refactor: enhance making of safe isaac gym (#117)

.pre-commit-config.yaml

@@ -8,7 +8,7 @@ ci:
 default_stages: [commit, push, manual]
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.4.0
+    rev: v4.6.0
     hooks:
       - id: check-symlinks
       - id: destroyed-symlinks
@@ -45,7 +45,7 @@ repos:
           )
       - id: double-quote-string-fixer
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.0.285
+    rev: v0.4.1
     hooks:
       - id: ruff
         args: [--fix, --exit-non-zero-on-fix]
@@ -54,15 +54,15 @@ repos:
             ^safety_gymnasium/tasks/safe_isaac_gym/
           )
   - repo: https://github.com/PyCQA/isort
-    rev: 5.12.0
+    rev: 5.13.2
     hooks:
       - id: isort
   - repo: https://github.com/psf/black
-    rev: 23.7.0
+    rev: 24.4.0
     hooks:
       - id: black-jupyter
   - repo: https://github.com/asottile/pyupgrade
-    rev: v3.10.1
+    rev: v3.15.2
     hooks:
       - id: pyupgrade
         args: [--py38-plus] # sync with requires-python
@@ -71,7 +71,7 @@ repos:
             ^examples/
           )
   - repo: https://github.com/pycqa/flake8
-    rev: 6.1.0
+    rev: 7.0.0
     hooks:
       - id: flake8
         additional_dependencies:
@@ -88,7 +88,7 @@ repos:
             ^docs/conf.py$
           )
   - repo: https://github.com/codespell-project/codespell
-    rev: v2.2.5
+    rev: v2.2.6
     hooks:
       - id: codespell
         additional_dependencies: [".[toml]"]

docs/environments/safe_multi_agent/multi_goal.rst

@@ -39,7 +39,7 @@ Agents are mandated to approach and reach the targets that **match their own col
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _MultiGoal0:
 

docs/environments/safe_navigation/button.rst

@@ -41,7 +41,7 @@ Specific Setting
 Episode End
 ------------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _Button0:
 

docs/environments/safe_navigation/circle.rst

@@ -36,7 +36,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 500: ``Trucated = True``.
+- When episode length is greater than 500: ``Truncated = True``.
 
 .. _Circle0:
 

docs/environments/safe_navigation/goal.rst

@@ -36,7 +36,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _Goal0:
 

docs/environments/safe_navigation/push.rst

@@ -53,7 +53,7 @@ Specific Setting
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated == True``.
+- When episode length is greater than 1000: ``Truncated == True``.
 
 .. _Push0:
 

docs/environments/safe_vision/building_button.rst

@@ -44,7 +44,7 @@ Specific Setting
 Episode End
 ------------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _BuildingButton0:
 

docs/environments/safe_vision/building_goal.rst

@@ -38,7 +38,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _BuildingGoal0:
 

docs/environments/safe_vision/building_push.rst

@@ -57,7 +57,7 @@ Specific Setting
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _BuildingPush0:
 

docs/environments/safe_vision/fading_easy.rst

@@ -48,7 +48,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _FadingEasy0:
 

docs/environments/safe_vision/fading_hard.rst

@@ -49,7 +49,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _FadingHard0:
 

docs/environments/safe_vision/formula_one.rst

@@ -35,7 +35,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 1000: ``Trucated = True``.
+- When episode length is greater than 1000: ``Truncated = True``.
 
 .. _FormulaOne0:
 

docs/environments/safe_vision/race.rst

@@ -35,7 +35,7 @@ Rewards
 Episode End
 -----------
 
-- When episode length is greater than 500: ``Trucated = True``.
+- When episode length is greater than 500: ``Truncated = True``.
 - Upon the agent's arrival at the "Goal": ``goal_achieved = True``.
 
