app.py

import os
import pandas as pd
import streamlit as st
from sklearn.linear_model import LinearRegression

def predict_hotel_price(train_features_path, train_label_path, test_features_path):
    # Baca data dari file train_features.csv
    train_features = pd.read_csv(train_features_path)

    # Baca data dari file train_label.csv
    train_label = pd.read_csv(train_label_path)

    # Gabungkan kedua dataframe berdasarkan indeks
    df_merged = pd.concat([train_features, train_label], axis=1)

    # Tambahkan kolom 'id' di paling kiri dengan menggunakan range indeks
    df_merged.insert(0, 'ID', range(len(df_merged)))

    # Simpan dataframe ke dalam file CSV
    df_merged.to_csv('merged_data.csv', index=False)

    # Baca file merged_data.csv sebagai hasil prapemrosesan
    hasil_features = pd.read_csv('merged_data.csv')

    # Prapemrosesan data pada kolom rating dengan mengubah format string menjadi float
    hasil_features['rating'] = hasil_features['rating'].apply(lambda x: float(x.split()[0]) if isinstance(x, str) and len(x.split())>0 and x.split()[0].replace('.','').isdigit() else None)
    hasil_features['Price'] = hasil_features['Price'].apply(lambda x: float(x.replace(',', '').replace('avg/night', '')) if isinstance(x, str) else x)

    # Menghilangkan missing value pada kolom rating
    hasil_features.dropna(subset=['rating'], inplace=True)
    hasil_features = hasil_features.drop(['facilities', 'location'], axis=1)

    # Membuat model Linear Regression
    model = LinearRegression()

    # Melatih model dengan dataset train
    model.fit(hasil_features.drop(['ID', 'Price'], axis=1), hasil_features['Price'])

    # Membaca dataset test dan menghapus kolom facilities, location, dan ID
    test_features = pd.read_csv(test_features_path)
    test_features = test_features.drop(['facilities', 'location', 'ID'], axis=1)

    # Prapemrosesan data pada kolom rating dengan mengubah format string menjadi float
    test_features['rating'] = test_features['rating'].apply(lambda x: float(x.split()[0]) if isinstance(x, str) else x)

    # Melakukan prediksi terhadap dataset test
    predictions = model.predict(test_features)

    # Convert predictions to a pandas dataframe
    predictions_df = pd.DataFrame(predictions, columns=['Price'])

    # Add the 'ID' column using square bracket notation
    predictions_df.insert(loc=0, column='ID', value=range(len(predictions_df)))

    # mengubah nilai kolom Price menjadi bilangan bulat
    predictions_df['Price'] = predictions_df['Price'].astype(int)

    # Membuat file CSV dari dataframe predictions_df
    predictions_df.to_csv('predictions.csv', index=False)
    return predictions_df

def main():
    st.title("Hotel Price Prediction With Linear Regression")
    st.write("Memprediksi Harga Hotel Berdasarkan Rating")

    # Membuat list nama file dari direktori yang berisi file input
    input_dir = 'dataset'
    input_files = os.listdir(input_dir)

    # Mengubah list nama file menjadi opsi dropdown
    train_features_path = st.selectbox("Train Features = 'Berisi Fitur-Fitur Dari Data Latih'", [os.path.join(input_dir, file) for file in input_files])
    train_label_path = st.selectbox("Train Label = 'Berisi Label Dari Data Latih'", [os.path.join(input_dir, file) for file in input_files])
    test_features_path = st.selectbox("Test Features = 'Berisi Fitur-Fitur Dari Data Uji'", [os.path.join(input_dir, file) for file in input_files])

    # Menjalankan fungsi predict_hotel_price dan menampilkan hasilnya
    if st.button("Prediksi Hasil Harga"):
        predictions_df = predict_hotel_price(train_features_path, train_label_path, test_features_path)
        st.write(predictions_df)
        st.download_button(
            label="Download Hasil Prediksi CSV",
            data=predictions_df.to_csv(index=False),
            file_name="predictions.csv",
            mime="text/csv"
        )

if __name__ == '__main__':
    main()