ignored some errors, ensured compatibility with python 3.7

nordmann2023.py

@@ -321,7 +321,7 @@ class Nordmann(
         )
 
         base_paths: Dict[str, str]
-        base_paths = dl_manager.download_and_extract(
+        base_paths = dl_manager.download_and_extract( # pyright: ignore
             url_or_urls=urls
         )
 
@@ -340,7 +340,7 @@ class Nordmann(
 
         return generators
 
-    def _generate_examples(
+    def _generate_examples( # pyright: ignore
         self,
         candidates: Dict[str, Candidate],
         base_paths: Dict[str, str]

utils.py

@@ -1,19 +1,19 @@
-from typing import Dict, TypeVar, Callable, List, Hashable, Literal, Union, Optional, Tuple, Collection, Iterable
+from typing import Dict, TypeVar, Callable, List, Hashable, Union, Optional, Tuple, Collection, Iterable
 from xml.etree import ElementTree
 import unicodedata
 from pathlib import Path
 import re
 
 Paths = Tuple[str, ...]
-Language = Literal['de', 'en']
+Language = str # Literal['de', 'en']
 Translation = Dict[
     Language, str
 ]
 Sample = Dict[
-    Literal['translation'], Translation
+    str, Translation # Literal['translation'], Translation
 ]
 SampleBatch = Dict[
-    Literal['translation'], List[Translation]
+    str, List[Translation] # Literal['translation'], List[Translation]
 ]
 
 _H1 = TypeVar('_H1', bound=Hashable)
@@ -102,9 +102,7 @@ unicode_control_characters = (
 def normalize(
     strip_whitespaces: bool,
     clean_control_characters: bool,
-    enforce_unicode_form: Optional[
-        Literal['NFC', 'NFKC', 'NFD', 'NFKD']
-    ] = None
+    enforce_unicode_form: Optional[str] = None # Optional[Literal['NFC', 'NFKC', 'NFD', 'NFKD']] = None
 ):
     regex_pattern = re.compile(
         pattern='[' + unicode_control_characters + ']+'
@@ -136,12 +134,12 @@ def normalize(
             translation = dict_map(
                 input=translation, map_fn=lambda key, value: (
                     key, unicodedata.normalize(
-                        enforce_unicode_form, value
+                        enforce_unicode_form, value # pyright: ignore
                     )
                 )
             )
 
-        sample['translation'] = translation
+        sample['translation'] = translation # pyright: ignore
         return sample
 
     return normalize_fn