import os
import json
import argparse
from collections import defaultdict

rocky_version = "8.10"

def create_sitemap(directory):
    sitemap = defaultdict(lambda: defaultdict(dict))
    for root, dirs, files in os.walk(directory):
        for file in files:
            full_filepath = os.path.join(root, file)
            filepath = full_filepath.split(rocky_version, 1)[-1]
            
            # Exclude any path containing 'index.html'
            if 'index.html' in filepath or 'sitemap.json' in filepath or 'sitemap.xml' in filepath or 'list.json' in filepath or 'list.json.br' in filepath:
                continue
            
            filepath_parts = filepath.split('/')
            package_name = filepath_parts[1]
            man_type = filepath_parts[2]
            man_type_number = man_type.lstrip('man') if man_type.startswith('man') else man_type
            command_file = filepath_parts[3]
            command = command_file.split('.html', 1)[0]
            
            if filepath.startswith('/'):
                filepath = filepath[1:]
            
            fullname = f"{package_name} - {command}({man_type_number})"
            
            # Add command details to sitemap
            sitemap[package_name][command] = {
                "url": filepath,
                "mantype": man_type,
                "fullname": fullname
            }
    
    return sitemap

def convert_sitemap_to_json(sitemap, minify=False):
    if minify:
        return json.dumps(sitemap, separators=(',', ':'))
    return json.dumps(sitemap, indent=4)

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Generate sitemap JSON.')
    parser.add_argument('directory', type=str, help='Directory to scan for HTML files')
    parser.add_argument('--minify', action='store_true', help='Export minified JSON')
    args = parser.parse_args()

    sitemap = create_sitemap(args.directory)
    json_output = convert_sitemap_to_json(sitemap, minify=args.minify)
    
    print(json_output)