Output directory for extracted frames')
    parser.add_argument('--frames-per-file', type=int, default=3, 
                       help='Maximum frames to extract per WebP file (default: 3)')
    parser.add_argument('--seed', type=int, help='Random seed for reproducible results')
    
    args = parser.parse_args()
    
    if args.k <= 0:
        print("Error: k must be a positive integer")
        sys.exit(1)
    
    # Set random seed if provided
    if args.seed is not None:
        random.seed(args.seed)
        print(f"Using random seed: {args.seed}")
    
    # Set up paths
    script_dir = Path(__file__).parent
    shots_dir = script_dir / "data" / "shots"
    output_dir = Path(args.output_dir)
    
    # Check if shots directory exists
    if not shots_dir.exists():
        print(f"Error: Shots directory not found: {shots_dir}")
        sys.exit(1)
    
    # Create output directory if it doesn't exist
    output_dir.mkdir(parents=True, exist_ok=True)
    print(f"Output directory: {output_dir}")
    
    # Find all WebP files
    webp_files = list(shots_dir.glob('sample-*.webp'))
    
    if not webp_files:
        print(f"No WebP files found in {shots_dir}")
        sys.exit(1)
    
    print(f"Found {len(webp_files)} WebP files available")
    
    # Calculate how many files we need to process
    frames_per_file = args.frames_per_file
    files_needed = (args.k + frames_per_file - 1) // frames_per_file  # Ceiling division
    files_needed = min(files_needed, len(webp_files))
    
    # Randomly select WebP files
    selected_files = random.sample(webp_files, files_needed)
    print(f"Randomly selected {len(selected_files)} WebP files")
    
    total_extracted = 0
    extracted_files = []
    
    # Process each selected WebP file
    for i, webp_file in enumerate(selected_files):
        remaining_frames = args.k - total_extracted
        if remaining_frames <= 0:
            break
        
        # Calculate how many frames to extract from this file
        frames_from_this_file = min(frames_per_file, remaining_frames)
        
        print(f"\nProcessing ({i+1}/{len(selected_files)}): {webp_file.name}")
        print(f"Extracting up to {frames_from_this_file} frames...")
        
        files = extract_random_frames(webp_file, output_dir, frames_from_this_file)
        extracted_files.extend(files)
        total_extracted += len(files)
    
    print(f"\n=== Summary ===")
    print(f"Target frames: {args.k}")
    print(f"Frames extracted: {total_extracted}")
    print(f"Files processed: {len(selected_files)}")
    print(f"Output directory: {output_dir}")
    
    if total_extracted < args.k:
        print(f"Warning: Only extracted {total_extracted} frames out of {args.k} requested")
    
    print(f"\nExtracted files:")
    for filename in extracted_files:
        print(f"  {filename}")


if __name__ == "__main__":
    main()