show-calc/tv_show_analyzer.py

#!/usr/bin/env python3
"""
TV Show Directory Analyzer
Scans a TV show directory structure and generates a CSV report with:
- Show metadata
- Season folder validation
- Directory structure issues
- Size metrics and ratios
"""

import os
import sys
import csv
from pathlib import Path
from collections import defaultdict


def get_folder_size_gb(path):
    """Calculate total size of a folder in GB"""
    total_size = 0
    try:
        for dirpath, dirnames, filenames in os.walk(path):
            for filename in filenames:
                filepath = os.path.join(dirpath, filename)
                try:
                    total_size += os.path.getsize(filepath)
                except (OSError, FileNotFoundError):
                    pass
    except (PermissionError, OSError):
        pass
    return total_size / (1024 ** 3)


def count_episodes_in_folder(folder_path):
    """Count files in a folder (proxy for episode count)"""
    try:
        return len([f for f in os.listdir(folder_path) 
                   if os.path.isfile(os.path.join(folder_path, f))])
    except (PermissionError, OSError):
        return 0


def is_season_folder(folder_name):
    """Check if folder name matches Season pattern"""
    name_lower = folder_name.lower().strip()
    # Matches: Season 1, season 1, Season1, S01, etc.
    if name_lower.startswith('season'):
        return True
    if name_lower.startswith('s') and len(name_lower) <= 4:
        try:
            int(name_lower[1:])
            return True
        except ValueError:
            pass
    return False


def is_featurette_folder(folder_name):
    """Check if folder appears to be featurettes"""
    name_lower = folder_name.lower()
    return 'featurette' in name_lower or 'special' in name_lower or 'bonus' in name_lower


def analyze_show_directory(show_path):
    """
    Analyze a single show directory
    Returns dict with analysis results
    """
    show_name = os.path.basename(show_path)
    
    try:
        subdirs = [d for d in os.listdir(show_path) 
                  if os.path.isdir(os.path.join(show_path, d))]
    except (PermissionError, OSError) as e:
        return {
            'show_name': show_name,
            'total_size_gb': 0,
            'season_folders': 0,
            'non_season_folders': [],
            'episode_count': 0,
            'size_per_episode': 0,
            'issues': f'Permission/Access Error: {str(e)}'
        }
    
    season_folders = []
    non_season_folders = []
    total_episodes = 0
    
    for subdir in subdirs:
        subdir_path = os.path.join(show_path, subdir)
        
        if is_season_folder(subdir):
            season_folders.append(subdir)
            total_episodes += count_episodes_in_folder(subdir_path)
        else:
            non_season_folders.append(subdir)
    
    # Calculate size
    total_size_gb = get_folder_size_gb(show_path)
    
    # Calculate size per episode
    size_per_episode = total_size_gb / total_episodes if total_episodes > 0 else 0
    
    # Identify issues
    issues = []
    if not season_folders:
        issues.append("NO_SEASON_FOLDERS")
    
    non_flagged_issues = [f for f in non_season_folders if not is_featurette_folder(f)]
    if non_flagged_issues:
        issues.append(f"UNEXPECTED_FOLDERS: {', '.join(non_flagged_issues)}")
    
    featurette_folders = [f for f in non_season_folders if is_featurette_folder(f)]
    
    return {
        'show_name': show_name,
        'total_size_gb': round(total_size_gb, 2),
        'season_folders': len(season_folders),
        'season_folder_names': ', '.join(sorted(season_folders)) if season_folders else 'NONE',
        'non_season_folders': ', '.join(non_season_folders) if non_season_folders else 'None',
        'featurette_folders': ', '.join(featurette_folders) if featurette_folders else 'None',
        'episode_count': total_episodes,
        'size_per_episode': round(size_per_episode, 3),
        'issues': '; '.join(issues) if issues else 'OK'
    }


def main():
    if len(sys.argv) > 1:
        directory = sys.argv[1]
    else:
        directory = input("Enter the path to your TV shows directory: ").strip()
    
    directory = os.path.expanduser(directory)
    
    if not os.path.isdir(directory):
        print(f"Error: Directory not found: {directory}")
        sys.exit(1)
    
    print(f"Scanning directory: {directory}")
    print("This may take a while for large collections...\n")
    
    # Get all show directories (first level subdirectories)
    try:
        show_folders = [d for d in os.listdir(directory) 
                       if os.path.isdir(os.path.join(directory, d))]
    except (PermissionError, OSError) as e:
        print(f"Error accessing directory: {e}")
        sys.exit(1)
    
    if not show_folders:
        print("No directories found in the specified path.")
        sys.exit(1)
    
    results = []
    for i, show_folder in enumerate(sorted(show_folders), 1):
        show_path = os.path.join(directory, show_folder)
        print(f"[{i}/{len(show_folders)}] Processing: {show_folder}")
        
        analysis = analyze_show_directory(show_path)
        results.append(analysis)
    
    # Write to CSV
    output_file = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'show_analysis.csv')
    
    fieldnames = [
        'show_name',
        'total_size_gb',
        'season_folders',
        'season_folder_names',
        'non_season_folders',
        'featurette_folders',
        'episode_count',
        'size_per_episode',
        'issues'
    ]
    
    try:
        with open(output_file, 'w', newline='', encoding='utf-8') as csvfile:
            writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
            writer.writeheader()
            writer.writerows(results)
        
        print(f"\n✓ Analysis complete!")
        print(f"✓ Results saved to: {output_file}")
        print(f"✓ Analyzed {len(results)} shows")
        
    except (PermissionError, IOError) as e:
        print(f"Error writing CSV file: {e}")
        sys.exit(1)


if __name__ == '__main__':
    main()