Compatibility with dm_env and libEGL

digideep/environment/dmc2gym/spec2space.py

@@ -5,7 +5,7 @@
 
 from gym import spaces
 # spaces.Dict | spaces.Tuple | spaces.Box | spaces.Discrete | spaces.MultiDiscrete | spaces.MultiBinary
-from dm_control.rl import specs
+from dm_env import specs
 import numpy as np
 import collections
 # import warnings
@@ -19,22 +19,22 @@ def spec2space_single(spec):
         :obj:`gym.spaces`: The ``gym`` equivalent ``spaces``.
     """
 
-    if (type(spec) is specs.BoundedArraySpec) and (spec.dtype == np.int):
+    if (type(spec) is specs.BoundedArray) and (spec.dtype == np.int):
     # Discrete
         # warnings.warn("The DMC environment uses a discrete action space!")
         if spec.minimum == 0:
             return spaces.Discrete(spec.maximum)
         else:
             raise ValueError("The environment's minimum values must be zero in the Discrete case!")
     # Box
-    elif type(spec) is specs.BoundedArraySpec:
+    elif type(spec) is specs.BoundedArray:
         _min = np.broadcast_to(spec.minimum, shape=spec.shape)
         _max = np.broadcast_to(spec.maximum, shape=spec.shape)
         # if clip_inf:
         #     _min = np.clip(_min, -sys.float_info.max, sys.float_info.max)
         #     _max = np.clip(_max, -sys.float_info.max, sys.float_info.max)
         return spaces.Box(_min, _max, dtype=np.float32)        
-    elif type(spec) is specs.ArraySpec:
+    elif type(spec) is specs.Array:
         return spaces.Box(-np.inf, np.inf, shape=spec.shape, dtype=np.float32)
     else:
         raise ValueError('Unknown spec in spec2space_single!')
@@ -45,7 +45,7 @@ def spec2space(spec):
     Caution:
         Currently it supports ``spaces.Discrete``, ``spaces.Box``, and ``spaces.Dict`` as outputs.
     """
-    if isinstance(spec, specs.ArraySpec) or isinstance(spec, specs.BoundedArraySpec):
+    if isinstance(spec, specs.Array) or isinstance(spec, specs.BoundedArray):
         return spec2space_single(spec)
     elif isinstance(spec, collections.OrderedDict):
         space = collections.OrderedDict()

digideep/environment/dmc2gym/wrapper.py

@@ -14,13 +14,19 @@
 import six
 
 from .spec2space import spec2space
-from .viewer import Viewer
+
+from dm_control._render import BACKEND
+from dm_control._render import constants
+if BACKEND==constants.GLFW:
+    from .viewer import Viewer
+else:
+    Viewer=None
 
 # TODO: flatten_observation does not work with nested dictionaries.
 from dm_control.rl.control import flatten_observation
 # from dm_control.rl.control import FLAT_OBSERVATION_KEY
 from dm_control.rl.control import PhysicsError
-from dm_control.rl import specs
+from dm_env import specs
 
 
 def _spec_from_observation(observation):
@@ -31,7 +37,7 @@ def _spec_from_observation(observation):
         elif isinstance(value, dict):
             raise NotImplementedError("'dict' types in observations are not supported as they may not preserve order. Use OrderedDict instead.")
         else:
-            result[key] = specs.ArraySpec(value.shape, value.dtype, name=key)
+            result[key] = specs.Array(value.shape, value.dtype, name=key)
     return result
 
 
@@ -217,7 +223,10 @@ def _get_viewer(self, mode):
             if mode == "rgb_array":
                 self.viewer = self.dmcenv.physics.render
             elif mode == "human":
-                self.viewer = Viewer(dmcenv=self.dmcenv, width=640, height=480)
+                if Viewer:
+                    self.viewer = Viewer(dmcenv=self.dmcenv, width=640, height=480)
+                else:
+                    self.viewer = None
             self._viewers[mode] = self.viewer
         return self.viewer
 

digideep/pipeline/session.py

@@ -111,8 +111,9 @@ def __init__(self, root_path):
             self.state['path_base_sessions'] = os.path.join(os.path.split(checkpoint_path)[0], "evaluations")
         else:
             # OK, we are loading from a checkpoint, just create session from scratch.
-            self.state['path_root_session']  = self.args["session_path"]
-            self.state['path_base_sessions'] = os.path.join(self.state['path_root_session'], 'digideep_sessions')
+            # self.state['path_root_session']  = self.args["session_path"]
+            # self.state['path_base_sessions'] = os.path.join(self.state['path_root_session'], 'digideep_sessions')
+            self.state['path_base_sessions'] = self.args["session_path"]
 
 
         # 1. Creating 'path_base_sessions', i.e. '/tmp/digideep_sessions':
@@ -334,7 +335,7 @@ def parse_arguments(self):
         parser.add_argument('--play', action="store_true", help="Will play the stored policy.")
         parser.add_argument('--dry-run', action="store_true", help="If used no footprints will be stored on disc whatsoever.")
         ## Session
-        parser.add_argument('--session-path', metavar=('<path>'), default='/tmp', type=str, help="The path to store the sessions. Default is in /tmp")
+        parser.add_argument('--session-path', metavar=('<path>'), default='/tmp/digideep_sessions', type=str, help="The path to store the sessions. Default is in /tmp")
         parser.add_argument('--session-name', metavar=('<name>'), default='',     type=str, help="A default name for the session. Random name if not provided.")
         parser.add_argument('--save-modules', metavar=('<path>'), default=[], nargs='+', type=str, help="The modules to be stored in the session.")
         parser.add_argument('--log-level', metavar=('<n>'), default=1, type=int, help="The logging level: 0 (debug and above), 1 (info and above), 2 (warn and above), 3 (error and above), 4 (fatal and above)")