prediction.py

import streamlit as st
import pandas as pd 
import numpy as np
import tensorflow_hub as tf_hub
from tensorflow.keras.models import load_model
import re
import nltk
nltk.download('punkt')
from nltk.tokenize import word_tokenize
nltk.download('stopwords')
from nltk.corpus import stopwords
nltk.download('wordnet')
from nltk.stem import WordNetLemmatizer

# Load the model
loaded_model = load_model('model_rnn')

# Create a dictionary to map the labels to the categories
label_dict = {0: 'Uang Masuk', 1: 'Uang Keluar', 2: 'Pinjaman', 3: 'Tagihan', 4: 'Top Up',
             5: 'Biaya & Lainnya', 6: 'Transportasi', 7: 'Pendidikan', 8: 'Hadiah & Amal',
             9: 'Belanja', 10: 'Hiburan',11: 'Makanan & Minuman', 12: 'Kesehatan',
             13: 'Perawatan Diri', 14: 'Hobi & Gaya Hidup', 15: 'Pencairan Investasi',
             16: 'Tabungan & Investasi'}

def preprocessing(text):
    ''' 
    Preprocessing text by applying lowercasing, normalization, tokenization, stopword removal, and lemmatization
    '''
    # Lowercase the text
    text = text.lower()

    # Normalize the text
    text = re.sub(r'\d+', '', text)  # Remove numbers
    text = re.sub(r'[^\w\s]', '', text)  # Remove punctuation
    text = re.sub(r'\s+', ' ', text).strip()  # Remove whitespaces

    # Tokenize the text
    tokens = word_tokenize(text)

    # Get the English stopwords
    stop_words = set(stopwords.words('indonesian'))
    stop_words.update(['the', 'yg', 'gk', 'nyagak', 'pake', 'pakai', 'i', "and"])

    # Remove stopwords
    tokens = [word for word in tokens if word not in stop_words]

    # Lemmatize the text
    lemmatizer = WordNetLemmatizer()
    tokens = [lemmatizer.lemmatize(word) for word in tokens]

    # Combine tokens back into a single string
    text = ' '.join(tokens)

    return text

def run():
    st.title('Notes Categorization')

    default = "konser twice"

    user_input = st.text_area("Enter the notes text here:", default, height=50)

    if st.button('Predict'):
        # Apply the function to the 'Text' column in the data
        text_processed = preprocessing(user_input)

        # The model expects input data in batch, even if just predicting on one sample
        # So, I'll add an extra dimension with np.expand_dims
        preprocessed_notes = np.expand_dims(text_processed, axis=0)

        # get the prediction
        predictions = loaded_model.predict(preprocessed_notes)

        # get the class with the highest probability
        predicted_class = np.argmax(predictions[0])

        # Decode the predicted class into the original category
        predicted_category = label_dict[predicted_class]

        st.write(f'The predicted category is: {predicted_category}')

if __name__ == '__main__':
    main()