@@ -59,7 +59,7 @@ def discrete_environment_example():
5959 config ["repeats_in_sequences" ] = False
6060
6161 config ["generate_random_mdp" ] = True
62- env = RLToyEnv (** config ) # Calls env.reset() automatically. So, in general,
62+ env = RLToyEnv (** config ) # Calls env.reset()[0] automatically. So, in general,
6363 # there is no need to call it after this.
6464
6565 # The environment maintains an augmented state which contains the underlying
@@ -73,7 +73,7 @@ def discrete_environment_example():
7373 "the transition:"
7474 )
7575 action = env .action_space .sample ()
76- next_state , reward , done , info = env .step (action )
76+ next_state , reward , done , trunc , info = env .step (action )
7777 print ("sars', done =" , state , action , reward , next_state , done )
7878
7979 env .close ()
@@ -113,7 +113,7 @@ def discrete_environment_image_representations_example():
113113 "the transition:"
114114 )
115115 action = env .action_space .sample ()
116- next_state_image , reward , done , info = env .step (action )
116+ next_state_image , reward , done , trunc , info = env .step (action )
117117 augmented_state_dict = env .get_augmented_state ()
118118 next_state = augmented_state_dict ["curr_state" ] # Underlying MDP state holds
119119 # the current discrete state.
@@ -159,7 +159,7 @@ def discrete_environment_diameter_image_representations_example():
159159 "the transition:"
160160 )
161161 action = env .action_space .sample ()
162- next_state_image , reward , done , info = env .step (action )
162+ next_state_image , reward , done , trunc , info = env .step (action )
163163 augmented_state_dict = env .get_augmented_state ()
164164 next_state = augmented_state_dict ["curr_state" ] # Underlying MDP state holds
165165 # the current discrete state.
@@ -192,14 +192,14 @@ def continuous_environment_example_move_to_a_point():
192192 config ["reward_function" ] = "move_to_a_point"
193193
194194 env = RLToyEnv (** config )
195- state = env .reset ().copy ()
195+ state = env .reset ()[ 0 ] .copy ()
196196
197197 print (
198198 "Taking a step in the environment with a random action and printing "
199199 "the transition:"
200200 )
201201 action = env .action_space .sample ()
202- next_state , reward , done , info = env .step (action )
202+ next_state , reward , done , trunc , info = env .step (action )
203203 print ("sars', done =" , state , action , reward , next_state , done )
204204
205205 env .close ()
@@ -231,7 +231,7 @@ def continuous_environment_example_move_to_a_point_irrelevant_image():
231231 config ["relevant_indices" ] = [0 , 1 ]
232232
233233 env = RLToyEnv (** config )
234- state = env .reset ()
234+ state = env .reset ()[ 0 ]
235235 augmented_state_dict = env .get_augmented_state ()
236236 state = augmented_state_dict ["curr_state" ].copy () # Underlying MDP state holds
237237 # the current continuous state.
@@ -241,7 +241,7 @@ def continuous_environment_example_move_to_a_point_irrelevant_image():
241241 "the transition:"
242242 )
243243 action = env .action_space .sample ()
244- next_state_image , reward , done , info = env .step (action )
244+ next_state_image , reward , done , trunc , info = env .step (action )
245245 augmented_state_dict = env .get_augmented_state ()
246246 next_state = augmented_state_dict ["curr_state" ].copy () # Underlying MDP state holds
247247 # the current continuous state.
@@ -274,14 +274,14 @@ def continuous_environment_example_move_along_a_line():
274274 config ["reward_function" ] = "move_along_a_line"
275275
276276 env = RLToyEnv (** config )
277- state = env .reset ().copy ()
277+ state = env .reset ()[ 0 ] .copy ()
278278
279279 print (
280280 "Taking a step in the environment with a random action and printing "
281281 "the transition:"
282282 )
283283 action = env .action_space .sample ()
284- next_state , reward , done , info = env .step (action )
284+ next_state , reward , done , trunc , info = env .step (action )
285285 print ("sars', done =" , state , action , reward , next_state , done )
286286
287287 env .close ()
@@ -305,12 +305,12 @@ def grid_environment_example():
305305
306306 for i in range (len (actions )):
307307 action = actions [i ]
308- next_obs , reward , done , info = env .step (action )
308+ next_obs , reward , done , trunc , info = env .step (action )
309309 next_state = env .get_augmented_state ()["augmented_state" ][- 1 ]
310310 print ("sars', done =" , state , action , reward , next_state , done )
311311 state = next_state
312312
313- env .reset ()
313+ env .reset ()[ 0 ]
314314 env .close ()
315315
316316
@@ -334,12 +334,12 @@ def grid_environment_image_representations_example():
334334
335335 for i in range (len (actions )):
336336 action = actions [i ]
337- next_obs , reward , done , info = env .step (action )
337+ next_obs , reward , done , trunc , info = env .step (action )
338338 next_state = env .get_augmented_state ()["augmented_state" ][- 1 ]
339339 print ("sars', done =" , state , action , reward , next_state , done )
340340 state = next_state
341341
342- env .reset ()
342+ env .reset ()[ 0 ]
343343 env .close ()
344344
345345 display_image (next_obs )
@@ -356,18 +356,18 @@ def atari_wrapper_example():
356356 }
357357
358358 from mdp_playground .envs import GymEnvWrapper
359- import gym
359+ import gymnasium as gym
360360
361361 ae = gym .make ("QbertNoFrameskip-v4" )
362362 env = GymEnvWrapper (ae , ** config )
363- state = env .reset ()
363+ state = env .reset ()[ 0 ]
364364
365365 print (
366366 "Taking 10 steps in the environment with a random action and printing the transition:"
367367 )
368368 for i in range (10 ):
369369 action = env .action_space .sample ()
370- next_state , reward , done , info = env .step (action )
370+ next_state , reward , done , trunc , info = env .step (action )
371371 print (
372372 "s.shape a r s'.shape, done =" ,
373373 state .shape ,
@@ -403,18 +403,18 @@ def mujoco_wrapper_example():
403403 # of the Mujoco base_class.
