import os
from dotenv import load_dotenv
from pathlib import Path

import mlflow 
import streamlit as st
import pandas as pd
from pycaret.classification import *

# dotenv_path = Path('.env')
# load_dotenv(dotenv_path=dotenv_path)

# load model
# os.environ['MLFLOW_TRACKING_USERNAME'] = 'fandanabil1379'
# os.environ['MLFLOW_TRACKING_PASSWORD'] = 'dadc32f6246f307c2fe4928f3074068f628b79ba'
# mlflow.set_tracking_uri('https://dagshub.com/fandanabil1379/loan_prediction.mlflow')
# model = mlflow.sklearn.load_model(f"models:/v1.0.1/Production")

# model = load_model('model')

# mlflow.set_tracking_uri('https://dagshub.com/fandanabil1379/loan_prediction.mlflow')
# logged_model = 'runs:/a45156c210f149b2abbfe15e5b824cc4/model'

# # Load model as a PyFuncModel.
# loaded_model = mlflow.pyfunc.load_model(logged_model)

import joblib
loaded_model = joblib.load('model.pkl')

def run():
   # init
   st.set_page_config(page_title="Loan Default Prediction App")
   st.title('Loan Default Prediction')

   uploaded_file = st.file_uploader("Choose a file", type={"csv"})
   if uploaded_file is not None:
      # do prediction
      data = pd.read_csv(uploaded_file)
      prediction = loaded_model.predict(data)

      # show the result
      st.write(prediction)

      # download the result
      # csv = prediction.to_csv(index=False).encode('utf-8')
      # if st.download_button('Download Prediction', csv, 'prediction.csv'):
      #    st.write('Thanks for downloading!')

run()