Update app.py

app.py

@@ -7,15 +7,15 @@ from pydub import AudioSegment
 from pydub.generators import Sine
 import io
 
-MODEL_NAME = "openai/whisper-tiny"
+MODEL_NAME = "openai/whisper-large-v3"
 BATCH_SIZE = 8
-# device = 0 if torch.cuda.is_available() else "cpu"
+device = 0 if torch.cuda.is_available() else "cpu"
 
 pipe = pipeline(
     task="automatic-speech-recognition",
     model=MODEL_NAME,
     chunk_length_s=30,
-    # device=device,
+    device=device,
 )
 
 arabic_bad_Words = pd.read_csv("arabic_bad_words_dataset.csv")
@@ -23,14 +23,11 @@ english_bad_Words = pd.read_csv("english_bad_words_dataset.csv")
 
 
 def clean_english_word(word):
-    # Use regex to remove special characters, punctuation, and spaces around words
     cleaned_text = re.sub(r'^[\s\W_]+|[\s\W_]+$', '', word)
-    return cleaned_text
+    return cleaned_text.lower()
 
 def clean_arabic_word(word):
-    # Define a regex pattern to match any non-Arabic letter character
     pattern = r'[^\u0600-\u06FF]'
-    # Replace any character matching the pattern with an empty string
     cleaned_word = re.sub(pattern, '', word)
     return cleaned_word
 
@@ -43,14 +40,15 @@ def classifier(word_list_with_timestamp, language):
         list_to_search = set(english_bad_Words["words"])
         for item in word_list_with_timestamp:
             word = clean_english_word(item['text'])
-            if word in list_to_search:
+            if word.lower() in list_to_search:
                 foul_words.append(word)
                 negative_timestamps.append(item['timestamp'])
+                break
     else:
         list_to_search = list(arabic_bad_Words["words"])
         for item in word_list_with_timestamp:
             word = clean_arabic_word(item['text'])
-            for word_in_list in list_to_search:
+            for word in list_to_search:
                 if word_in_list == word:
                     foul_words.append(word)
                     negative_timestamps.append(item['timestamp'])
@@ -65,7 +63,6 @@ def generate_bleep(duration_ms, frequency=1000):
 
 def mute_audio_range(audio_filepath, ranges, bleep_frequency=800):
     audio = AudioSegment.from_file(audio_filepath)
-
     for range in ranges:
         start_time = range[0] - 0.1
         end_time = range[-1] + 0.1
@@ -123,10 +120,32 @@ examples = [
         ["arabic_english_audios/audios/arabic_hate_audio_1.mp3", 'Arabic', 'transcribe', 'word'],
         ["arabic_english_audios/audios/arabic_hate_audio_2.flac", 'Arabic', 'transcribe', 'word'],
         ["arabic_english_audios/audios/arabic_hate_audio_3.flac", 'Arabic', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/arabic_hate_audio_31.mp3", 'Arabic', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/arabic_hate_audio_32.mp3", 'Arabic', 'transcribe', 'word'],
         ["arabic_english_audios/audios/english_audio_1.wav", 'English', 'transcribe', 'word'],
         ["arabic_english_audios/audios/english_audio_2.mp3", 'English', 'transcribe', 'word'],
         ["arabic_english_audios/audios/english_audio_3.mp3", 'English', 'transcribe', 'word'],
+    
+        ["arabic_english_audios/audios/english_audio_31.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_32.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_33.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_34.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_35.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_36.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_37.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_38.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_39.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_40.mp3", 'English', 'transcribe', 'word'],
         ["arabic_english_audios/audios/english_audio_4.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_41.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_42.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_43.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_44.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_45.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_46.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_48.mp3", 'English', 'transcribe', 'word'],
+        ["arabic_english_audios/audios/english_audio_49.mp3", 'English', 'transcribe', 'word'],
+        
         ["arabic_english_audios/audios/english_audio_5.mp3", 'English', 'transcribe', 'word'],
         ["arabic_english_audios/audios/english_audio_6.wav", 'English', 'transcribe', 'word']
     ]