app.py

import streamlit as st
import pandas as pd
import re

st.set_page_config(layout='wide')

def load_data():
    return pd.read_csv(filepath_or_buffer='./data.csv')

df = load_data()

if 'idx' not in st.session_state:
    st.session_state.idx = 0

st.markdown("<h1 style='text-align: center;'>Investigation Legal Documents Dataset Checker</h1>", unsafe_allow_html=True)


col_1, col_2, col_3, col_4, col_5, col_6, col_7, col_8, col_9, col_10 = st.columns([1, 1, 1, 1, 1, 1, 1, 1, 1, 1])
btn_prev = col_1.button(label='Previous sample', use_container_width=True)
btn_next = col_2.button(label='Next sample', use_container_width=True)
btn_save = col_3.button(label='Save changes', use_container_width=True)

if btn_prev:
    if st.session_state.idx > 0:
        st.session_state.idx -= 1

if btn_next:
    if st.session_state.idx < len(df) - 1:
        st.session_state.idx += 1

st.markdown(f"<h3 style='text-align: center;'>Sample: {st.session_state.idx+1}/{len(df)}</h3>", unsafe_allow_html=True)

context = st.text_area(label='Your context: ', value=df['contexts'][st.session_state.idx], height=300)
question = st.text_area(label='Your question: ', value=df['questions'][st.session_state.idx], height=100)
answer = st.text_area(label='Your answer: ', value=df['answers'][st.session_state.idx], height=100)

if answer.strip() and context.strip():
    highlighted_context = re.sub(re.escape(answer), "<mark>" + answer + "</mark>", context, flags=re.IGNORECASE)
    st.markdown(highlighted_context, unsafe_allow_html=True)

if btn_save:
    df.loc[st.session_state.idx, 'contexts'] = context
    df.loc[st.session_state.idx, 'questions'] = question
    df.loc[st.session_state.idx, 'answers'] = answer
    
    df.to_csv('./data.csv', index=False)