fix bugs for recognizing upper or lower case

utils/in_place_oov_tokens.py

@@ -2,6 +2,7 @@
 import sys
 import argparse
 from utilities import (
+    opentext,
     process_tltchool_gigaspeech_interregnum_tokens
 )
 
@@ -19,27 +20,8 @@
                     default="Librispeech-model-mct-tdnnf/data/lang/words.txt",
                     type=str)
 
-parser.add_argument("--conf",
-                    default="Librispeech-model-mct-tdnnf/data/lang/words.txt",
-                    type=str)
-
 args = parser.parse_args()
 
-
-def opentext(file):
-    s = set()
-    t = dict()
-    with open(file, "r") as f:
-        for l in f.readlines():
-            _l = l.split()
-            uttid = _l[0]
-            words = _l[1:]
-            t[uttid] = " ".join(words)
-            # remove the utt_id part
-            for word in words:
-                s.add(word.lower())
-    return list(s), t
-
 def openwords(file):
     s = dict()
     with open(file, "r") as f: