API_Python_1.0.0.py

#!/usr/bin/env python
# coding: utf-8

# import statements
from fastapi import FastAPI, HTTPException
import json
import numpy as np
import pickle
import datetime

# Import the airport encodings file
f = open('airport_encodings.json')
 
# returns JSON object as a dictionary
airports = json.load(f)

def create_airport_encoding(airport: str, airports: dict) -> np.array:
    """
    create_airport_encoding is a function that creates an array the length of all arrival airports from the chosen
    departure aiport.  The array consists of all zeros except for the specified arrival airport, which is a 1.  

    Parameters
    ----------
    airport : str
        The specified arrival airport code as a string
    airports: dict
        A dictionary containing all of the arrival airport codes served from the chosen departure airport
        
    Returns
    -------
    np.array
        A NumPy array the length of the number of arrival airports.  All zeros except for a single 1 
        denoting the arrival airport.  Returns None if arrival airport is not found in the input list.
        This is a one-hot encoded airport array.

    """
    temp = np.zeros(len(airports))
    if airport in airports:
        temp[airports.get(airport)] = 1
        temp = temp.T
        return temp
    else:
        return None

# TODO:  write the back-end logic to provide a prediction given the inputs
# requires finalized_model.pkl to be loaded 
# the model must be passed a NumPy array consisting of the following:
# (polynomial order, encoded airport array, departure time as seconds since midnight, arrival time as seconds since midnight)
# the polynomial order is 1 unless you changed it during model training in Task 2
# YOUR CODE GOES HERE

# TODO:  write the API endpoints.  
# YOUR CODE GOES HERE