import streamlit as st
from pybanking.churn_prediction import model_churn
import pickle
import sklearn.metrics as metrics
from mlxtend.plotting import plot_confusion_matrix
import matplotlib.pyplot as plt

st.set_page_config(page_title="Customer Churn Prediction Model")

st.title('Customer Churn Prediction Model')
# x = st.slider('Select a value')

st.subheader('This is the Sample Data')

df = model_churn.get_data()

st.dataframe(df.head(5))

model_names = [
     "Logistic_Regression",
     "Support_Vector_Machine",
     "Support_Vector_Machine_Optimized",
     "Decision_Tree",
     "Neural_Network",
     "Random_Forest",
     "Pycaret_Best"
     ]

option = st.selectbox(
     'Select a model to be used',
     model_names
     )

st.write("Model Loaded :  ", option)

X, y = model_churn.preprocess_inputs(df, option)

model = pickle.load(open(option+'.pkl', 'rb'))

option2 = st.selectbox(
     'Which dataset would you like to use for prediction?',
     ['Sample Dataset', 'Upload Custom']
     )

if option2 == 'Upload custom':
     model = model_churn.train(df, model)

st.dataframe(X.head(5))

y_pred = model.predict(X)

st.write("Accuracy:",metrics.accuracy_score(y, y_pred))
st.write("Precision:",metrics.precision_score(y, y_pred))
st.write("Recall:",metrics.recall_score(y, y_pred))

fig, ax = plot_confusion_matrix(conf_mat=metrics.confusion_matrix(y, y_pred), figsize=(6, 6), cmap=plt.cm.Reds, colorbar=True)
plt.xlabel('Predictions', fontsize=18)
plt.ylabel('Actuals', fontsize=18)
plt.title('Confusion Matrix', fontsize=18)
st.pyplot(fig)