 .. _Race0:

safety_gymnasium/bases/base_task.py

@@ -376,9 +376,9 @@ def build_staged_goal_position(self) -> None:
             self.staged_goal.get_next_goal_xy(),  # pylint: disable=no-member
         )
         # Move goal geom to new layout position
-        self.world_info.world_config_dict['geoms']['staged_goal']['pos'][
-            :2
-        ] = self.world_info.layout['staged_goal']
+        self.world_info.world_config_dict['geoms']['staged_goal']['pos'][:2] = (
+            self.world_info.layout['staged_goal']
+        )
         self._set_goal(self.world_info.layout['staged_goal'], 'staged_goal')
         mujoco.mj_forward(self.model, self.data)  # pylint: disable=no-member
 

safety_gymnasium/tasks/safe_isaac_gym/envs/cfgs/shadow_hand_catch_over_2_underarm_safe_finger.yaml → safety_gymnasium/tasks/safe_isaac_gym/envs/cfgs/ShadowHandCatchOver2underarm_Safe_finger.yaml

@@ -22,7 +22,7 @@ env:
   resetDofPosRandomInterval: 0.2
   resetDofVelRandomInterval: 0.0
 
-  distRewardScale: 50.0
+  distRewardScale: 50
   transition_scale: 0.05
   orientation_scale: 0.5
 

safety_gymnasium/tasks/safe_isaac_gym/envs/tasks/ShadowHandCatchOver2underarm_Safe_finger.py

@@ -10,8 +10,8 @@
 
 import numpy as np
 import torch
-from isaacgym import gymapi, gymtorch
-from matplotlib.pyplot import axis
+import yaml
+from isaacgym import gymapi, gymtorch, gymutil
 
 from safety_gymnasium.tasks.safe_isaac_gym.envs.tasks.hand_base.base_task import BaseTask
 from safety_gymnasium.tasks.safe_isaac_gym.utils.torch_jit_utils import *
@@ -20,16 +20,30 @@
 class ShadowHandCatchOver2Underarm_Safe_finger(BaseTask):
     def __init__(
         self,
-        cfg,
         sim_params,
         physics_engine,
         device_type,
         device_id,
         headless,
+        num_envs=None,
+        cfg=None,
         agent_index=[[[0, 1, 2, 3, 4, 5]], [[0, 1, 2, 3, 4, 5]]],
         is_multi_agent=False,
     ):
-        self.cfg = cfg
+        yaml_path = (
+            os.path.abspath(__file__).replace('envs/tasks', 'envs/cfgs').replace('.py', '.yaml')
+        )
+        with open(yaml_path) as f:
+            self.cfg = yaml.load(f, Loader=yaml.FullLoader)
+        if cfg:
+            self.cfg.update(cfg)
+
+        if num_envs:
+            self.cfg['env']['numEnvs'] = num_envs
+
+        if 'sim' in self.cfg:
+            gymutil.parse_sim_config(self.cfg['sim'], sim_params)
+
         self.sim_params = sim_params
         self.physics_engine = physics_engine
         self.agent_index = agent_index
@@ -743,9 +757,9 @@ def compute_full_state(self, asymm_obs=False):
         )
 
         fingertip_obs_start = 72  # 168 = 157 + 11
-        self.obs_buf[
-            :, fingertip_obs_start : fingertip_obs_start + num_ft_states
-        ] = self.fingertip_state.reshape(self.num_envs, num_ft_states)
+        self.obs_buf[:, fingertip_obs_start : fingertip_obs_start + num_ft_states] = (
+            self.fingertip_state.reshape(self.num_envs, num_ft_states)
+        )
         self.obs_buf[
             :,
             fingertip_obs_start
@@ -775,12 +789,12 @@ def compute_full_state(self, asymm_obs=False):
 
         # another_hand
         another_hand_start = action_obs_start + 26
-        self.obs_buf[
-            :, another_hand_start : self.num_shadow_hand_dofs + another_hand_start
-        ] = unscale(
-            self.shadow_hand_another_dof_pos,
-            self.shadow_hand_dof_lower_limits,
-            self.shadow_hand_dof_upper_limits,
+        self.obs_buf[:, another_hand_start : self.num_shadow_hand_dofs + another_hand_start] = (
+            unscale(
+                self.shadow_hand_another_dof_pos,
+                self.shadow_hand_dof_lower_limits,
+                self.shadow_hand_dof_upper_limits,
+            )
         )
         self.obs_buf[
             :,
@@ -814,9 +828,9 @@ def compute_full_state(self, asymm_obs=False):
         )
 
         hand_another_pose_start = fingertip_another_obs_start + 95
-        self.obs_buf[
-            :, hand_another_pose_start : hand_another_pose_start + 3
-        ] = self.hand_positions[self.another_hand_indices, :]
+        self.obs_buf[:, hand_another_pose_start : hand_another_pose_start + 3] = (
+            self.hand_positions[self.another_hand_indices, :]
+        )
         self.obs_buf[:, hand_another_pose_start + 3 : hand_another_pose_start + 4] = get_euler_xyz(
             self.hand_orientations[self.another_hand_indices, :]
         )[0].unsqueeze(-1)

safety_gymnasium/tasks/safe_isaac_gym/envs/tasks/ShadowHandCatchOver2underarm_Safe_joint.py

