app.py

from flask import Flask, render_template, request
import pickle
from datetime import datetime
import numpy as np
import pandas as pd
app = Flask(__name__, template_folder='templates', static_folder='images')
data = pd.read_csv('./dataset/maxdata.csv')
modelfile = open('./model/rfbest.pkl', 'rb')
model = pickle.load(modelfile)
label = pickle.load(modelfile)

@app.route('/')
def home():
    return render_template('index.html')

@app.route('/predict', methods=['POST'])
def predict():
    val = [x for x in request.form.values()]
    s = val[0]
    d = val[1]
    c = val[2]
    ct1 = val[3]
    ct2 = val[4]
    df = data[(data.source == s) & (data.destination == d)]
    if c == '2':
        n = ct2
    elif c == '1':
        n = ct1
    if df.empty:
        res = -1
    elif s == '--Choose a Source--' or d == '--Choose a Destination--' or c == '--Choose a Cab--' or (ct1 == '--Choose a Cab type--' and ct2 == '--Choose a Cab type--'):
        res = -1
    else:
        distance = df[(df.name == n)]['distance'].iloc[0]
        price = df[(df.name == n)]['price'].iloc[0]
        ct = np.where(label == n)
        t = datetime.now()
        hour = int(t.strftime('%H'))
        wd = t.weekday()
        if wd > 4:
            is_weekday = 0
        else:
            is_weekday = 1
        if 7 <= hour <= 19:
            is_oh = 1
        else:
            is_oh = 0
        features = np.array([distance, price, ct[0][0], is_weekday, is_oh]).reshape(1, 5)
        pred = model.predict(features)
        res = round(pred[0], 2)
    return render_template('index.html', result=res)

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