Update app.py

app.py

@@ -1,97 +1,62 @@
+
 # Import required libraries
 import os
 import pandas as pd
 import streamlit as st
-# from transformers import DistilBertTokenizer, DistilBertForSequenceClassification
-from transformers import DistilBertTokenizerFast, DistilBertModel
-# from transformers import DistilBertTokenizerFast, DistilBertForSequenceClassification
-
 from transformers import pipeline
 from sentence_transformers import SentenceTransformer, util
 import requests
 import json
+from pyngrok import ngrok
 
-# Configure Hugging Face API token securely
+# Set up Hugging Face API token
 api_key = os.getenv("HF_API_KEY")
 
 # Load the CSV dataset
-try:
-    data = pd.read_csv('genetic-Final.csv')  # Ensure the dataset filename is correct
-except FileNotFoundError:
-    st.error("Dataset file not found. Please upload it to this directory.")
-
-# Load DistilBERT Tokenizer and Model
-# tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased')
-# model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased')
-
-# Load DistilBERT tokenizer and model (without classification layer)
-tokenizer = DistilBertTokenizerFast.from_pretrained("distilbert-base-uncased")
-model = DistilBertModel.from_pretrained("distilbert-base-uncased")
-
-query = "What is fructose-1,6-bisphosphatase deficiency?"
-
-# Tokenize input
-inputs = tokenizer(query, return_tensors="pt", padding=True, truncation=True)
-
-# Get model output (embeddings)
-with torch.no_grad():
-    outputs = model(**inputs)
-
-# Extract embeddings (last hidden state)
-embeddings = outputs.last_hidden_state.mean(dim=1)  # Averaging over token embeddings
-
-# Use the embeddings for further processing or retrieval
-print(embeddings)
-
-
-# Preprocessing the dataset (if needed)
-if 'combined_description' not in data.columns:
-    data['combined_description'] = (
-        data['Symptoms'].fillna('') + " " +
-        data['Severity Level'].fillna('') + " " +
-        data['Risk Assessment'].fillna('') + " " +
-        data['Treatment Options'].fillna('') + " " +
-        data['Suggested Medical Tests'].fillna('') + " " +
-        data['Minimum Values for Medical Tests'].fillna('') + " " +
-        data['Emergency Treatment'].fillna('')
-    )
-
-# Initialize Sentence Transformer model for RAG-based retrieval
+data = pd.read_csv('genetic-Final.csv')
+
+# Drop unnecessary columns (Unnamed columns)
+data = data.drop(columns=['Unnamed: 0', 'Unnamed: 11', 'Unnamed: 12', 'Unnamed: 13'])
+
+# Combine relevant columns into one combined description field
+data['combined_description'] = (
+    data['Symptoms'].fillna('') + " " +
+    data['Severity Level'].fillna('') + " " +
+    data['Risk Assessment'].fillna('') + " " +
+    data['Treatment Options'].fillna('') + " " +
+    data['Suggested Medical Tests'].fillna('') + " " +
+    data['Minimum Values for Medical Tests'].fillna('') + " " +
+    data['Emergency Treatment'].fillna('')
+)
+
+# Initialize the Sentence Transformer model for embeddings
 retriever_model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
 
-# Define a function to get embeddings using DistilBERT
+# Function to safely generate embeddings for each row
 def generate_embedding(description):
-    if description:
-        inputs = tokenizer(description, return_tensors='pt', truncation=True, padding=True, max_length=512)
-        outputs = model(**inputs)
-        return outputs.logits.detach().numpy().flatten()
+    if description:  # Check if the description is not empty or NaN
+        return retriever_model.encode(description).tolist()  # Convert the numpy array to list
     else:
         return []
 
 # Generate embeddings for the combined description
-if 'embeddings' not in data.columns:
-    data['embeddings'] = data['combined_description'].apply(generate_embedding)
+data['embeddings'] = data['combined_description'].apply(generate_embedding)
 
-# Function to retrieve relevant information based on user query
+# Function to retrieve relevant information from CSV dataset based on user query
 def get_relevant_info(query, top_k=3):
     query_embedding = retriever_model.encode(query)
     similarities = [util.cos_sim(query_embedding, doc_emb)[0][0].item() for doc_emb in data['embeddings']]
     top_indices = sorted(range(len(similarities)), key=lambda i: similarities[i], reverse=True)[:top_k]
     return data.iloc[top_indices]
 
-# Function to generate response using DistilBERT (integrating with the model)
-def generate_response(input_text, relevant_info):
-    # Concatenate the relevant information as context for the model
-    context = "\n".join(relevant_info['combined_description'].tolist())
-    input_with_context = f"Context: {context}\n\nUser Query: {input_text}"
+# Function to generate response using Hugging Face Model API
+def generate_response(input_text):
+    api_url = "https://api-inference.huggingface.co/models/m42-health/Llama3-Med42-8B"
+    headers = {"Authorization": f"Bearer {os.environ['HUGGINGFACEHUB_API_TOKEN']}"}
+    payload = {"inputs": input_text}
 
-    # Simple logic for generating a response using DistilBERT-based model
-    inputs = tokenizer(input_with_context, return_tensors='pt', truncation=True, padding=True, max_length=512)
-    outputs = model(**inputs)
-    logits = outputs.logits.detach().numpy().flatten()
-    response = tokenizer.decode(logits.argmax(), skip_special_tokens=True)
-    
-    return response
+    response = requests.post(api_url, headers=headers, json=payload)
+    return json.loads(response.content.decode("utf-8"))[0]["generated_text"]
 
 # Streamlit UI for the Chatbot
 def main():
@@ -112,19 +77,20 @@ def main():
         relevant_info = get_relevant_info(user_query)
         st.write("#### Relevant Medical Information:")
         for i, row in relevant_info.iterrows():
-            st.write(f"- {row['combined_description']}")  # Adjust to show meaningful info
+            st.write(f"- {row['combined_description']}")
 
-        # Generate a response from DistilBERT model
-        response = generate_response(user_query, relevant_info)
+        # Generate a response from the Llama3-Med42-8B model
+        response = generate_response(user_query)
         st.write("#### Model's Response:")
         st.write(response)
 
     # Process the uploaded file (if any)
     if uploaded_file:
-        # Display analysis of the uploaded report file (process based on file type)
+        # Display analysis of the uploaded report file
         st.write("### Uploaded Report Analysis:")
         report_text = "Extracted report content here"  # Placeholder for file processing logic
         st.write(report_text)
 
+# Start Streamlit app in Colab using ngrok
 if __name__ == "__main__":
     main()