import streamlit as st
import pandas as pd
import plotly.express as px
import pickle

# Charger le jeu de données initial
data = pd.read_csv('concrete.csv')

# Créer un dictionnaire de correspondance entre les noms de colonnes et les noms de variables
column_name_mapping = {
    'cement': 'cement',
    'slag': 'slag',
    'ash': 'ash',
    'water': 'water',
    'superplastic': 'superplastic',
    'fineagg': 'fineagg',
    'age': 'age'
}

# Charger le modèle de régression enregistré avec pickle
with open('best_model.pkl', 'rb') as f:
    best_model = pickle.load(f)

# Configuration de la mise en page
st.set_page_config(layout='wide', initial_sidebar_state='expanded')

# Définir le style CSS
with open('style.css') as f:
    st.markdown(f'<style>{f.read()}</style>', unsafe_allow_html=True)

# Barre latérale avec les paramètres de visualisation et de prédiction
with st.sidebar:
    st.title('Concrete Data APP')
    st.image('concrete.jpg')

    st.subheader('Heat map parameter')
    cement_value = st.selectbox('Age', data['age'].unique())

    st.subheader('Donut chart parameter')
    slag_value = st.selectbox('Select data', data['slag'].unique())

    st.subheader('Line chart parameters')
    plot_data = st.multiselect('Select data', ['cement', 'slag', 'ash'], ['cement', 'slag', 'ash'])
    plot_height = st.slider('Specify plot height', 200, 500, 250)

    st.subheader('Prediction')
    do_prediction = st.checkbox('Perform prediction')

    st.subheader('Upload Dataset')
    uploaded_file = st.file_uploader("Choose a CSV file", type="csv")
    if uploaded_file is not None:
        try:
            data = pd.read_csv(uploaded_file)
        except pd.errors.EmptyDataError:
            st.error("Error: The uploaded CSV file is empty or could not be read.")
        except Exception as e:
            st.error(f"An error occurred: {str(e)}")

# Row A
st.markdown('### Metrics')
col1, col2, col3 = st.columns(3)

# Nombre d'échantillons
num_samples = 905
col1.metric("Number of Samples", 905)

# strength moyen
average_strength = data['superplastic'].mean()
col2.metric("Average superplastic", f"{average_strength:.2f}")

# cement moyen
average_cement = data['cement'].mean()
col3.metric("Average Cement", f"{average_cement:.2f}")

# Row B
st.markdown('### Donut chart')
top_10_ash = data['ash'].value_counts().head(10).index.tolist()
top_10_data = data[data['ash'].isin(top_10_ash)]
donut_fig = px.pie(top_10_data, names='ash', hole=0.5)
st.plotly_chart(donut_fig, use_container_width=True)

# Row C
st.markdown('### Line chart')
st.line_chart(data[plot_data], height=plot_height)

# Prediction Section
if do_prediction:
    st.markdown('### Predicted strength')
    cement = st.number_input('Cement', min_value=100.0, max_value=600.0, value=300.0, step=1.0, key='cement')
    slag = st.number_input('Slag', min_value=0.0, max_value=300.0, value=100.0, step=1.0, key='slag')
    ash = st.number_input('ash', min_value=0.0, max_value=200.0, value=50.0, step=1.0, key='ash')
    water = st.number_input('Water', min_value=100.0, max_value=300.0, value=200.0, step=1.0, key='water')
    superplastic = st.number_input('Superplastic', min_value=0.0, max_value=20.0, value=5.0, step=0.1, key='superplastic')
    coarseagg = st.number_input('Coarse Aggregate', min_value=600.0, max_value=1200.0, value=1000.0, step=1.0, key='coarseagg')
    fineagg = st.number_input('Fine Aggregate', min_value=500.0, max_value=1000.0, value=600.0, step=1.0, key='fineagg')
    age = st.number_input('Age', min_value=1, max_value=365, value=28, step=1, key='age')

    input_data = pd.DataFrame({
        'cement': [cement],
        'slag': [slag],
        'ash': [ash],
        'water': [water],
        'superplastic': [superplastic],
        'coarseagg': [coarseagg],
        'fineagg': [fineagg],
        'age': [age]
    })

    if st.button('Predict'):
        predicted_strength = best_model.predict(input_data)
        st.write(f'Predicted strength: {predicted_strength[0]:.2f}')