ahdsoft/Persian-Topic-Modeling

1import streamlit as st2# import numpy as np3import pandas as pd4from topic_modeling import TopicModeling5st.set_page_config(page_title='تحلیل‌گر متن عهد', page_icon = './ahd_logo.png', layout = 'wide')6@st.cache_resource7def get_model():8    tp_model = TopicModeling()9    return tp_model10 11tp_model = get_model()12 13 14 15col1, col2, col3 = st.columns(3)16with col2:17    st.title("تحلیل اسناد متنی")18 19# Upload CSV file20uploaded_file = st.file_uploader("آپلود فایل")21if uploaded_file is not None:22    filename = uploaded_file.name23    if filename.endswith('.xlsx'):24        df = pd.read_excel(uploaded_file)25    elif filename.endswith('.csv'):26        df = pd.read_csv(uploaded_file)27    else:28        raise ValueError('Unsupported file format')29    30    # Show first 10 rows of dataframe31    st.write(df.head(10))32    33    # Select columns to use for topic modeling34    cols = st.multiselect("ستون‌های متنی موردنظر را انتخاب نمایید", df.columns)35    ratio = st.slider('چند درصد از کل دادگان پردازش شود',min_value=0, max_value=100) 36    col1, col2, col3 , col4, col5 = st.columns(5)37    with col3:38        done_button = st.button("پردازش دادگان")39    if done_button:40        # print('colssssssssssssss ', cols)41        # Concatenate selected text columns42        df = df[cols]43        df = df.head(int(len(df) * (ratio/100)))44        df = df.dropna()45        # text = df.apply(lambda x:' '.join(x), axis=1)46 47        # Run topic modeling function48        col1, col2, col3 = st.columns(3)49        with col2:50            data_progress = st.spinner('در حال پردازش دادگان')51            with data_progress:52                docs = tp_model.add_data(df)53            st.success('پردازش دادگان با موفقیت به پایان رسید')  54            # print('before docs')55            with st.spinner('در حال آموزش مدل'):56                # print('fittttttttttt')57                tp_model.fit(docs)58            st.success('آموزش پایان یافت')59        col1, col2, col3 = st.columns(3)60        with col3:61            st.title(" فضای تاپیک‌ها ")62            st.header("")63        fig = tp_model.get_vis_topics()64        st.plotly_chart(65            fig, 66            use_container_width=True,67            theme="streamlit",  # ✨ Optional, this is already set by default!68        )69        col1, col2, col3 = st.columns(3)70        with col3:71            st.title(" کلمات هر تاپیک ")72            st.header("")73        fig = tp_model.get_barchart()74        st.plotly_chart(75            fig, 76            use_container_width=True,77            theme="streamlit",  # ✨ Optional, this is already set by default!78        )79 80 81        col1, col2, col3 = st.columns(3)82        with col3:83            st.title("لیست تاپیک‌ها")84            st.header("")85        topics_info = tp_model.get_topic_info()86        st.write(topics_info)87 88        col1, col2, col3 = st.columns(3)89        with col3:90            st.title(" ابر کلمات ")91            st.header("")92        # # figs = tp_model.get_wordcloud()93        # topic_counts = len(tp_model.topic_model.get_topic_info())94        # print('topic count ', topic_counts)95        # if topic_counts > 15:96        #     topic_counts = 1597        # for topic_index in range(topic_counts):98        #     fig = tp_model.get_wordcloud_by_topic(topic_index)99        #     if fig:100        figs = tp_model.get_wordcloud()101        for fig in figs:102            st.header("")103            st.markdown('topic:')104            # st.title(f'topic:{topic_index}')105            st.pyplot(fig)106 107 108 109 110 111        112