Update app.py

app.py

@@ -5,6 +5,7 @@ import difflib
 import pandas as pd
 from Levenshtein import distance as lev_distance
 import whisper
+import string
 
 # Load Whisper model once (choose "small" or "medium" for better results)
 #model = whisper.load_model("small")
@@ -30,6 +31,8 @@ def get_error_type(asr_word, correct_word):
 
 def compare_hindi_sentences(expected, transcribed):
     expected_words = expected.strip().split()
+    expected_clean = expected.translate(str.maketrans('', '', string.punctuation))
+    expected_words = expected_clean.strip().split()
     transcribed_words = transcribed.strip().split()
     matcher = difflib.SequenceMatcher(None, transcribed_words, expected_words)
     errors = []