mtf2json: added decoder for mixed encodings in MTF files (#6)

mtf2json/mtf2json.py

@@ -4,6 +4,7 @@
 """
 import json
 import re
+import codecs
 from math import ceil
 from pathlib import Path
 from typing import Dict, Any, Tuple, Union, Optional, List, cast, TextIO
@@ -71,6 +72,14 @@ class ConversionError(Exception):
 string_keys = ['model']
 
 
+def mixed_decoder(error: UnicodeError) -> Tuple[str, int]:
+    bs: bytes = error.object[error.start: error.end]
+    return bs.decode("cp1252"), error.start + 1
+
+
+codecs.register_error("mixed", mixed_decoder)
+
+
 def __rename_keys(obj: Any) -> Any:
     """
     Rename the keys in the given object according to
@@ -794,7 +803,7 @@ def read_mtf(path: Path) -> Dict[str, Any]:
     mech_data: Dict[str, Any] = {}
 
     current_section = None
-    with open(path, 'r') as file:
+    with open(path, 'r', encoding='utf8', errors='mixed') as file:
         __check_compat(file)
         for line in file:
             line = line.strip()