404404 try :
405405 from mdp_playground .envs import get_mujoco_wrapper
406- from gym .envs .mujoco .half_cheetah_v3 import HalfCheetahEnv
406+ from gymnasium .envs .mujoco .half_cheetah_v3 import HalfCheetahEnv
407407
408408 wrapped_mujoco_env = get_mujoco_wrapper (HalfCheetahEnv )
409409
410410 env = wrapped_mujoco_env (** config )
411- state = env .reset ()
411+ state = env .reset ()[ 0 ]
412412
413413 print (
414414 "Taking a step in the environment with a random action and printing the transition:"
415415 )
416416 action = env .action_space .sample ()
417- next_state , reward , done , info = env .step (action )
417+ next_state , reward , done , trunc , info = env .step (action )
418418 print ("sars', done =" , state , action , reward , next_state , done )
419419
420420 env .close ()
@@ -440,22 +440,22 @@ def minigrid_wrapper_example():
440440 }
441441
442442 from mdp_playground .envs .gym_env_wrapper import GymEnvWrapper
443- import gym
443+ import gymnasium as gym
444444
445- from gym_minigrid .wrappers import RGBImgPartialObsWrapper , ImgObsWrapper
445+ from minigrid .wrappers import RGBImgPartialObsWrapper , ImgObsWrapper
446446
447447 env = gym .make ("MiniGrid-Empty-8x8-v0" )
448448 env = RGBImgPartialObsWrapper (env ) # Get pixel observations
449449 env = ImgObsWrapper (env ) # Get rid of the 'mission' field
450450
451451 env = GymEnvWrapper (env , ** config )
452- obs = env .reset () # This now produces an RGB tensor only
452+ obs = env .reset ()[ 0 ] # This now produces an RGB tensor only
453453
454454 print (
455455 "Taking a step in the environment with a random action and printing the transition:"
456456 )
457457 action = env .action_space .sample ()
458- next_obs , reward , done , info = env .step (action )
458+ next_obs , reward , done , trunc , info = env .step (action )
459459 print (
460460 "s.shape ar s'.shape, done =" ,
461461 obs .shape ,
@@ -481,17 +481,17 @@ def procgen_wrapper_example():
481481 }
482482
483483 from mdp_playground .envs .gym_env_wrapper import GymEnvWrapper
484- import gym
484+ import gymnasium as gym
485485
486486 env = gym .make ("procgen:procgen-coinrun-v0" )
487487 env = GymEnvWrapper (env , ** config )
488- obs = env .reset ()
488+ obs = env .reset ()[ 0 ]
489489
490490 print (
491491 "Taking a step in the environment with a random action and printing the transition:"
492492 )
493493 action = env .action_space .sample ()
494- next_obs , reward , done , info = env .step (action )
494+ next_obs , reward , done , trunc , info = env .step (action )
495495 print (
496496 "s.shape ar s'.shape, done =" ,
497497 obs .shape ,
@@ -577,7 +577,7 @@ def procgen_wrapper_example():
577577
578578 # Using gym.make() example 1
579579 import mdp_playground
580- import gym
580+ import gymnasium as gym
581581
582582 gym .make ("RLToy-v0" )
583583
@@ -591,6 +591,6 @@ def procgen_wrapper_example():
591591 "maximally_connected" : True ,
592592 }
593593 )
594- env .reset ()
594+ env .reset ()[ 0 ]
595595 for i in range (10 ):
596596 print (env .step (env .action_space .sample ()))
0 commit comments