src/stwebm_parquet_wa.py

# -*- coding: utf-8 -*-
"""
Created on Sun Dec 11 19:51:02 2022
Modified on Sun Aug 24 22:45:00 2025 to streamline code
@author: ruben
"""

import streamlit as st
from biconWebStmParquetWa import regex_search as rs

def init_session_state():
    """Initializes the session state variables."""
    if 'queryresults' not in st.session_state:
        st.session_state.queryresults = None
    if 'page' not in st.session_state:
        st.session_state.page = 1
    if 'maxpage' not in st.session_state:
        st.session_state.maxpage = 0
    if 'minpage' not in st.session_state:
        st.session_state.minpage = 1
    if 'datasize' not in st.session_state:
        st.session_state.datasize = 0
    if 'chunksize' not in st.session_state:
        st.session_state.chunksize = 20
    if 'success_messages' not in st.session_state:
        st.session_state.success_messages = []

def reset_session_state():
    """Resets the session state for a new session."""
    st.session_state.queryresults = None
    st.session_state.page = 1
    st.session_state.maxpage = 0
    st.session_state.minpage = 1
    st.session_state.datasize = 0
    st.session_state.success_messages = []

def buildTable(page):
    """Builds the HTML table for the current page of results."""
    slices = st.session_state.queryresults
    datasize = st.session_state.datasize
    table = '<table width="100%">' # Corrected: escaped double quotes within the string
    n = st.session_state.chunksize
    for j in range(n):
        index = (page - 1) * n + j
        if index >= datasize:
            break
        try:
            corpus, score, en, zh = slices[page - 1][j].split('\t') # Corrected: escaped tab character
        except:
            continue
        table += (
            '<tr>'
            f'<td>{corpus}</td><td colspan=2>{score}</td>'
            '</tr>'
            '<tr>'
            f'<td>{index + 1}</td><td width="45%" valign="top">{en}</td><td width="50%" valign="top">{zh}</td>' # Corrected: escaped double quotes within the string
            '</tr>'
        )
    table += '</table>'
    return table

def main():
    """Main function to run the Streamlit app."""
    appTitle = '臺大法規雙語查詢系統'
    sources = ('NTURegs', 'VOA')

    st.set_page_config(
        page_title=appTitle,
        layout='wide',
        initial_sidebar_state='auto',
        menu_items={
            'Get Help': 'https://streamlit.io/',
            'Report a bug': 'https://github.com',
            'About': f'**{appTitle}**\nCopyright (c) Ruben G. Tsui' # Corrected: escaped newline character
        }
    )

    page_style = '''
        <style>
            .css-o18uir.e16nr0p33 {
              margin-top: -125px;
            }
            .reportview-container .css-1lcbmhc .css-1outpf7 {
                padding-top: -125px;
            }
            .reportview-container .main .block-container{
                padding-top: 0rem;
                padding-right: 0rem;
                padding-left: 0rem;
                padding-bottom: 0rem;}
            p.europe {
				font-family: Source Pro, Consolas, LingWai TC, Menlo, Courier New, Arial;
                font-size: 16px;
			}
			p.cjk {
				font-family: Microsoft Jhenghei, Source Han Sans, Noto Sans CJK TC Regular, Hiragino Sans CNS, LantingHei TC, Source Han Serif;
                font-size: 18px;
			}
        </style>
    '''
    st.markdown(page_style, unsafe_allow_html=True)

    st.sidebar.subheader(appTitle)

    table_placeholder = st.empty()

    with st.sidebar:
        #if st.button("New Session"):
        #    reset_session_state()
        #    table_placeholder.empty()

        query = st.text_area('輸入搜尋字串').strip()
        multicorpora = st.multiselect('選擇語料庫（可複選）', sources, ['NTURegs'])

        colc, cold = st.columns([1, 1])
        with colc:
            submit_button = st.button('搜尋')
        with cold:
            regex_search = st.radio("Regex search", ["Yes", "Always"], horizontal=True)

        cola, colb = st.columns([1, 1])
        with cola:
            size = st.selectbox('筆數上限', [10, 20, 50, 100, 200, 500, 5000], index=2)
        with colb:
            case_sensitive = st.radio("Case sensitive", ["No", "Yes"], horizontal=True)

        st.session_state.chunksize = st.slider("每頁筆數", 1, 50, 20)

        # Build user interface
        col1, col2, col3, col4 = st.columns([1, 1, 1, 1])
        with col1:
            first_button = st.button('First')
        with col2:
            prev_button = st.button('Prev')
        with col3:
            next_button = st.button('Next')
        with col4:
            last_button = st.button('Last')

    # Navigation
    if next_button and st.session_state.page < st.session_state.maxpage:
        st.session_state.page += 1
    if prev_button and st.session_state.page > st.session_state.minpage:
        st.session_state.page -= 1
    if first_button:
        st.session_state.page = st.session_state.minpage
    if last_button:
        st.session_state.page = st.session_state.maxpage

    if submit_button:
        reset_session_state()
        all_results = []
        success_messages = []
        success_messages.append('No. of matches found: ')
        for c in multicorpora:
            selectedCorpus = sources.index(c)
            results = rs(query, c=selectedCorpus, max_matches=size, case_sensitive=(case_sensitive == "Yes"), literal=False)
            success_messages.append(f'[{c}]: {len(results)}')
            all_results.extend(results)
        
        st.session_state.success_messages = success_messages

        datasize = len(all_results)
        n = st.session_state.chunksize
        slices = [all_results[i:i + n] for i in range(0, datasize, n)]
        pagesize = len(slices)

        st.session_state.datasize = datasize
        st.session_state.maxpage = pagesize
        st.session_state.queryresults = slices

    if st.session_state.queryresults is not None:
        table_placeholder.empty()
        with table_placeholder.container():
            if st.session_state.success_messages:
                messages = " | ".join(st.session_state.success_messages)
                st.markdown(f'''
                <div style="
                    border: 1px solid yellow;
                    padding: 5px;
                    border-radius: 5px;
                    font-size: 0.9em;
                    margin-bottom: 10px;
                    background-color: black;
                    color: yellow;
                    margin-top: -25px;">
                    {messages}
                </div>
                ''', unsafe_allow_html=True)

            st.markdown(f"page {st.session_state.page} of {st.session_state.maxpage}")
            table = buildTable(st.session_state.page)
            st.markdown(table, unsafe_allow_html=True)

if __name__ == '__main__':
    init_session_state()
    main()