app.py

import streamlit as st
import pandas as pd


def load_data():
    return pd.read_csv("benchmark_data.csv")

def case_insensitive_search(data, query, column):
    if query: 
        return data[data[column].str.lower().str.contains(query.lower())]
    return data

def display_table(data, rows_per_page=10):
    container = st.container()
    with container:
        height = min(40 + rows_per_page * 38, 800)  
        st.dataframe(data, height=height)


def main():
    st.title("Multihop-RAG Benchmark Space")

    data = load_data()

    st.sidebar.header("Search Options")
    chat_model_query = st.sidebar.text_input("Search by Chat Model")
    embedding_model_query = st.sidebar.text_input("Search by Embedding Model")
    chunk_query = st.sidebar.text_input("Search by Chunk") 

    if chat_model_query:
        data = case_insensitive_search(data, chat_model_query, 'chat_model')
    if embedding_model_query:
        data = case_insensitive_search(data, embedding_model_query, 'embedding_model')
    if chunk_query:  
        data = case_insensitive_search(data, chunk_query, 'chunk')


    st.header("Benchmark Results")
    st.write("Displaying results for MRR@10, Hit@10, and Accuracy across different frameworks, embedding models, chat models, and chunks.")
    display_table(data) 


    if st.sidebar.checkbox("Show Metrics Distribution"):
        st.subheader("Metrics Distribution")
        st.bar_chart(data[['MRR@10', 'Hit@10', 'Accuracy']]) 


    st.sidebar.header("Citation")
    st.sidebar.info(
        "Please cite this dataset as:\n"
        "Author et al. (2024). Multihop-RAG Benchmark Dataset. Retrieved from [Source URL]."
    )
    st.markdown("---")
    st.caption("For citation, please use: 'Author et al. (2024), Multihop-RAG Benchmark Dataset, Retrieved from [Source URL].'")

if __name__ == "__main__":
    main()