clean up process text for cer function

Signed-off-by: Paarth Neekhara <paarth.n@gmail.com>

Author: paarthneekhara

nemo/collections/tts/parts/utils/helpers.py

@@ -808,30 +808,18 @@ def g2p_backward_compatible_support(g2p_target: str) -> str:
 
 
 def process_text_for_cer(input_text):
-    """
-    Normalizes text for CER/WER calculation.
-    """
-    # Convert text to lowercase
-    lower_case_text = input_text.lower()
-
-    # Remove commas from text
-    no_comma_text = lower_case_text.replace(",", "")
-    # Replace "-" with spaces
-    no_dash_text = no_comma_text.replace("-", " ")
-    no_dash_text = no_dash_text.replace("'", "")
-    no_dash_text = no_dash_text.replace(";", "")
-    no_dash_text = no_dash_text.replace(".", "")
-
-    # Replace double spaces with single space
-    single_space_text = " ".join(no_dash_text.split())
-
-    single_space_text = single_space_text.translate(str.maketrans('', '', string.punctuation))
-
-    # Handle some common errors in ASR transcripts
-    single_space_text = single_space_text.replace("h t t p", "http")
-    single_space_text = single_space_text.replace("w w w", "www")
-
-    return single_space_text
+    """Normalizes text for CER/WER calculation."""
+    text = input_text.lower()
+    for char in [",", "'", ";", "."]:
+        text = text.replace(char, "")
+    text = text.replace("-", " ")
+    text = " ".join(text.split())
+    # Strip any remaining punctuation characters
+    text = text.translate(str.maketrans('', '', string.punctuation))
+    # Fix common ASR transcript artifacts
+    text = text.replace("h t t p", "http")
+    text = text.replace("w w w", "www")
+    return text
 
 
 def transcribe_with_whisper(