set upper limit on gymnasium version

deprl/custom_distributed.py

@@ -52,7 +52,7 @@ def __init__(
         if env_args is not None:
             [x.merge_args(env_args) for x in self.environments]
             [x.apply_args() for x in self.environments]
-        self._max_episode_steps = max_episode_steps
+        self.max_episode_steps = max_episode_steps
         self.observation_space = self.environments[0].observation_space
         self.action_space = self.environments[0].action_space
         self.name = self.environments[0].name
@@ -85,7 +85,7 @@ def step(self, actions):
             muscle = self.environments[i].muscle_states
             self.lengths[i] += 1
             # Timeouts trigger resets but are not true terminations.
-            reset = term or self.lengths[i] == self._max_episode_steps
+            reset = term or self.lengths[i] == self.max_episode_steps
             next_observations.append(ob)
             rewards.append(rew)
             resets.append(reset)
@@ -138,7 +138,7 @@ def __init__(
         self.build_dict = build_dict
         self.worker_groups = worker_groups
         self.workers_per_group = workers_per_group
-        self._max_episode_steps = max_episode_steps
+        self.max_episode_steps = max_episode_steps
         self.env_args = env_args
         self.header = header
 
@@ -171,7 +171,7 @@ def initialize(self, seed):
                 "output_queue": self.output_queue,
                 "group_seed": group_seed,
                 "build_dict": self.build_dict,
-                "max_episode_steps": self._max_episode_steps,
+                "max_episode_steps": self.max_episode_steps,
                 "index": i,
                 "workers": self.workers_per_group,
                 "env_args": self.env_args
@@ -267,7 +267,7 @@ def distribute(
     if "header" in tonic_conf:
         exec(tonic_conf["header"])
     dummy_environment = build_env_from_dict(build_dict)
-    max_episode_steps = dummy_environment._max_episode_steps
+    max_episode_steps = dummy_environment.max_episode_steps
     del dummy_environment
 
     if parallel < 2:

deprl/env_wrappers/dm_wrapper.py

@@ -26,7 +26,7 @@ def muscle_activity(self):
         return self.unwrapped.environment.physics.data.act
 
     @property
-    def _max_episode_steps(self):
+    def max_episode_steps(self):
         return self.unwrapped.max_episode_steps
 
 

deprl/env_wrappers/gym_wrapper.py

@@ -56,7 +56,7 @@ def muscle_activity(self):
         return self.unwrapped.sim.data.act
 
     @property
-    def _max_episode_steps(self):
+    def max_episode_steps(self):
         if hasattr(self.unwrapped, "max_episode_steps"):
             return self.unwrapped.max_episode_steps
         else:

deprl/vendor/tonic/environments/builders.py

@@ -84,12 +84,10 @@ def build_environment(
 
     # Get the default time limit.
     if max_episode_steps == "default":
-        if hasattr(environment, "_max_episode_steps"):
-            max_episode_steps = environment._max_episode_steps
+        if hasattr(environment, "max_episode_steps"):
+            max_episode_steps = environment.max_episode_steps
         elif hasattr(environment, "horizon"):
             max_episode_steps = environment.horizon
-        elif hasattr(environment, "max_episode_steps"):
-            max_episode_steps = environment.max_episode_steps
 
         else:
             logger.log("No max episode steps found, setting them to 1000")

pyproject.toml

@@ -14,7 +14,7 @@ numpy = "^1.22.4"
 termcolor = "^2.2.0"
 pandas = "^2.0.1"
 gdown = "^5.1.0"
-gymnasium = "*"
+gymnasium = "<=0.30"
 wandb = "^0.15.4"
 # torch = {version="2.1.0", source="pytorch-cpu"}
 torch = ">=2.1.0"