support openAI gym.make

.gitignore

@@ -1,6 +1,9 @@
 # Compiled python modules.
 *.pyc
 
+# results foler
+/results/
+
 # Setuptools distribution folder.
 /dist/
 

setup.py

@@ -2,8 +2,8 @@
 
 
 setup(name='simglucose',
-      version='0.1.4',
-      description='A Type-1 Diabetes Simulator as a Reinforcement Learning Environment',
+      version='0.1.5',
+      description='A Type-1 Diabetes Simulator as a Reinforcement Learning Environment in OpenAI gym or rllab (python implementation of UVa/Padova Simulator)',
       url='https://github.com/jxx123/simglucose',
       author='Jinyu Xie',
       author_email='[email protected]',
@@ -14,7 +14,8 @@
           'numpy',
           'scipy',
           'matplotlib',
-          'pathos'
+          'pathos',
+          'gym'
       ],
       include_package_data=True,
       zip_safe=False)

simglucose/__init__.py

@@ -0,0 +1,6 @@
+from gym.envs.registration import register
+
+register(
+    id='simglucose-v0',
+    entry_point='simglucose.envs:T1DSimEnv',
+)

simglucose/actuator/pump.py

@@ -1,6 +1,7 @@
 import pandas as pd
 import pkg_resources
 import logging
+import numpy as np
 
 INSULIN_PUMP_PARA_FILE = pkg_resources.resource_filename(
     'simglucose', 'params/pump_params.csv')
@@ -21,17 +22,17 @@ def withName(cls, name):
 
     def bolus(self, amount):
         bol = amount * self.U2PMOL  # convert from U/min to pmol/min
-        bol = round(bol / self._params['inc_bolus']
-                    ) * self._params['inc_bolus']
+        bol = np.round(bol / self._params['inc_bolus']
+                       ) * self._params['inc_bolus']
         bol = bol / self.U2PMOL     # convert from pmol/min to U/min
         bol = min(bol, self._params['max_bolus'])
         bol = max(bol, self._params['min_bolus'])
         return bol
 
     def basal(self, amount):
         bas = amount * self.U2PMOL  # convert from U/min to pmol/min
-        bas = round(bas / self._params['inc_basal']
-                    ) * self._params['inc_basal']
+        bas = np.round(bas / self._params['inc_basal']
+                       ) * self._params['inc_basal']
         bas = bas / self.U2PMOL     # convert from pmol/min to U/min
         bas = min(bas, self._params['max_basal'])
         bas = max(bas, self._params['min_basal'])

simglucose/controller/basal_bolus_ctrller.py

@@ -22,10 +22,11 @@ def __init__(self, target=140):
     def policy(self, observation, reward, done, **kwargs):
         sample_time = kwargs.get('sample_time', 1)
         pname = kwargs.get('patient_name')
+        meal = kwargs.get('meal')
 
         action = self._bb_policy(
             pname,
-            observation.CHO,
+            meal,
             observation.CGM,
             sample_time)
         return action

