server.py

# import the nessecary pieces from Flask
from flask import Flask,render_template, request,jsonify,Response
import pickle
import pandas as pd

#Create the app object that will route our calls
app = Flask(__name__)
# Add a single endpoint that we can use for testing
@app.route('/', methods = ['GET'])
def home():
    return render_template('home.html')

@app.route('/mpg', methods = ['GET'])
def mpg():
    return render_template('mpg.html')

model = pickle.load(open('linreg.p','rb'))

@app.route('/inference',  methods = ['POST'])
def inference():
    req = request.get_json()
    print(req)
    c,h,w = req['cylinders'],req['horsepower'],req['weight']
    prediction = list(model.predict([[c,h,w]]))
    return jsonify({'c':c,'h': h,'w':w,'prediction':prediction[0] })

@app.route('/plot',  methods = ['GET'])
def plot():
    df = pd.read_csv('cars.csv')
    data = list(zip(df.mpg,df.weight))
    return jsonify(data)


#When run from command line, start the server
if __name__ == '__main__':
    app.run(host ='0.0.0.0', port = 3333, debug = True)