Update app.py

app.py

@@ -1,95 +1,95 @@
 import gradio as gr
 from transformers import GPT2LMHeadModel, GPT2Tokenizer, pipeline
-
-# Initialize the GPT2 model and tokenizer
-tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
-model = GPT2LMHeadModel.from_pretrained("gpt2")
-
-# Initialize the Whisper GPT model
-translation_pipeline =  pipeline("automatic-speech-recognition", model="openai/whisper-large-v2")
-
-# Geriatric Depression Scale Quiz Questions
-questions = [
-"Are you basically satisfied with your life?",
-"Have you dropped many of your activities and interests?",
-"Do you feel that your life is empty?",
-"Do you often get bored?",
-"Are you in good spirits most of the time?",
-"Are you afraid that something bad is going to happen to you?",
-"Do you feel happy most of the time?",
-"Do you often feel helpless?",
-"Do you prefer to stay at home, rather than going out and doing things?",
-"Do you feel that you have more problems with memory than most?",
-"Do you think it is wonderful to be alive now?",
-"Do you feel worthless the way you are now?",
-"Do you feel full of energy?",
-"Do you feel that your situation is hopeless?",
-"Do you think that most people are better off than you are?"   
-]
-
-def ask_questions(answers):
-"""Calculate score based on answers."""
-score = 0
-for answer in answers:
-    if answer.lower() == 'yes':
-        score += 1
-    elif answer.lower() != 'no':
-        raise ValueError(f"Invalid answer: {answer}")
-return score
-
-def understand_answers(audio_answers):
-"""Convert audio answers to text using the Whisper ASR model."""
-asr_pipeline = pipeline("automatic-speech-recognition", model="openai/whisper-large-v2")
-text_answers = []
-for audio in audio_answers:
-    transcript = asr_pipeline(audio)
-    text_answers.append(transcript[0]['generated_text'])
-return text_answers
-
-def modified_summarize(answers):
-"""Summarize answers using the GPT2 model."""
-answers_str = " ".join(answers)
-inputs = tokenizer.encode("summarize: " + answers_str, return_tensors='pt')
-summary_ids = model.generate(inputs, max_length=150, num_beams=5, early_stopping=True)
-return tokenizer.decode(summary_ids[0], skip_special_tokens=True)
-
-
-def assistant(*audio_answers):
-"""Convert audio answers to text, evaluate and provide a summary."""
-text_answers = understand_answers(audio_answers)
-summarized_text = modified_summarize(text_answers)
-score = ask_questions(text_answers)
-return summarized_text, f"Your score is: {score}/{len(questions)}", text_answers  # Return text_answers as well
-
-# Create the Gradio Blocks interface with button click
-
-def update():
-audio_answers = [audio.value for audio in inp]  # Using inp as it collects all the audio inputs
-# Handling the three returned values from the assistant function
-summarized_text, score_string, text_answers = assistant(*audio_answers)
-out_last_transcription.value = summarized_text  # Displaying the summarized text
-out_score.value = score_string  # Displaying the score
-
-with gr.Blocks() as demo:
-gr.Markdown("Start recording your responses below and then click **Run** to see the transcription and your score.")
-
-# Clearly initializing Inputs and Outputs lists for the button click
-inp = []
-
-# Using Column to nest questions
-with gr.Column(scale=1, min_width=600):
-    for i, question in enumerate(questions):
-        gr.Markdown(f"**Question {i+1}:** {question}")
-        audio_input = gr.Audio(source="microphone")
-        inp.append(audio_input)
-
-# Two output textboxes: one for the last transcribed answer and another for the score
-out_last_transcription = gr.Textbox(label="Last Transcribed Answer", placeholder="Last transcribed answer will appear here.")
-out_score = gr.Textbox(label="Score", placeholder="Your score will appear here.")
