Update app.py

app.py

@@ -41,188 +41,9 @@ os.system('pip install ./indic-trans/.')
 #hf_writer = gr.HuggingFaceDatasetSaver(HF_TOKEN, "asr_demo")
 
 from indictrans import Transliterator
-trn = Transliterator(source='ori', target='eng', build_lookup=True)
-def transcribe_odiya(speech):
-    text = p1(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-
-###################################
-def transcribe_odiya_model2(speech):
-    text = odia_model2(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-
-def transcribe_odiya_eng_model2(speech):
-    trn = Transliterator(source='ori', target='eng', build_lookup=True)
-    text = odia_model2(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-
-########################################
-def cleanhtml(raw_html):
-    cleantext = re.sub(r'<.*?>', '', raw_html)
-    return cleantext
-
-def transcribe_hindi(speech):
-    text = p2(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-def transcribe_kannada(speech):
-    text = p3(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-def transcribe_telugu(speech):
-    text = p4(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-    
-def transcribe_bangala(speech):
-    text = p5(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-
-def transcribe_assamese_LM(speech):
-    text = p8(speech)["text"]
-    text = cleanhtml(text)
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-    
-def transcribe_assamese_model2(speech):
-    text = p7(speech)["text"]
-    text = cleanhtml(text)
-    if text is None:
-        return "Error: ASR returned None"
-    return text
-    
-def transcribe_odiya_eng(speech):
-    trn = Transliterator(source='ori', target='eng', build_lookup=True)
-    text = p1(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-    
-def transcribe_ban_eng(speech):
-    trn = Transliterator(source='ben', target='eng', build_lookup=True)
-    text = p5(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-
-def transcribe_hin_eng(speech):
-    trn = Transliterator(source='hin', target='eng', build_lookup=True)
-    text = p2(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-
-def transcribe_kan_eng(speech):
-    trn = Transliterator(source='kan', target='eng', build_lookup=True)
-    text = p3(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-    
-def transcribe_tel_eng(speech):
-    trn = Transliterator(source='tel', target='eng', build_lookup=True)
-    text = p4(speech)["text"]
-    if text is None:
-        return "Error: ASR returned None"
-    sentence = trn.transform(text)
-    if sentence is None:
-        return "Error: Transliteration returned None"
-    replaced_words = replace_words(sentence)
-    processed_sentence = process_doubles(replaced_words)
-    return process_transcription(processed_sentence)
-
 
-def process_transcription(input_sentence):
-    word_to_code_map = {}
-    code_to_word_map = {}
-
-    transcript_1 = sentence_to_transcript(input_sentence, word_to_code_map)
-    if transcript_1 is None:
-        return "Error: Transcript conversion returned None"
-
-    numbers = text2int(transcript_1)
-    if numbers is None:
-        return "Error: Text to number conversion returned None"
+###########################################
 
-    code_to_word_map = {v: k for k, v in word_to_code_map.items()}
-    text = transcript_to_sentence(numbers, code_to_word_map)
-    return text
-
-def sel_lng(lng, mic=None, file=None):
-    if mic is not None:
-        audio = mic
-    elif file is not None:
-        audio = file
-    else:
-        return "You must either provide a mic recording or a file"
-    
-    if lng == "Odiya":
-        return transcribe_odiya(audio)
-    elif lng == "Odiya-trans":
-        return transcribe_odiya_eng(audio)
-    elif lng == "Hindi-trans":
-        return transcribe_hin_eng(audio)
-    elif lng == "Hindi":
-        return transcribe_hindi(audio)
-    elif lng == "Kannada-trans":
-       return transcribe_kan_eng(audio)
-    elif lng == "Kannada":
-       return transcribe_kannada(audio)
-    elif lng == "Telugu-trans":
-       return transcribe_tel_eng(audio)
-    elif lng == "Telugu":
-       return transcribe_telugu(audio)
-    elif lng == "Bangala-trans":
-       return transcribe_ban_eng(audio)
-    elif lng == "Bangala":
-       return transcribe_bangala(audio)
-    elif lng == "Assamese-LM":
-       return transcribe_assamese_LM(audio)
-    elif lng == "Assamese-Model2":
-       return transcribe_assamese_model2(audio)
-    elif lng == "Odia_model2":
-       return transcribe_odiya_model2(audio)
-    elif lng == "Odia_trans_model2":
-       return transcribe_odiya_eng_model2(audio)
-        
-    
 # Function to replace incorrectly spelled words
 def replace_words(sentence):
     replacements = [
@@ -425,6 +246,191 @@ code_to_word_map = {}
 # Create reverse mapping
 code_to_word_map = {v: k for k, v in word_to_code_map.items()}
 
