import streamlit as st
from code.DiseaseModel import DiseaseModel
from code.helper import prepare_symptoms_array

# Create disease class and load ML model
disease_model = DiseaseModel()
disease_model.load_xgboost('model/xgboost_model.json')

# Set page width to wide
st.set_page_config(layout='wide')

# Custom CSS for background color and text color
st.markdown(
    """
    <style>
        .stApp {
            background-color: #efefef !important;
            color: black !important;
        }
        h1, h2, h3, h4, h5, h6, div, span, label {
            color: black !important;
            background-color:#efefef;
        }
        button, .stButton>button {
            color: black !important;
        }
        
        /* Hide Streamlit header, menu, and footer */
        header, iframe, .css-18ni7ap.e8zbici2, .css-164nlkn.e1g8pov64 { 
            display: none !important; 
        }

        /* Transparent navbar */
        .css-18ni7ap.e8zbici2 {
            background: transparent !important;
        }
    </style>
    """,
    unsafe_allow_html=True
)

# Create sidebar
st.sidebar.markdown('# The Health AI ')
st.sidebar.markdown("This web app uses a machine learning model to predict diseases based on a set of symptoms using Scikit-learn, Python and Streamlit.")
st.sidebar.markdown("Author: S N V S KOMAL")

# Title
st.write('# Symptoms to Disease Prediction')

symptoms = st.multiselect('What are your symptoms?', options=disease_model.all_symptoms)

X = prepare_symptoms_array(symptoms)

# Trigger XGBoost model
if st.button('Predict'):
    # Run the model with the python script
    prediction, prob = disease_model.predict(X)
    st.write(f'## Disease: {prediction} with {prob*100:.2f}% probability')

    tab1, tab2= st.tabs(["Description", "Precautions"])

    with tab1:
        st.write(disease_model.describe_predicted_disease())

    with tab2:
        precautions = disease_model.predicted_disease_precautions()
        for i in range(4):
            st.write(f'{i+1}. {precautions[i]}')