simglucose/envs/__init__.py

@@ -0,0 +1 @@
+from simglucose.envs.simglucose_gym_env import T1DSimEnv

simglucose/envs/simglucose_gym_env.py

@@ -0,0 +1,89 @@
+from simglucose.simulation.env import T1DSimEnv as _T1DSimEnv
+from simglucose.patient.t1dpatient import T1DPatient
+from simglucose.sensor.cgm import CGMSensor
+from simglucose.actuator.pump import InsulinPump
+from simglucose.simulation.scenario_gen import RandomScenario
+from simglucose.controller.base import Action
+import pandas as pd
+import numpy as np
+import pkg_resources
+import gym
+from gym import error, spaces, utils
+from gym.utils import seeding
+from datetime import datetime
+
+PATIENT_PARA_FILE = pkg_resources.resource_filename(
+    'simglucose', 'params/vpatient_params.csv')
+
+
+class T1DSimEnv(gym.Env):
+    '''
+    A wrapper of simglucose.simulation.env.T1DSimEnv to support gym API
+    '''
+    metadata = {'render.modes': ['human']}
+
+    def __init__(self, patient_name=None):
+        '''
+        patient_name must be 'adolescent#001' to 'adolescent#010',
+        or 'adult#001' to 'adult#010', or 'child#001' to 'child#010'
+        '''
+        if patient_name is None:
+            # patient_name = self.pick_patient()
+
+            # have to hard code the patient_name, gym has some interesting
+            # error when choosing the patient
+            patient_name = 'adolescent#001'
+        patient = T1DPatient.withName(patient_name)
+        sensor = CGMSensor.withName('Dexcom')
+        scenario = RandomScenario(start_time=datetime(2018, 1, 1, 0, 0, 0))
+        pump = InsulinPump.withName('Insulet')
+        self.env = _T1DSimEnv(patient, sensor, pump, scenario)
+
+    @staticmethod
+    def pick_patient():
+        patient_params = pd.read_csv(PATIENT_PARA_FILE)
+        while True:
+            print('Select patient:')
+            for j in range(len(patient_params)):
+                print('[{0}] {1}'.format(j + 1, patient_params['Name'][j]))
+            try:
+                select = int(input('>>> '))
+            except ValueError:
+                print('Please input a number.')
+                continue
+
+            if select < 1 or select > len(patient_params):
+                print('Please input 1 to {}'.format(len(patient_params)))
+                continue
+
+            return select
+
+    def _step(self, action):
+        # This gym only controls basal insulin
+        act = Action(basal=action, bolus=0)
+        return self.env.step(act)
+
+    def _reset(self):
+        return self.env.reset()
+
+    def _seed(self, seed=None):
+        rng, seed1 = seeding.np_random(seed=seed)
+        # Derive a random seed. This gets passed as a uint, but gets
+        # checked as an int elsewhere, so we need to keep it below
+        # 2**31.
+        seed2 = seeding.hash_seed(seed1 + 1) % 2**31
+        self.env.sensor.seed = seed1
+        self.env.scenario.seed = seed2
+        return [seed1, seed2]
+
+    def _render(self, mode='human', close=False):
+        self.env.render(close=close)
+
+    @property
+    def action_space(self):
+        ub = self.env.pump._params['max_basal']
+        return spaces.Box(low=0, high=ub, shape=(1,))
+
+    @property
+    def observation_space(self):
+        return spaces.Box(low=0, high=np.inf, shape=(1,))

simglucose/sensor/cgm.py

@@ -15,7 +15,7 @@ def __init__(self, params, seed=None):
         self.name = params.Name
         self.sample_time = params.sample_time
         self.seed = seed
-        self.reset()
+        self._last_CGM = 0
 
     @classmethod
     def withName(cls, name, **kwargs):
@@ -37,6 +37,15 @@ def measure(self, patient):
         # Zero-Order Hold
         return self._last_CGM
 
+    @property
+    def seed(self):
+        return self._seed
+
+    @seed.setter
+    def seed(self, seed):
+        self._seed = seed
+        self._noise_generator = CGMNoise(self._params, seed=seed)
+
     def reset(self):
         logger.debug('Resetting CGM sensor ...')
         self._noise_generator = CGMNoise(self._params, seed=self.seed)

simglucose/simulation/env.py

@@ -5,6 +5,7 @@
 from datetime import timedelta
 import logging
 from collections import namedtuple
+from simglucose.simulation.rendering import Viewer
 
 rllab = True
 try:
@@ -36,7 +37,7 @@ def step(self, action):
         def reset(self):
             raise NotImplementedError
 
-Observation = namedtuple('Observation', ['CHO', 'CGM'])
+Observation = namedtuple('Observation', ['CGM'])
 logger = logging.getLogger(__name__)
 
 
@@ -50,21 +51,7 @@ def __init__(self,
         self.sensor = sensor
         self.pump = pump
         self.scenario = scenario
-        self.sample_time = self.sensor.sample_time
-
-        # Initial Recording
-        BG = self.patient.observation.Gsub
-        horizon = 0
-        LBGI, HBGI, risk = risk_index([BG], horizon)
-        CGM = self.sensor.measure(self.patient)
-        self.time_hist = [self.scenario.start_time]
-        self.BG_hist = [BG]
-        self.CGM_hist = [CGM]
-        self.risk_hist = [risk]
-        self.LBGI_hist = [LBGI]
-        self.HBGI_hist = [HBGI]
-        self.CHO_hist = []
-        self.insulin_hist = []
+        self._reset()
 
     @property
     def time(self):
@@ -122,25 +109,22 @@ def step(self, action):
         self.HBGI_hist.append(HBGI)
 
