app.py

from environment import grid, print_grid
from value_iteration import value_iteration
from q_learning import q_learning

from flask import Flask, request, render_template
import json

app = Flask(__name__)
env = grid()


@app.route('/valueiteration')
def dp():
    policy, V, stats = value_iteration(env, (1, 0), (3, 3))
    print_grid(env, policy)
    print_grid(env, V)
    return render_template('value_iteration.html', data=json.dumps(stats))


@app.route('/qlearning', methods=['GET'])
def td():
    episodes = int(request.args.get('episodes', 300))
    epsilon = float(request.args.get('epsilon', 0.05))
    alpha = float(request.args.get('alpha', 0.5))
    policy, V, stats = q_learning(
        env, episodes, epsilon, alpha, (1, 0), (3, 3))
    print_grid(env, policy)
    print_grid(env, V)
    # omit first 10 episodes as they are very long for visualization
    return render_template('q_learning.html', data=json.dumps(stats[10:]))


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