app.py

import streamlit as st
import pandas as pd
from langchain_core.messages import AIMessage, HumanMessage

from azure_openai import qt, get_response
from retriver import search_and_reconstruct

def read_file(file):
    """
    Reads the content of a text file and returns it as a string.

    :param approver: The type of approver.
    :return: The content of the file as a string.
    """
    fp = f"assets/{file}.md"
    try:
        with open(fp, 'r', encoding='utf-8') as file:
            content = file.read()
        return content
    except FileNotFoundError:
        print(f"The file at {fp} was not found.")
    except IOError:
        print(f"An error occurred while reading the file at {fp}.")


QTESystemMessage = read_file("QTESystemMessage")
RAGSystemMessage = read_file("RAGSystemMessage")
RAGUserMessage = read_file("RAGUserMessage")
k = 5
pagesReturned = 3
temp1 = 0.5
tokens1 = 200
temp2 = 0.6
tokens2 = 2000
asset = "GSKGlossary"

# app config
st.set_page_config(page_title="Medical Sales Toolbox", page_icon="🤖")
st.title("Medical Sales Toolbox :toolbox:")

# session state
if "chat_history" not in st.session_state:
    st.session_state.chat_history = [
        AIMessage(content="Hello, I am the Medical Sales Assistant. How can I help you?"),
    ]

# conversation
for message in st.session_state.chat_history:
    if isinstance(message, AIMessage):
        with st.chat_message("AI"):
            st.write(message.content)
    elif isinstance(message, HumanMessage):
        with st.chat_message("Human"):
            st.write(message.content)

# user input
user_query = st.chat_input("Type your message here...")
if user_query is not None and user_query != "":
    st.session_state.chat_history.append(HumanMessage(content=user_query))

    with st.chat_message("Human"):
        st.markdown(user_query)

   
    qte = qt(QTESystemMessage, st.session_state.chat_history, temp1, tokens1, asset)
    st.text("Contextualised Query")
    st.caption(qte)
    knowledge = search_and_reconstruct(qte, k, pagesReturned)
    
    if knowledge:
        
        # Prepare the data for the table
        table_data = {
            "Title": [entry['Title'] for entry in knowledge],
            "Score (%)": [f"{int(entry.get('Score', 0) * 100)}%" for entry in knowledge],  # Convert to percentage and remove decimals
            "Page": [entry['PageNumber'] for entry in knowledge],
            # "Grounding Text": [entry['ReconstructedText'] for entry in knowledge]
        }

        # Create a dataframe for displaying as a table
    
        df = pd.DataFrame(table_data)
         # Calculate the mean score
        mean_score = sum(entry.get('Score', 0) for entry in knowledge) / len(knowledge)

        # Display the mean score as a Streamlit text element
        
        # Display the table in the sidebar
        st.text("Knowledge Base Results")
        st.text(f"Average Accuracy Score: {mean_score * 100:.2f}%")
        st.dataframe(df)  # Adjust height as needed
    else:
        st.write("No relevant knowledge base results found.")
    
    with st.chat_message("AI"):
        response = st.write_stream(get_response(st.session_state.chat_history, qte, knowledge, temp2, tokens2, RAGSystemMessage, RAGUserMessage, asset))
        st.session_state.chat_history.append(AIMessage(content=response))