Update app.py (#1)

- Update app.py (122a817fc3e71b709355262c7a64c37518c15c05)

app.py

@@ -4,6 +4,7 @@ from langchain_community.vectorstores import Chroma
 from langchain_community.embeddings import HuggingFaceBgeEmbeddings
 from groq import Groq
 from dotenv import load_dotenv
+import requests
 
 # Initialize Streamlit page configuration
 st.set_page_config(page_title="Medical Knowledge Assistant", layout="wide")
@@ -29,125 +30,117 @@ if not api_key:
 # Initialize the app
 st.title("Medical Knowledge Assistant")
 
-try:
-    # Set up the embeddings
-    model_name = "BAAI/bge-large-en"
-    model_kwargs = {'device': 'cpu'}
-    encode_kwargs = {'normalize_embeddings': False}
-    embeddings = HuggingFaceBgeEmbeddings(
-        model_name=model_name,
-        model_kwargs=model_kwargs,
-        encode_kwargs=encode_kwargs
-    )
+# Google Drive file ID (use your own file ID)
+file_id = '1lVlF8dYsNFPzrNGqn7jiJos7qX49jmi0'  # Replace with your Google Drive file ID
+destination_path = '/tmp/Embedded_Med_books'  # Temporary location to store the vector store
 
-    # Load the vector store from the local drive
-    script_dir = os.path.dirname(os.path.abspath(__file__))
-    persist_directory = os.path.join(script_dir, 'Embedded_Med_books')
+# Function to download file from Google Drive
+def download_from_drive(file_id, destination_path):
+    """Download the vector store file from Google Drive."""
+    url = f'https://drive.google.com/uc?export=download&id={file_id}'
+    response = requests.get(url)
+    if response.status_code == 200:
+        with open(destination_path, 'wb') as f:
+            f.write(response.content)
+        return destination_path
+    else:
+        st.error("Failed to download the file from Google Drive.")
+        return None
 
-    
-    # Debug information
-    st.sidebar.header("Debug Information")
-    st.sidebar.write("Vector store path:", persist_directory)
-    
-    with st.sidebar:
-        st.write("API Key Loaded:", "Yes" if api_key else "No")
+# Check if the vector store file exists, and download it if necessary
+if not os.path.exists(destination_path):
+    st.warning("Downloading the vector store from Google Drive...")
+    download_from_drive(file_id, destination_path)
+    st.success("Vector store downloaded successfully!")
 
-    # Check vector store directory
-    if not os.path.exists(persist_directory):
-        st.error(f"Vector store directory not found at: {persist_directory}")
-        if st.button("Create Directory"):
-            os.makedirs(persist_directory)
-            st.success("Directory created!")
-    
-    try:
-        vector_store = Chroma(
-            persist_directory=persist_directory,
-            embedding_function=embeddings
-        )
-    except Exception as e:
-        st.error(f"Error loading vector store: {e}")
+# Set up embeddings
+model_name = "BAAI/bge-large-en"
+model_kwargs = {'device': 'cpu'}
+encode_kwargs = {'normalize_embeddings': False}
+embeddings = HuggingFaceBgeEmbeddings(
+    model_name=model_name,
+    model_kwargs=model_kwargs,
+    encode_kwargs=encode_kwargs
+)
 
-    # vector_store = Chroma(
-    #     persist_directory=persist_directory,
-    #     embedding_function=embeddings
-    # )
-    
-    retriever = vector_store.as_retriever(search_kwargs={'k': 1})
-    
-    # Initialize Groq client
-    client = Groq(api_key=api_key)
+# Load the vector store from the downloaded file
+vector_store = Chroma(
+    persist_directory=destination_path,
+    embedding_function=embeddings
+)
+retriever = vector_store.as_retriever(search_kwargs={'k': 1})
 
-    # Streamlit input
-    query = st.text_input("Enter your medical question here:")
+# Initialize Groq client
+client = Groq(api_key=api_key)
 
-    def query_with_groq(query, retriever):
-        try:
-            # Retrieve relevant documents
-            docs = retriever.get_relevant_documents(query)
-            context = "\n".join([doc.page_content for doc in docs])
+# Streamlit input
+query = st.text_input("Enter your medical question here:")
 
