erinak/test3

1import numpy as np2import pandas as pd3from gensim.corpora import Dictionary, MmCorpus4from gensim.models import LdaModel, Word2Vec5import matplotlib.pyplot as plt6import streamlit as st7from pyLDAvis import prepared_data_to_html8import pyLDAvis.gensim_models as gensimvis9 10# 生データ・コーパス・辞書・モデルのロード11df = pd.read_csv("./raw_corpus.csv")12corpus = MmCorpus('./corpus.mm')13dict = Dictionary.load(f'./livedoor_demo.dict')14lda = LdaModel.load('./lda_demo.model')15 16st.caption("生データ一覧")17st.dataframe(df.iloc[:100])18 19st.caption("記事のカテゴリ")20fig, ax = plt.subplots()21count = df[["CATEGORY", "DOCUMENT"]].groupby("CATEGORY").count()22count.plot.pie(y="DOCUMENT", ax=ax, ylabel="", legend=False)23st.pyplot(fig)24 25# pyLDAvisによるトピックの可視化26vis = gensimvis.prepare(lda, corpus, dict)27html_string = prepared_data_to_html(vis)28st.components.v1.html(html_string, width=1300, height=800)29