-
-# Button with click event
-btn = gr.Button("Run")
-btn.click(fn=update, inputs=inp, outputs=[out_last_transcription, out_score])
-
-
-demo.launch()
+    
+    # Initialize the GPT2 model and tokenizer
+    tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
+    model = GPT2LMHeadModel.from_pretrained("gpt2")
+    
+    # Initialize the Whisper GPT model
+    translation_pipeline =  pipeline("automatic-speech-recognition", model="openai/whisper-large-v2")
+    
+    # Geriatric Depression Scale Quiz Questions
+    questions = [
+    "Are you basically satisfied with your life?",
+    "Have you dropped many of your activities and interests?",
+    "Do you feel that your life is empty?",
+    "Do you often get bored?",
+    "Are you in good spirits most of the time?",
+    "Are you afraid that something bad is going to happen to you?",
+    "Do you feel happy most of the time?",
+    "Do you often feel helpless?",
+    "Do you prefer to stay at home, rather than going out and doing things?",
+    "Do you feel that you have more problems with memory than most?",
+    "Do you think it is wonderful to be alive now?",
+    "Do you feel worthless the way you are now?",
+    "Do you feel full of energy?",
+    "Do you feel that your situation is hopeless?",
+    "Do you think that most people are better off than you are?"   
+    ]
+    
+    def ask_questions(answers):
+        """Calculate score based on answers."""
+        score = 0
+        for answer in answers:
+            if answer.lower() == 'yes':
+                score += 1
+            elif answer.lower() != 'no':
+                raise ValueError(f"Invalid answer: {answer}")
+        return score
+    
+    def understand_answers(audio_answers):
+        """Convert audio answers to text using the Whisper ASR model."""
+        asr_pipeline = pipeline("automatic-speech-recognition", model="openai/whisper-large-v2")
+        text_answers = []
+        for audio in audio_answers:
+            transcript = asr_pipeline(audio)
+            text_answers.append(transcript[0]['generated_text'])
+        return text_answers
+    
+    def modified_summarize(answers):
+        """Summarize answers using the GPT2 model."""
+        answers_str = " ".join(answers)
+        inputs = tokenizer.encode("summarize: " + answers_str, return_tensors='pt')
+        summary_ids = model.generate(inputs, max_length=150, num_beams=5, early_stopping=True)
+        return tokenizer.decode(summary_ids[0], skip_special_tokens=True)
+    
+    
+    def assistant(*audio_answers):
+        """Convert audio answers to text, evaluate and provide a summary."""
+        text_answers = understand_answers(audio_answers)
+        summarized_text = modified_summarize(text_answers)
+        score = ask_questions(text_answers)
+        return summarized_text, f"Your score is: {score}/{len(questions)}", text_answers  # Return text_answers as well
+    
+    # Create the Gradio Blocks interface with button click
+    
+    def update():
+        audio_answers = [audio.value for audio in inp]  # Using inp as it collects all the audio inputs
+        # Handling the three returned values from the assistant function
+        summarized_text, score_string, text_answers = assistant(*audio_answers)
+        out_last_transcription.value = summarized_text  # Displaying the summarized text
+        out_score.value = score_string  # Displaying the score
+    
+    with gr.Blocks() as demo:
+        gr.Markdown("Start recording your responses below and then click **Run** to see the transcription and your score.")
+        
+        # Clearly initializing Inputs and Outputs lists for the button click
+        inp = []
+        
+        # Using Column to nest questions
+        with gr.Column(scale=1, min_width=600):
+            for i, question in enumerate(questions):
+                gr.Markdown(f"**Question {i+1}:** {question}")
+                audio_input = gr.Audio(source="microphone")
+                inp.append(audio_input)
+        
+        # Two output textboxes: one for the last transcribed answer and another for the score
+        out_last_transcription = gr.Textbox(label="Last Transcribed Answer", placeholder="Last transcribed answer will appear here.")
+        out_score = gr.Textbox(label="Score", placeholder="Your score will appear here.")
+        
+        # Button with click event
+        btn = gr.Button("Run")
+        btn.click(fn=update, inputs=inp, outputs=[out_last_transcription, out_score])
+    
+    
+    demo.launch()