app.py

import streamlit as st
import joblib
import datetime
import snscrape.modules.twitter as sntwitter
import pandas as pd
from tensorflow.keras.models import load_model
import nltk
import re
import tensorflow as tf
import numpy as np
from nltk.corpus import stopwords
from nltk.tokenize import word_tokenize
from Sastrawi.Stemmer.StemmerFactory import StemmerFactory
import matplotlib.pyplot as plt
import seaborn as sns

nltk.download('stopwords')
nltk.download('punkt')
nltk.download('wordnet')
nltk.download('omw-1.4')
nlp = load_model('model_lstm.tf')

def run():
    # Creating list to append tweet data to
    attributes_container = []

    stock = st.selectbox('Pick a stock:', ('BBNI', 'BBRI', 'BBTN', 'BMRI'))
    
    today = datetime.datetime.now()
    today = today.strftime('%Y-%m-%d')
    yesterday = datetime.datetime.now() - datetime.timedelta(days=1)
    yesterday = yesterday.strftime('%Y-%m-%d')

    # Using TwitterSearchScraper to scrape data and append tweets to list
    for i,tweet in enumerate(sntwitter.TwitterSearchScraper(stock.lower() + ' since:' + yesterday + ' until:' + today).get_items()):
        attributes_container.append([tweet.user.username, tweet.date, tweet.likeCount, tweet.sourceLabel, tweet.content])

    # Creating a dataframe to load the list
    tweets_df = pd.DataFrame(attributes_container, columns=["User", "Date Created", "Number of Likes", "Source of Tweet", "Tweet"])
    tweets = tweets_df[tweets_df['Tweet'].str.contains(stock)]
    tweets_daily = pd.DataFrame(pd.to_datetime(tweets['Date Created']).dt.tz_localize(None))
    tweets['Date Created'] = tweets_daily
    tweets['Date Created'] = pd.to_datetime(tweets['Date Created']).dt.date
    tweets

    if st.button('Predict sentiment'):
        df = pd.DataFrame()
        
        # Mendefinisikan stopwords bahasa Indonesia
        idn = list(set(stopwords.words('indonesian')))
        factory = StemmerFactory()
        stemmer = factory.create_stemmer()

        def text_process(text):
            # Mengubah Teks ke Lowercase
            text = text.lower()
            
            # Menghilangkan Mention
            text = re.sub("@[A-Za-z0-9_]+", " ", text)
            
            # Menghilangkan Hashtag
            text = re.sub("#[A-Za-z0-9_]+", " ", text)
            
            # Menghilangkan \n
            text = re.sub(r"\\n", " ",text)
            
            # Menghilangkan Whitespace
            text = text.strip()

            # Menghilangkan Link
            text = re.sub(r"http\S+", " ", text)
            text = re.sub(r"www.\S+", " ", text)

            # Menghilangkan yang Bukan Huruf seperti Emoji, Simbol Matematika (seperti μ), dst
            text = re.sub("[^A-Za-z\s']", " ", text)

            # Menghilangkan RT
            text = re.sub("rt", " ",text)

            # Melakukan Tokenisasi
            tokens = word_tokenize(text)

            # Menghilangkan Stopwords
            text = ' '.join([word for word in tokens if word not in idn])

            # Melakukan Stemming
            text = ' '.join([stemmer.stem(word) for word in text.split()])
            
            return text

        df['Tweet_processed'] = tweets['Tweet'].apply(lambda x: text_process(x))
        pred = np.argmax(nlp.predict(df['Tweet_processed']), axis=-1)
        pred_df = pd.DataFrame(pred, columns=['label'])

        pred_df['label'] = pred_df.replace([0,1,2], ['positive','neutral','negative'])

        def PieComposition(dataframe, column):
            palette_color = sns.color_palette('pastel')
            data = {}
            freq = {}
            datalen = len(dataframe[column].unique())
            x = np.arange(datalen)
            dq = dataframe[column].unique()
            for i in x:
                data[i] = dq[i]
                freq[i] = dataframe[column][dataframe[column] == dq[i]].value_counts().sum()
            data = list(data.values())
            freq = list(freq.values())
            fig = plt.figure(figsize=(15, 5))
            plt.pie(freq, labels = data, colors=palette_color, autopct='%.0f%%')
            plt.show()
            st.pyplot(fig)

        PieComposition(pred_df, 'label')

if __name__ == '__main__':
    run()