Update app.py

app.py

@@ -1,101 +1,91 @@
-import os
 import streamlit as st
-import pandas as pd
-from groq import Groq
-import sounddevice as sd
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+import torch
 import numpy as np
-import wavio
-
-# Set your API key (replace with secure method for production)
-API_KEY = "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"
-
-# Initialize Groq client
-client = Groq(api_key=API_KEY)
-
-# Streamlit app layout
-st.set_page_config(page_title="PhoneX: Personalized Reading App", layout="wide")
-st.title("Welcome to PhoneX: Your Personalized Reading App")
-st.sidebar.title("Navigation")
-
-# User Registration
-user_id = st.sidebar.text_input("Enter your User ID:")
-if st.sidebar.button("Register"):
-    if user_id:
-        st.sidebar.success(f"User {user_id} registered successfully!")
-    else:
-        st.sidebar.error("Please enter a valid User ID.")
-
-# Diagnostic Test
-st.header("Diagnostic Test")
-assessment_text = st.text_area("Read the following text aloud:", "Sample text for assessment: The quick brown fox jumps over the lazy dog.")
-if st.button("Submit Assessment"):
-    if assessment_text:
-        # Call the LLAMA model for feedback based on assessment
-        chat_completion = client.chat.completions.create(
-            messages=[
-                {
-                    "role": "user",
-                    "content": f"Evaluate the following assessment: {assessment_text}",
-                }
-            ],
-            model="llama3-8b-8192",
-        )
-        st.success("Assessment submitted successfully!")
-        st.write("LLAMA Model Output:", chat_completion.choices[0].message.content)
-    else:
-        st.error("Please enter the text to assess.")
-
-# Personalized Reading Material
-st.header("Generate Personalized Reading Material")
-if st.button("Generate Reading Material"):
-    generated_text = "Generated personalized text focusing on phonetic patterns: 'sh', 'ch', and 'th'."  # Placeholder for personalized content
-    st.text_area("Your Personalized Reading Material:", generated_text, height=200)
-
-# Audio Recording Feature
-st.header("Record Your Voice")
-duration = st.number_input("Duration of recording (seconds)", min_value=1, max_value=60, value=5)
-if st.button("Record Audio"):
-    st.write("Recording...")
-    recording = sd.rec(int(duration * 44100), samplerate=44100, channels=2)
-    sd.wait()  # Wait until the recording is finished
-    wavio.write("recording.wav", recording, 44100, sampwidth=2)  # Save as WAV file
-    st.success("Recording completed!")
-
-# Upload Recorded Audio for Transcription
-if st.file_uploader("Upload your audio file (m4a):", type=["m4a"]) is not None:
-    if st.button("Transcribe Audio"):
-        filename = "recording.wav"  # Use the recorded file for transcription
-        with open(filename, "rb") as file:
-            transcription = client.audio.transcriptions.create(
-                file=(filename, file.read()),
-                model="whisper-large-v3",
-                response_format="verbose_json",
-            )
-            st.write("Transcription:", transcription.text)
-
-# Feedback Collection
-st.header("Feedback Collection")
-feedback = st.text_area("Your Feedback:")
-if st.button("Submit Feedback"):
-    if feedback:
-        # Placeholder for storing feedback (e.g., to Google Sheets)
-        st.success("Feedback submitted successfully!")
-    else:
-        st.error("Please provide your feedback.")
-
-# Visualization (User Progress)
-st.header("User Progress Visualization")
-# Placeholder for progress data visualization
-progress_data = pd.DataFrame({'Task': ['Reading', 'Feedback'], 'Scores': [85, 90]})
-st.bar_chart(progress_data.set_index('Task'))
-
-# Error Handling Example
-try:
-    # Your main logic here
-    pass
-except Exception as e:
-    st.error(f"An error occurred: {e}")
-
-# Footer
-st.sidebar.markdown("---")
-st.sidebar.markdown("Created by Your Name")
+import tempfile
+import wave
+
+# Load Wav2Vec2 model and processor
+processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
+
+# Streamlit App
+st.title("Phonics/Personalized Reading App")
+st.write("Record your audio and we will transcribe it.")
+
+# Audio recording using HTML5
+record_button = st.button("Record Audio")
+
+if record_button:
+    st.markdown("""
+        <audio id="audio" controls></audio>
+        <button id="start" onclick="startRecording()">Start Recording</button>
+        <button id="stop" onclick="stopRecording()" disabled>Stop Recording</button>
+        <script>
+            let mediaRecorder;
+            let audioChunks = [];
+
+            async function startRecording() {
+                const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
+                mediaRecorder = new MediaRecorder(stream);
+                mediaRecorder.ondataavailable = event => {
+                    audioChunks.push(event.data);
+                };
+
+                mediaRecorder.onstop = () => {
+                    const audioBlob = new Blob(audioChunks, { type: 'audio/wav' });
+                    const audioUrl = URL.createObjectURL(audioBlob);
+                    const audioElement = document.getElementById('audio');
+                    audioElement.src = audioUrl;
+
+                    // Prepare to send audio to server
+                    const formData = new FormData();
+                    formData.append('audio', audioBlob, 'recording.wav');
+
+                    fetch('/upload', {
+                        method: 'POST',
+                        body: formData
+                    }).then(response => response.json()).then(data => {
+                        st.session_state.transcription = data.transcription;
+                        st.experimental_rerun();  // Refresh the app to show the transcription
+                    });
+                };
+
+                mediaRecorder.start();
+                document.getElementById('start').disabled = true;
+                document.getElementById('stop').disabled = false;
+            }
+
+            function stopRecording() {
+                mediaRecorder.stop();
+                document.getElementById('start').disabled = false;
+                document.getElementById('stop').disabled = true;
+            }
+        </script>
+    """, unsafe_allow_html=True)
+
+# Display the transcription
+if 'transcription' in st.session_state:
+    st.write("Transcription:")
+    st.write(st.session_state.transcription)
+
+# Handle audio file upload
+uploaded_file = st.file_uploader("Or upload your audio file", type=["wav", "mp3"])
+
+if uploaded_file is not None:
+    # Save uploaded audio file to a temporary file
+    with tempfile.NamedTemporaryFile(delete=True) as temp_file:
+        temp_file.write(uploaded_file.read())
+        temp_file.flush()
+        
+        # Process the audio file for transcription
+        audio_input = processor(temp_file.name, sampling_rate=16000, return_tensors="pt", padding=True)
+
+        with torch.no_grad():
+            logits = model(audio_input.input_values).logits
+
+        predicted_ids = torch.argmax(logits, dim=-1)
+        transcription = processor.batch_decode(predicted_ids)
+
+        st.session_state.transcription = transcription[0]  # Store transcription
+        st.experimental_rerun()  # Refresh the app to show the transcription