         # Compute reward, and decide whether game is over
-        # reward = - np.log(risk)
-        # reward = 10 - risk
         if len(self.risk_hist) > 1:
             reward = self.risk_hist[-2] - self.risk_hist[-1]
         else:
             reward = - self.risk_hist[-1]
         done = BG < 70 or BG > 350
-        obs = Observation(CHO=CHO, CGM=CGM)
+        obs = Observation(CGM=CGM)
         return Step(observation=obs,
                     reward=reward,
                     done=done,
                     sample_time=self.sample_time,
-                    patient_name=self.patient.name)
+                    patient_name=self.patient.name,
+                    meal=CHO)
 
-    def reset(self):
-        self.patient.reset()
-        self.sensor.reset()
-        self.pump.reset()
-        self.scenario.reset()
+    def _reset(self):
+        self.sample_time = self.sensor.sample_time
+        self.viewer = None
 
         BG = self.patient.observation.Gsub
         horizon = 0
@@ -155,12 +139,27 @@ def reset(self):
         self.CHO_hist = []
         self.insulin_hist = []
 
+    def reset(self):
+        self.patient.reset()
+        self.sensor.reset()
+        self.pump.reset()
+        self.scenario.reset()
+        self._reset()
+        CGM = self.sensor.measure(self.patient)
+        obs = Observation(CGM=CGM)
+        return Step(observation=obs,
+                    reward=0,
+                    done=False,
+                    sample_time=self.sample_time,
+                    patient_name=self.patient.name,
+                    meal=0)
+
     @property
     def action_space(self):
         if rllab:
-            ub = self.pump._params['max_basal'] + \
-                self.pump._params['max_bolus']
-            return Box(low=0, high=ub, shape=(1,))
+            ub = np.array([self.pump._params['max_basal'],
+                           self.pump._params['max_bolus']])
+            return Box(low=np.array([0, 0]), high=ub)
         else:
             pass
 
@@ -171,51 +170,17 @@ def observation_space(self):
         else:
             pass
 
-    def render(self, axes, lines):
-        logger.info('Rendering ...')
-
-        lines[0].set_xdata(self.time_hist)
-        lines[0].set_ydata(self.BG_hist)
-
-        lines[1].set_xdata(self.time_hist)
-        lines[1].set_ydata(self.CGM_hist)
-
-        axes[0].draw_artist(axes[0].patch)
-        axes[0].draw_artist(lines[0])
-        axes[0].draw_artist(lines[1])
-
-        adjust_ylim(axes[0], min(min(self.BG_hist), min(self.CGM_hist)), max(
-            max(self.BG_hist), max(self.CGM_hist)))
-
-        lines[2].set_xdata(self.time_hist[:-1])
-        lines[2].set_ydata(self.CHO_hist)
-
-        axes[1].draw_artist(axes[1].patch)
-        axes[1].draw_artist(lines[2])
-
-        adjust_ylim(axes[1], min(self.CHO_hist), max(self.CHO_hist))
-
-        lines[3].set_xdata(self.time_hist[:-1])
-        lines[3].set_ydata(self.insulin_hist)
-
-        axes[2].draw_artist(axes[2].patch)
-        axes[2].draw_artist(lines[3])
-        adjust_ylim(axes[2], min(self.insulin_hist), max(self.insulin_hist))
-
-        lines[4].set_xdata(self.time_hist)
-        lines[4].set_ydata(self.LBGI_hist)
-
-        lines[5].set_xdata(self.time_hist)
-        lines[5].set_ydata(self.HBGI_hist)
+    def render(self, close=False):
+        if close:
+            if self.viewer is not None:
+                self.viewer.close()
+                self.viewer = None
+            return
 
-        lines[6].set_xdata(self.time_hist)
-        lines[6].set_ydata(self.risk_hist)
+        if self.viewer is None:
+            self.viewer = Viewer(self.scenario.start_time, self.patient.name)
 
-        axes[3].draw_artist(axes[3].patch)
-        axes[3].draw_artist(lines[4])
-        axes[3].draw_artist(lines[5])
-        axes[3].draw_artist(lines[6])
-        adjust_ylim(axes[3], min(self.risk_hist), max(self.risk_hist))
+        self.viewer.render(self.show_history())
 
     def show_history(self):
         df = pd.DataFrame()