@@ -10,8 +10,8 @@
 
 import numpy as np
 import torch
-from isaacgym import gymapi, gymtorch
-from matplotlib.pyplot import axis
+import yaml
+from isaacgym import gymapi, gymtorch, gymutil
 
 from safety_gymnasium.tasks.safe_isaac_gym.envs.tasks.hand_base.base_task import BaseTask
 from safety_gymnasium.tasks.safe_isaac_gym.utils.torch_jit_utils import *
@@ -20,16 +20,30 @@
 class ShadowHandCatchOver2Underarm_Safe_joint(BaseTask):
     def __init__(
         self,
-        cfg,
         sim_params,
         physics_engine,
         device_type,
         device_id,
         headless,
+        num_envs=None,
+        cfg=None,
         agent_index=[[[0, 1, 2, 3, 4, 5]], [[0, 1, 2, 3, 4, 5]]],
         is_multi_agent=False,
     ):
-        self.cfg = cfg
+        yaml_path = (
+            os.path.abspath(__file__).replace('envs/tasks', 'envs/cfgs').replace('.py', '.yaml')
+        )
+        with open(yaml_path) as f:
+            self.cfg = yaml.load(f, Loader=yaml.FullLoader)
+        if cfg:
+            self.cfg.update(cfg)
+
+        if num_envs:
+            self.cfg['env']['numEnvs'] = num_envs
+
+        if 'sim' in self.cfg:
+            gymutil.parse_sim_config(self.cfg['sim'], sim_params)
+
         self.sim_params = sim_params
         self.physics_engine = physics_engine
         self.agent_index = agent_index
@@ -731,9 +745,9 @@ def compute_full_state(self, asymm_obs=False):
         )
 
         fingertip_obs_start = 72  # 168 = 157 + 11
-        self.obs_buf[
-            :, fingertip_obs_start : fingertip_obs_start + num_ft_states
-        ] = self.fingertip_state.reshape(self.num_envs, num_ft_states)
+        self.obs_buf[:, fingertip_obs_start : fingertip_obs_start + num_ft_states] = (
+            self.fingertip_state.reshape(self.num_envs, num_ft_states)
+        )
         self.obs_buf[
             :,
             fingertip_obs_start
@@ -763,12 +777,12 @@ def compute_full_state(self, asymm_obs=False):
 
         # another_hand
         another_hand_start = action_obs_start + 26
-        self.obs_buf[
-            :, another_hand_start : self.num_shadow_hand_dofs + another_hand_start
-        ] = unscale(
-            self.shadow_hand_another_dof_pos,
-            self.shadow_hand_dof_lower_limits,
-            self.shadow_hand_dof_upper_limits,
+        self.obs_buf[:, another_hand_start : self.num_shadow_hand_dofs + another_hand_start] = (
+            unscale(
+                self.shadow_hand_another_dof_pos,
+                self.shadow_hand_dof_lower_limits,
+                self.shadow_hand_dof_upper_limits,
+            )
         )
         self.obs_buf[
             :,
@@ -802,9 +816,9 @@ def compute_full_state(self, asymm_obs=False):
         )
 
         hand_another_pose_start = fingertip_another_obs_start + 95
-        self.obs_buf[
-            :, hand_another_pose_start : hand_another_pose_start + 3
-        ] = self.hand_positions[self.another_hand_indices, :]
+        self.obs_buf[:, hand_another_pose_start : hand_another_pose_start + 3] = (
+            self.hand_positions[self.another_hand_indices, :]
+        )
         self.obs_buf[:, hand_another_pose_start + 3 : hand_another_pose_start + 4] = get_euler_xyz(
             self.hand_orientations[self.another_hand_indices, :]
         )[0].unsqueeze(-1)