app.py

# Library Load Model
import pandas as pd
import numpy as np
import pickle
import streamlit as st
# Library Pre-Processing
import nltk
import re
import tensorflow as tf
from nltk.corpus import stopwords
from nltk.tokenize import word_tokenize
from nltk.stem import WordNetLemmatizer
nltk.download('punkt')
nltk.download('stopwords')
nltk.download('wordnet')

model_path= 'model'

# Load model
model = tf.keras.models.load_model(model_path)

# Define Stopwords
## Load Stopwords from NLTK
from nltk.corpus import stopwords
stop_words_en = stopwords.words("english")

print('Stopwords from NLTK')
print(len(stop_words_en), stop_words_en)
print('')

## Create A New Stopwords
new_stop_words = ['aye', 'mine', 'have']

# Define Lemmatizer
lemmatizer = WordNetLemmatizer()

## Merge Stopwords
stop_words_en = stop_words_en + new_stop_words
stop_words_en = list(set(stop_words_en))
print('Out Final Stopwords')
print(len(stop_words_en), stop_words_en)

# Create A Function for review Preprocessing

def review_preprocessing(review):
  # Case folding
    review = review.lower()

  # Mention removal
    review = re.sub("@[A-Za-z0-9_]+", " ", review)

  # Hashtags removal
    review = re.sub("#[A-Za-z0-9_]+", " ", review)

  # Newline removal (\n)
    review = re.sub(r"\\n", " ",review)

  # Whitespace removal
    review = review.strip()

  # URL removal
    review = re.sub(r"http\S+", " ", review)
    review = re.sub(r"www.\S+", " ", review)

  # Non-letter removal (such as emoticon, symbol (like μ, $, 兀), etc
    review = re.sub("[^A-Za-z\s']", " ", review)
    review = re.sub("['ï']", " ", review)
    review = re.sub("['¿']", " ", review)
    review = re.sub("['½']", " ", review)
    review = re.sub("['ý']", " ", review)
  # Tokenization
    tokens = word_tokenize(review)

  # Stopwords removal
    tokens = [word for word in tokens if word not in stop_words_en]

  # Lemmetize
    tokens = [lemmatizer.lemmatize(word) for word in tokens]

  # Combining Tokens
    review = ' '.join(tokens)

    return review

def preprocess_text(text):
    '''
    Function to preprocess text by cleaning, removing stopwords, and lemmatizing.

    Parameters:
        text (str): The input text to be preprocessed.

    Returns:
        str: The preprocessed text.
    '''
    text = review_preprocessing(text)
    return text

def run():
    # membuat title
    st.title("DETECTION RATING BASED ON MCDONALD'S CUSTOMER REVIEW")
    st.subheader('Detecting Reviews')
    st.markdown('---')
    # Buat form
    with st.form(key='review'):
        st.write("## Customers' Review")
        # URL input
        text = st.text_input("Enter The Review:")
        submitted = st.form_submit_button('Predict')
        # Perform prediction
        if submitted:
                df_inf = {'preprocessing_review': text}
                df_inf = pd.DataFrame([df_inf])
                df_inf['preprocessing_review'] = df_inf['preprocessing_review'].apply(lambda x: review_preprocessing(x))
              

                y_pred_inf = model.predict(df_inf['preprocessing_review'])
                y_pred_inf = np.argmax(df_inf['preprocessing_review'], axis = -1)

                # Display the prediction result
                if y_pred_inf == 0:
                    st.subheader("Prediction: Negative Comment with Rating 1 Star - 2 Stars")
                elif y_pred_inf == 1:
                    st.subheader("Prediction: Neutral Comment with Rating 3 Stars")
                else:
                    st.subheader("Prediction: Positive Comment with Rating 5 Stars")

                # Display the extracted text
                st.subheader("Extracted Text:")
                st.write(text)

if __name__ == '__main__':
    run()