fix_encoding.py

#!/usr/bin/env python3
"""
Script to fix encoding of SGML files from ISO-8859-1 to UTF-8
The original files were encoded in ISO-8859-1, which caused issues with
special characters like accented letters.
"""

import os
import glob
from pathlib import Path

def fix_file_encoding(file_path, source_encoding, target_encoding):
    """Convert a file from source encoding to target encoding"""
    try:
        with open(file_path, 'r', encoding=source_encoding) as file:
            content = file.read()
        
        with open(file_path, 'w', encoding=target_encoding) as file:
            file.write(content)
        
        print(f"Fixed encoding for: {file_path}")
        
    except Exception as e:
        print(f"Error processing {file_path}: {e}")

def main():
    # Get the directory where the script is located
    script_dir = Path(__file__).parent
    efe_dir = script_dir / 'efe'
    
    if not efe_dir.exists():
        print(f"Error: Directory '{efe_dir}' not found!")
        return
    
    # Find all SGML files in the directory
    sgml_files = glob.glob(str(efe_dir / "*.sgml"))
    if not sgml_files:
        print("No SGML files found in the efe directory!")
        return
    
    print(f"Found {len(sgml_files)} SGML files to process...")
    print("-" * 50)
    
    for file_path in sorted(sgml_files):
        # this is the important part
        fix_file_encoding(file_path, 'iso-8859-1', 'utf-8')
    
    print("-" * 50)
    print(f"Completed processing {len(sgml_files)} files.")

if __name__ == "__main__":
    main()