+###########################################
+
+def transcribe_odiya(speech):
+    text = p1(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+
+###################################
+def transcribe_odiya_model2(speech):
+    text = odia_model2(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+
+def transcribe_odiya_eng_model2(speech):
+    trn = Transliterator(source='ori', target='eng', build_lookup=True)
+    text = odia_model2(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+
+########################################
+def cleanhtml(raw_html):
+    cleantext = re.sub(r'<.*?>', '', raw_html)
+    return cleantext
+
+def transcribe_hindi(speech):
+    text = p2(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+def transcribe_kannada(speech):
+    text = p3(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+def transcribe_telugu(speech):
+    text = p4(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+    
+def transcribe_bangala(speech):
+    text = p5(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+
+def transcribe_assamese_LM(speech):
+    text = p8(speech)["text"]
+    text = cleanhtml(text)
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+    
+def transcribe_assamese_model2(speech):
+    text = p7(speech)["text"]
+    text = cleanhtml(text)
+    if text is None:
+        return "Error: ASR returned None"
+    return text
+    
+def transcribe_odiya_eng(speech):
+    trn = Transliterator(source='ori', target='eng', build_lookup=True)
+    text = p1(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+    
+def transcribe_ban_eng(speech):
+    trn = Transliterator(source='ben', target='eng', build_lookup=True)
+    text = p5(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+
+def transcribe_hin_eng(speech):
+    trn = Transliterator(source='hin', target='eng', build_lookup=True)
+    text = p2(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+
+def transcribe_kan_eng(speech):
+    trn = Transliterator(source='kan', target='eng', build_lookup=True)
+    text = p3(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+    
+def transcribe_tel_eng(speech):
+    trn = Transliterator(source='tel', target='eng', build_lookup=True)
+    text = p4(speech)["text"]
+    if text is None:
+        return "Error: ASR returned None"
+    sentence = trn.transform(text)
+    if sentence is None:
+        return "Error: Transliteration returned None"
+    replaced_words = replace_words(sentence)
+    processed_sentence = process_doubles(replaced_words)
+    return process_transcription(processed_sentence)
+
+
+def process_transcription(input_sentence):
+    word_to_code_map = {}
+    code_to_word_map = {}
+
+    transcript_1 = sentence_to_transcript(input_sentence, word_to_code_map)
+    if transcript_1 is None:
+        return "Error: Transcript conversion returned None"
+
+    numbers = text2int(transcript_1)
+    if numbers is None:
+        return "Error: Text to number conversion returned None"
+
+    code_to_word_map = {v: k for k, v in word_to_code_map.items()}
+    text = transcript_to_sentence(numbers, code_to_word_map)
+    return text
+
+def sel_lng(lng, mic=None, file=None):
+    if mic is not None:
+        audio = mic
+    elif file is not None:
+        audio = file
+    else:
+        return "You must either provide a mic recording or a file"
+    
+    if lng == "Odiya":
+        return transcribe_odiya(audio)
+    elif lng == "Odiya-trans":
+        return transcribe_odiya_eng(audio)
+    elif lng == "Hindi-trans":
+        return transcribe_hin_eng(audio)
+    elif lng == "Hindi":
+        return transcribe_hindi(audio)
+    elif lng == "Kannada-trans":
+       return transcribe_kan_eng(audio)
+    elif lng == "Kannada":
+       return transcribe_kannada(audio)
+    elif lng == "Telugu-trans":
+       return transcribe_tel_eng(audio)
+    elif lng == "Telugu":
+       return transcribe_telugu(audio)
+    elif lng == "Bangala-trans":
+       return transcribe_ban_eng(audio)
+    elif lng == "Bangala":
+       return transcribe_bangala(audio)
+    elif lng == "Assamese-LM":
+       return transcribe_assamese_LM(audio)
+    elif lng == "Assamese-Model2":
+       return transcribe_assamese_model2(audio)
+    elif lng == "Odia_model2":
+       return transcribe_odiya_model2(audio)
+    elif lng == "Odia_trans_model2":
+       return transcribe_odiya_eng_model2(audio)
+        
+    
+
+
 # Convert transcript back to sentence
 # reconstructed_sentence_1 = transcript_to_sentence(numbers, code_to_word_map)