findlarge

#!/usr/bin/env python3

"""
findlarge: List the largest files in your TV or Movies library.

This script scans either your TV or Movies folder (as defined by TV_PATH and MOVIES_PATH),
recursively listing all files except those in excluded folders or with excluded keywords.
It then prints the N largest files (default 10), showing their human-readable size and full path.

Features:
- Excludes specific folders and files with certain keywords (customizable).
- Uses colorized output for file sizes (requires colorama and humanize).
- User chooses which library to scan and how many files to list.
- Useful for finding space hogs or cleaning up your media library.

Usage:
    python3 findlarge
    # Follow the prompts to select TV or Movies and how many files to list.

Requirements:
    pip install colorama humanize
"""

import os
import sys
import humanize
from colorama import Fore, Style, init

# Initialize colorama
init(autoreset=True)

TV_PATH = "/Media/TV"
MOVIES_PATH = "/Media/Movies"

EXCLUDED_FOLDERS = ["Folder1"]

EXCLUDED_KEYWORDS = ["Keyword1"]

def get_choice():
    print("Which folder do you want to scan?")
    print("1. TV")
    print("2. Movies")
    choice = input("Enter 1 or 2: ")
    if choice == "1":
        return TV_PATH
    elif choice == "2":
        return MOVIES_PATH
    else:
        print("Invalid choice.")
        sys.exit(1)

def get_limit():
    try:
        count_str = input("How many largest files do you want to list? [10]: ").strip()
        count = int(count_str) if count_str else 10
        return max(1, count)
    except ValueError:
        print("Invalid number.")
        sys.exit(1)

def list_largest_files(base_path, excluded_folders, excluded_keywords, limit):
    file_list = []

    for root, dirs, files in os.walk(base_path):
        # Skip excluded folders
        if any(excluded.lower() in root.lower() for excluded in excluded_folders):
            continue

        for file in files:
            full_path = os.path.join(root, file)
            # Skip excluded keywords
            if any(keyword.lower() in file.lower() for keyword in excluded_keywords):
                continue
            try:
                size = os.path.getsize(full_path)
                file_list.append((full_path, size))
            except OSError:
                continue

    # Sort by size descending
    file_list.sort(key=lambda x: x[1], reverse=True)

    print(f"\nTop {limit} largest files in '{base_path}':\n")
    for path, size in file_list[:limit]:
        size_colored = Fore.CYAN + humanize.naturalsize(size, binary=True) + Style.RESET_ALL
        print(f"{size_colored:>10} - {path}")

if __name__ == "__main__":
    selected_path = get_choice()
    file_limit = get_limit()
    list_largest_files(selected_path, EXCLUDED_FOLDERS, EXCLUDED_KEYWORDS, file_limit)