-            # Call the Groq API with the query and context
-            completion = client.chat.completions.create(
-                model="llama3-70b-8192",
-                messages=[
-                    {
-                        "role": "system",
-                        "content": (
-                            "You are a knowledgeable medical assistant. For any medical term or disease, include comprehensive information covering: "
-                            "definitions, types, historical background, major theories, known causes, and contributing risk factors. "
-                            "Explain the genesis or theories on its origin, if applicable. Use a structured, thorough approach and keep language accessible. "
-                            "provide symptoms, diagnosis, and treatment and post operative care , address all with indepth explanation , with specific details and step-by-step processes where relevant. "
-                            "If the context does not adequately cover the user's question, respond with: 'I cannot provide an answer based on the available medical dataset.'"
-                        )
-                    },
-                    {
-                        "role": "system",
-                        "content": (
-                            "If the user asks for a medical explanation, ensure accuracy, don't include layman's terms if complex terms are used, "
-                            "and organize responses in a structured way."
-                        )
-                    },
-                    {
-                        "role": "system",
-                        "content": (
-                            "When comparing two terms or conditions, provide a clear, concise, and structured comparison. Highlight key differences in their "
-                            "definitions, symptoms, causes, diagnoses, and treatments with indepth explanation of each. If relevant, include any overlapping characteristics."
-                        )
-                    },
-                    {
-                        "role": "user",
-                        "content": f"{context}\n\nQ: {query}\nA:"
-                    }
-                ],
-                temperature=0.7,
-                max_tokens=3000,
-                stream=True
-            )
+def query_with_groq(query, retriever):
+    try:
+        # Retrieve relevant documents
+        docs = retriever.get_relevant_documents(query)
+        context = "\n".join([doc.page_content for doc in docs])
 
-            # Create a placeholder for streaming response
-            response_container = st.empty()
-            response = ""
-            
-            # Stream the response
-            for chunk in completion:
-                if chunk.choices[0].delta.content:
-                    response += chunk.choices[0].delta.content
-                    response_container.markdown(response)
-            
-            return response
-            
-        except Exception as e:
-            st.error(f"Error during query processing: {str(e)}")
-            return None
+        # Call the Groq API with the query and context
+        completion = client.chat.completions.create(
+            model="llama3-70b-8192",
+            messages=[
+                {
+                    "role": "system",
+                    "content": (
+                        "You are a knowledgeable medical assistant. For any medical term or disease, include comprehensive information covering: "
+                        "definitions, types, historical background, major theories, known causes, and contributing risk factors. "
+                        "Explain the genesis or theories on its origin, if applicable. Use a structured, thorough approach and keep language accessible. "
+                        "provide symptoms, diagnosis, and treatment and post operative care , address all with indepth explanation , with specific details and step-by-step processes where relevant. "
+                        "If the context does not adequately cover the user's question, respond with: 'I cannot provide an answer based on the available medical dataset.'"
+                    )
+                },
+                {
+                    "role": "system",
+                    "content": (
+                        "If the user asks for a medical explanation, ensure accuracy, don't include layman's terms if complex terms are used, "
+                        "and organize responses in a structured way."
+                    )
+                },
+                {
+                    "role": "system",
+                    "content": (
+                        "When comparing two terms or conditions, provide a clear, concise, and structured comparison. Highlight key differences in their "
+                        "definitions, symptoms, causes, diagnoses, and treatments with indepth explanation of each. If relevant, include any overlapping characteristics."
+                    )
+                },
+                {
+                    "role": "user",
+                    "content": f"{context}\n\nQ: {query}\nA:"
+                }
+            ],
+            temperature=0.7,
+            max_tokens=3000,
+            stream=True
+        )
 
-    if st.button("Get Answer"):
-        if query:
-            with st.spinner("Processing your query..."):
-                answer = query_with_groq(query, retriever)
-                if answer:
-                    st.success("Query processed successfully!")
-        else:
-            st.warning("Please enter a query.")
+        # Create a placeholder for streaming response
+        response_container = st.empty()
+        response = ""
+        
+        # Stream the response
+        for chunk in completion:
+            if chunk.choices[0].delta.content:
+                response += chunk.choices[0].delta.content
+                response_container.markdown(response)
+        
+        return response
+        
+    except Exception as e:
+        st.error(f"Error during query processing: {str(e)}")
+        return None
 
-except Exception as e:
-    st.error(f"Initialization error: {str(e)}")
+if st.button("Get Answer"):
+    if query:
+        with st.spinner("Processing your query..."):
+            answer = query_with_groq(query, retriever)
+            if answer:
+                st.success("Query processed successfully!")
+    else:
+        st.warning("Please enter a query.")