# flask, pandas, sci-kit learn, pickle-mixin
from flask import Flask, render_template, request
import pandas as pd
import pickle
import sklearn
import numpy as np

app =Flask(__name__)

model = pickle.load(open('LinearRegressionModel.pkl','rb'))
car = pd.read_csv('cleaned car.csv')

@app.route('/')
def index():
    companies = sorted(car['company'].unique())
    car_models = sorted(car['name'].unique())
    year = sorted(car['year'].unique(), reverse=True)
    fuel_type = car['fuel_type'].unique()
    companies.insert(0,'Select Company')
    return render_template('index.html', companies=companies, car_models=car_models,years=year,fuel_types=fuel_type)

@app.route('/predict',methods=['POST'])
def predict():
    company= request.form.get('company')
    car_model = request.form.get('car_model')
    year = int(request.form.get('year'))
    fuel_type = request.form.get('fuel_type')
    kms_driven = int(request.form.get('kilo_driven'))

    prediction = model.predict(pd.DataFrame([[car_model, company,year,kms_driven,fuel_type]], columns=['name','company','year','kms_driven','fuel_type']))
    
    return str(np.round(prediction[0],2))


if __name__=='__main__':
    app.run(debug=True)