scripts/quick_analysis.py

import os
from PIL import Image
from collections import Counter

def analyze_images(directory):
    analysis_results = {}
    
    for root, dirs, files in os.walk(directory):
        if files:
            model_folder_name = os.path.basename(root)
            if model_folder_name not in analysis_results:
                analysis_results[model_folder_name] = {
                    'image_count': 0,
                    'total_size': 0,
                    'resolutions': Counter()
                }
            
            for file in files:
                file_path = os.path.join(root, file)
                
                # Count the image
                analysis_results[model_folder_name]['image_count'] += 1
                
                # Calculate the size of the image
                try:
                    with Image.open(file_path) as img:
                        # Get the size of the image in bytes
                        file_size = os.path.getsize(file_path)
                        analysis_results[model_folder_name]['total_size'] += file_size
                        
                        # Get image dimensions
                        width, height = img.size
                        analysis_results[model_folder_name]['resolutions'][(width, height)] += 1
                except Exception as e:
                    print(f"Error reading file {file_path}: {e}")
    
    return analysis_results

def print_and_log_analysis_results(analysis_results, dataset_name, log_file):
    # Determine the maximum length of model names
    max_model_length = max(len(model) for model in analysis_results.keys())
    model_column_width = max(max_model_length, 20)  # Ensure at least 20 characters
    
    # Define column widths
    image_count_width = 12
    total_size_width = 14
    resolution_width = 25
    
    # Create header
    header = f"{'Model':<{model_column_width}} | {'Image Count':>{image_count_width}} | {'Total Size (MB)':>{total_size_width}} | {'Most Common Resolution':<{resolution_width}}"
    separator = "-" * (model_column_width + image_count_width + total_size_width + resolution_width + 7)  # 7 for separators
    
    result_lines = []
    result_lines.append(f"Analysis for {dataset_name}:\n")
    result_lines.append(header + "\n")
    result_lines.append(separator + "\n")
    
    for model, data in analysis_results.items():
        total_size_mb = data['total_size'] / (1024 * 1024)
        most_common_resolution = data['resolutions'].most_common(1)
        
        if most_common_resolution:
            common_res = f"{most_common_resolution[0][0][0]}x{most_common_resolution[0][0][1]} ({most_common_resolution[0][1]} images)"
        else:
            common_res = "None"
        
        result_lines.append(f"{model:<{model_column_width}} | {data['image_count']:>{image_count_width}} | {total_size_mb:>{total_size_width}.2f} | {common_res:<{resolution_width}}\n")
    
    result_lines.append("\n")
    
    # Print to console
    for line in result_lines:
        print(line, end='')
    
    # Write to log file
    with open(log_file, 'a') as f:
        f.writelines(result_lines)

def main():
    # Define directories
    generated_dir = 'resampledEvalSet'
    real_dir = 'real'
    log_file = 'analysis_results.txt'
    
    # Clear the log file (optional, comment out if you want to append)
    with open(log_file, 'w') as f:
        pass
    
    # Analyze generated images
    generated_analysis_results = analyze_images(generated_dir)
    print_and_log_analysis_results(generated_analysis_results, "Generated Images", log_file)
    
    # Analyze real images
    real_analysis_results = analyze_images(real_dir)
    print_and_log_analysis_results(real_analysis_results, "Real Images", log_file)

if __name__ == "__main__":
    main()