Update app.py

app.py

@@ -1,56 +1,54 @@
-# app.py
 import gradio as gr
 from transformers import pipeline
 import torch
+from TTS.api import TTS
 import tempfile
 import os
-from TTS.api import TTS
-import whisper
+import speech_recognition as sr
 
-# Load question-generation pipeline (use a lightweight model)
+# Load question generation model
 qg_pipeline = pipeline("text2text-generation", model="valhalla/t5-small-e2e-qg")
 
 # Load TTS model
-tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False)
-
-# Load Whisper STT model
-whisper_model = whisper.load_model("base")
+tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False, gpu=False)
 
-# Generate question and audio from input text
 def generate_question(text):
-    output = qg_pipeline("generate question: " + text, max_length=64, clean_up_tokenization_spaces=True)[0]['generated_text']
-    
-    # Save TTS audio to temp file
+    # Prompt for question generation
+    input_text = f"generate question: {text.strip()}"
+    question = qg_pipeline(input_text)[0]["generated_text"]
+
+    # Save spoken question as audio
     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as fp:
-        tts.tts_to_file(text=output, file_path=fp.name)
+        tts.tts_to_file(text=question, file_path=fp.name)
         audio_path = fp.name
 
-    return output, audio_path
+    return question, audio_path
+
+def transcribe_audio(audio_path):
+    recognizer = sr.Recognizer()
+    with sr.AudioFile(audio_path) as source:
+        audio_data = recognizer.record(source)
+        try:
+            return recognizer.recognize_google(audio_data)
+        except sr.UnknownValueError:
+            return "Sorry, I could not understand your answer."
+        except sr.RequestError:
+            return "Sorry, there was an error with the speech recognition service."
 
-# Transcribe user audio answer
-def transcribe_audio(audio):
-    audio = whisper.load_audio(audio)
-    audio = whisper.pad_or_trim(audio)
-    mel = whisper.log_mel_spectrogram(audio).to(whisper_model.device)
-    options = whisper.DecodingOptions()
-    result = whisper.decode(whisper_model, mel, options)
-    return result.text
+with gr.Blocks() as app:
+    gr.Markdown("### 📘 Enter your coursebook text below:")
 
-# Gradio interface
-with gr.Blocks() as demo:
-    gr.Markdown("### Voice Q&A Generator")
-    with gr.Row():
-        input_text = gr.Textbox(label="Coursebook Text")
-        generate_btn = gr.Button("Generate Question")
+    course_text = gr.Textbox(lines=6, label="Coursebook Text")
 
-    question_out = gr.Textbox(label="Generated Question")
-    audio_out = gr.Audio(label="AI Question (Audio)", type="filepath")
+    generate_btn = gr.Button("🎤 Generate Question and Speak")
+    question_output = gr.Textbox(label="Generated Question")
+    audio_output = gr.Audio(label="Question Audio", type="filepath")
 
-    with gr.Row():
-        user_audio = gr.Audio(source="microphone", type="filepath", label="Your Answer")
-        transcribed_text = gr.Textbox(label="Transcribed Answer")
+    user_audio = gr.Audio(label="Your Answer", type="filepath", sources=["microphone"])
+    transcribe_btn = gr.Button("📝 Transcribe Answer")
+    transcription_output = gr.Textbox(label="Transcribed Answer")
 
-    generate_btn.click(fn=generate_question, inputs=input_text, outputs=[question_out, audio_out])
-    user_audio.change(fn=transcribe_audio, inputs=user_audio, outputs=transcribed_text)
+    generate_btn.click(fn=generate_question, inputs=course_text, outputs=[question_output, audio_output])
+    transcribe_btn.click(fn=transcribe_audio, inputs=user_audio, outputs=transcription_output)
 
-demo.launch()
+app.launch()