Delete app.py

app.py

@@ -1,92 +0,0 @@
-import streamlit as st
-import sounddevice as sd
-import soundfile as sf
-import numpy as np
-import whisper
-import requests
-import json
-from datetime import datetime
-import os
-from dotenv import load_dotenv
-
-# Load environment variables
-load_dotenv()
-
-# Initialize Whisper model
-@st.cache_resource
-def load_whisper_model():
-    return whisper.load_model("base")
-
-# Mixtral API call function
-def get_clinical_notes(transcription):
-    API_URL = "https://api-inference.huggingface.co/models/mistralai/Mixtral-8x7B-Instruct-v0.1"
-    headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_TOKEN')}"}
-    
-    messages = [
-        {"role": "system", "content": "You are a medical assistant helping to generate clinical notes from doctor-patient conversations. Format the notes in a clear, professional structure."},
-        {"role": "user", "content": f"Generate clinical notes from this doctor-patient conversation: {transcription}"}
-    ]
-    
-    payload = {
-        "model": "mistralai/Mixtral-8x7B-Instruct-v0.1",
-        "messages": messages,
-        "max_tokens": 500,
-        "stream": False
-    }
-    
-    response = requests.post(API_URL, headers=headers, json=payload)
-    return response.json()['choices'][0]['message']['content']
-
-# Main app
-st.title("Medical Conversation Transcriber")
-st.markdown("### Created by Dr. Fernando Ly")
-st.markdown("This application helps medical professionals automatically generate clinical notes from patient conversations.")
-
-# Recording parameters
-SAMPLE_RATE = 16000
-recording_duration = st.slider("Recording duration (seconds)", 10, 300, 60)
-
-if "audio_data" not in st.session_state:
-    st.session_state.audio_data = None
-if "transcription" not in st.session_state:
-    st.session_state.transcription = None
-if "clinical_notes" not in st.session_state:
-    st.session_state.clinical_notes = None
-
-# Record button
-if st.button("Start Recording"):
-    with st.spinner("Recording..."):
-        audio_data = sd.rec(int(recording_duration * SAMPLE_RATE),
-                          samplerate=SAMPLE_RATE,
-                          channels=1,
-                          dtype=np.float32)
-        sd.wait()
-        st.session_state.audio_data = audio_data
-        
-        # Save audio temporarily
-        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-        filename = f"recording_{timestamp}.wav"
-        sf.write(filename, audio_data, SAMPLE_RATE)
-        
-        # Transcribe audio
-        model = load_whisper_model()
-        st.session_state.transcription = model.transcribe(filename)["text"]
-        
-        # Generate clinical notes
-        st.session_state.clinical_notes = get_clinical_notes(st.session_state.transcription)
-        
-        # Clean up audio file
-        os.remove(filename)
-
-# Display results
-if st.session_state.transcription:
-    st.subheader("Transcription")
-    st.write(st.session_state.transcription)
-    
-if st.session_state.clinical_notes:
-    st.subheader("Clinical Notes")
-    st.write(st.session_state.clinical_notes)
-
-# Footer
-st.markdown("---")
-st.markdown("*Note: This is an AI-assisted tool. Please review and verify all generated notes.*")