app.py

import os, pickle
import logging
from flask import Flask, jsonify, render_template, request

app = Flask(__name__)


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

@app.route('/api/predict', methods=['POST'])
def predict():
  try:
    # load and parse input
    data = request.json
    vector = [
        float(data['age']),
        data['travel'],
        data['department'],
        float(data['distance']),
        float(data['education']),
        data['gender'],
        float(data['satisfaction']),
        data['maritalstatus'],
        float(data['income']),
        data['overtime'],
        float(data['totalyears']),
        float(data['years']),
        float(data['lastpromotion'])
    ]
    # app.logging.info(f'vector: {vector}')
    # print(f'vector: {vector}\n')

    # load the model
    with open(os.path.join('data', 'logistic.pkcls'), 'rb') as file:
      model = pickle.load(file)

    predictions = model(vector, 1)
    # app.logging.info(f'predictions: {predictions}')
    # print(f'predictions: {predictions}\n')

    # send the response
    return jsonify({ "predictions": { "leave": predictions[0], "stay": predictions[1] } })
  except Exception as e:
    return jsonify({"error": repr(e)})


if __name__ == "__main__":
    app.logger.setLevel(logging.INFO)
    app.run(host="0.0.0.0", port=7860)