anveshak / app.py
ankanghosh's picture
Update app.py
9ac9057 verified
raw
history blame
4.62 kB
import streamlit as st
import time
# FIRST: Set page config before ANY other Streamlit command
st.set_page_config(page_title="Indian Spiritual RAG")
# Initialize ALL session state variables
if 'initialized' not in st.session_state:
st.session_state.initialized = False
if 'model' not in st.session_state:
st.session_state.model = None
if 'tokenizer' not in st.session_state:
st.session_state.tokenizer = None
if 'last_query' not in st.session_state:
st.session_state.last_query = ""
if 'submit_clicked' not in st.session_state:
st.session_state.submit_clicked = False
if 'query_answered' not in st.session_state:
st.session_state.query_answered = False # Tracks if query was answered
# THEN: Import your modules
from rag_engine import process_query, load_model
from utils import setup_all_auth
# Create a placeholder for initialization message
init_message = st.empty()
# βœ… Custom CSS for border & button fixes
st.markdown("""
<style>
/* βœ… Green border by default */
div[data-baseweb="input"] {
border: 2px solid #4CAF50 !important;
border-radius: 8px !important;
transition: border 0.3s ease-in-out;
}
/* πŸ”΄ Change to red while typing */
div[data-baseweb="input"]:focus-within {
border: 2px solid #FF5722 !important;
}
/* βœ… Ensure it turns green again AFTER submission */
.answered div[data-baseweb="input"] {
border: 2px solid #4CAF50 !important;
}
/* βœ… Fully green button */
.stButton>button {
background-color: #4CAF50 !important;
color: white !important;
border: none !important;
border-radius: 8px !important;
padding: 8px 16px;
font-size: 16px;
font-weight: bold;
cursor: pointer;
}
</style>
""", unsafe_allow_html=True)
# Handle initialization
if not st.session_state.initialized:
init_message.info("Hang in there! We are setting the system up for you. 😊")
try:
setup_all_auth()
load_model()
st.session_state.initialized = True
st.session_state.query_answered = False # Reset query state
init_message.success("System initialized successfully!")
time.sleep(1)
init_message.empty()
except Exception as e:
init_message.error(f"Error initializing: {str(e)}")
# Function to handle form submission
def handle_form_submit():
if st.session_state.query_input:
st.session_state.last_query = st.session_state.query_input
st.session_state.submit_clicked = True
st.session_state.query_answered = True # βœ… Ensure it resets to green
st.session_state.query_input = "" # βœ… Clear input
st.experimental_rerun() # βœ… Refresh UI properly
# βœ… Wrap the text input inside a div for styling
css_class = "answered" if st.session_state.query_answered else ""
st.markdown(f'<div class="{css_class}">', unsafe_allow_html=True)
# βœ… Create a form for user input
with st.form(key="query_form"):
query = st.text_input("Ask your question:", key="query_input", placeholder="Type here...")
submit_button = st.form_submit_button("Get Answer", on_click=handle_form_submit)
st.markdown("</div>", unsafe_allow_html=True) # βœ… Close styling div
# βœ… Display the last question after submission
if st.session_state.last_query:
st.markdown("### Current Question:")
st.info(st.session_state.last_query)
# Sliders for customization
col1, col2 = st.columns(2)
with col1:
top_k = st.slider("Number of sources:", 3, 10, 5)
with col2:
word_limit = st.slider("Word limit:", 50, 500, 200)
# βœ… Process the query if submitted
if st.session_state.submit_clicked and st.session_state.last_query:
st.session_state.submit_clicked = False # Reset flag
with st.spinner("Processing your question..."):
try:
result = process_query(st.session_state.last_query, top_k=top_k, word_limit=word_limit)
st.subheader("Answer:")
st.write(result["answer_with_rag"])
st.subheader("Sources:")
for citation in result["citations"].split("\n"):
st.write(citation)
st.session_state.query_answered = True # βœ… Mark query as answered
st.experimental_rerun() # βœ… Proper rerun to refresh UI
except Exception as e:
st.error(f"Error processing query: {str(e)}")
# Add helpful information
st.markdown("---")
st.markdown("""
### About this app
This application uses a Retrieval-Augmented Generation (RAG) system to answer questions about Indian spiritual texts.
It searches through a database of texts to find relevant passages and generates answers based on those passages.
""")