Init gym env with list of patients/meals

[Shurikal]

I added the possibility to create a gym env with multiple patients and meals. Every time the env is reseted, a new patient/meal is randomly sampled. I also changed the _create_env_from_random_state(custom_scenario) name and parameter to _create_env() because the custom_scenario is an instance variable.

from datetime import datetime
import gym
import simglucose
from simglucose.simulation.scenario import CustomScenario
import numpy as np

start_time = datetime(2018, 1, 1, 0, 0, 0)
meal_scenario_1 = CustomScenario(start_time=start_time, scenario=[(1,20)])
meal_scenario_2 = CustomScenario(start_time=start_time,scenario=[(3,15)])

patient_name = ['adult#001','adult#002','adult#003','adult#004','adult#005','adult#006','adult#007','adult#008','adult#009','adult#010']

gym.envs.register(
    id='env-v0',
    entry_point="simglucose.envs:T1DSimEnv",
    kwargs={'patient_name': patient_name,
           'custom_scenario': [meal_scenario_1, meal_scenario_2]}
)

env = gym.make('env-v0')

env.reset()

min_insulin = env.action_space.low
max_insulin = env.action_space.high

observation = env.reset()
for t in range(100):
    env.render(mode='human')
    #action = np.random.uniform(min_insulin, max_insulin)

    action = observation.CGM * 0.0005
    if observation.CGM < 120:
        action = 0

    #print(action)
    observation, reward, done, info = env.step(action)
    if done:
        print("Episode finished after {} timesteps".format(t + 1))
        break

[jxx123]

Thank you so much! This is some feature that I want for a while.