Fix for nltk#3294

Author: soras

nltk/tokenize/punkt.py

@@ -1760,13 +1760,13 @@ def load_punkt_params(lang_dir):
     pdec = PunktDecoder()
     # Make a new Parameters object:
     params = PunktParameters()
-    with open(f"{lang_dir}/collocations.tab") as f:
+    with open(f"{lang_dir}/collocations.tab", "r", encoding="utf-8") as f:
         params.collocations = pdec.tab2tups(f)
-    with open(f"{lang_dir}/sent_starters.txt") as f:
+    with open(f"{lang_dir}/sent_starters.txt", "r", encoding="utf-8") as f:
         params.sent_starters = pdec.txt2set(f)
-    with open(f"{lang_dir}/abbrev_types.txt") as f:
+    with open(f"{lang_dir}/abbrev_types.txt", "r", encoding="utf-8") as f:
         params.abbrev_types = pdec.txt2set(f)
-    with open(f"{lang_dir}/ortho_context.tab") as f:
+    with open(f"{lang_dir}/ortho_context.tab", "r", encoding="utf-8") as f:
         params.ortho_context = pdec.tab2intdict(f)
     return params