refactor: open pylint in pre-commit (#48)

Co-authored-by: Xuehai Pan <[email protected]>

.github/workflows/ci.yml

@@ -40,6 +40,10 @@ jobs:
         run: |
           python -m pip install -vvv --editable '.[lint]'
 
+      - name: Install safety_gymnasium
+        run: |
+          python -m pip install -vvv --editable 'envs/safety-gymnasium'
+
       - name: pre-commit
         run: |
           python -m pre_commit run --all-files

.pre-commit-config.yaml

@@ -28,18 +28,19 @@ repos:
     hooks:
       - id: black-jupyter
         stages: [commit, push, manual]
-  # - repo: local
-  #   hooks:
-  #     - id: pylint
-  #       name: pylint
-  #       entry: pylint
-  #       language: system
-  #       types: [python]
-  #       require_serial: true
-  #       stages: [commit, push, manual]
-  #       exclude: |
-  #         (?x)(
-  #           ^examples/|
-  #           ^tests/|
-  #           ^setup.py$
-  #         )
+  - repo: local
+    hooks:
+      - id: pylint
+        name: pylint
+        entry: pylint
+        language: system
+        types: [python]
+        require_serial: true
+        stages: [commit, push, manual]
+        exclude: |
+          (?x)(
+            ^examples/|
+            ^tests/|
+            ^setup.py$|
+            ^docs/source/conf.py$
+          )

docs/source/spelling_wordlist.txt

@@ -168,3 +168,16 @@ vel
 quaternion
 Quaternions
 Jacobian
+Lillicrap
+Erez
+Yuval
+Tassa
+Jiaming
+Ji
+Juntao
+Dai
+Linrui
+Binbin
+Zhou
+Pengfei
+Yaodong

envs/safety-gymnasium/safety_gymnasium/bases/__init__.py

@@ -14,5 +14,5 @@
 # ==============================================================================
 """Base classes."""
 
-from safety_gymnasium.bases.base_mujoco_task import BaseMujocoTask
-from safety_gymnasium.bases.base_task import BaseTask
+# from safety_gymnasium.bases.base_mujoco_task import BaseMujocoTask
+# from safety_gymnasium.bases.base_task import BaseTask

envs/safety-gymnasium/safety_gymnasium/bases/base_mujoco_task.py

@@ -18,7 +18,7 @@
 from copy import deepcopy
 from typing import Union
 
-import gymnasium
+# import gymnasium
 import mujoco
 import numpy as np
 from gymnasium.envs.mujoco.mujoco_rendering import RenderContextOffscreen, Viewer

envs/safety-gymnasium/safety_gymnasium/bases/base_task.py

@@ -25,7 +25,7 @@
 from safety_gymnasium.assets.mocaps import MOCAPS_REGISTER
 from safety_gymnasium.assets.objects import OBJS_REGISTER
 from safety_gymnasium.assets.robot import Robot
-from safety_gymnasium.bases import BaseMujocoTask
+from safety_gymnasium.bases.base_mujoco_task import BaseMujocoTask
 from safety_gymnasium.utils.common_utils import ResamplingError
 from safety_gymnasium.utils.task_utils import quat2mat, theta2vec
 

envs/safety-gymnasium/safety_gymnasium/tasks/button/button_level0.py

@@ -21,7 +21,7 @@
 import numpy as np
 from safety_gymnasium.assets.geoms import Buttons
 from safety_gymnasium.assets.group import GROUP
-from safety_gymnasium.bases import BaseTask
+from safety_gymnasium.bases.base_task import BaseTask
 
 
 # pylint: disable-next=too-many-instance-attributes

envs/safety-gymnasium/safety_gymnasium/tasks/goal/goal_level0.py

@@ -15,7 +15,7 @@
 """Goal level 0."""
 
 from safety_gymnasium.assets.geoms import Goal
-from safety_gymnasium.bases import BaseTask
+from safety_gymnasium.bases.base_task import BaseTask
 
 
 class GoalLevel0(BaseTask):

envs/safety-gymnasium/safety_gymnasium/tasks/push/push_level0.py

@@ -17,7 +17,7 @@
 import numpy as np
 from safety_gymnasium.assets.geoms import Goal
 from safety_gymnasium.assets.objects import PushBox
-from safety_gymnasium.bases import BaseTask
+from safety_gymnasium.bases.base_task import BaseTask
 
 
 class PushLevel0(BaseTask):

omnisafe/algorithms/off_policy/ddpg.py

@@ -30,12 +30,13 @@
 
 
 @registry.register
-class DDPG:
+class DDPG:  # pylint: disable=too-many-instance-attributes
     """Continuous control with deep reinforcement learning (DDPG) Algorithm.
 
     References:
         Paper Name: Continuous control with deep reinforcement learning.
-        Paper author: Timothy P. Lillicrap, Jonathan J. Hunt, Alexander Pritzel, Nicolas Heess, Tom Erez, Yuval Tassa, David Silver, Daan Wierstra.
+        Paper author: Timothy P. Lillicrap, Jonathan J. Hunt, Alexander Pritzel, Nicolas Heess,
+                      Tom Erez, Yuval Tassa, David Silver, Daan Wierstra.
         Paper URL: https://arxiv.org/abs/1509.02971
 
     """
@@ -96,7 +97,6 @@ def __init__(
         self.actor_critic = ConstraintActorQCritic(
             observation_space=self.env.observation_space,
             action_space=self.env.action_space,
-            scale_rewards=cfgs.scale_rewards,
             standardized_obs=cfgs.standardized_obs,
             model_cfgs=cfgs.model_cfgs,
         )
@@ -223,16 +223,16 @@ def compute_loss_v(self, data):
             data['obs_next'],
             data['done'],
         )
-        q = self.actor_critic.critic(obs, act)
+        q_value = self.actor_critic.critic(obs, act)
         # Bellman backup for Q function
         with torch.no_grad():
             act_targ, _ = self.ac_targ.actor.predict(obs, deterministic=True)
             q_targ = self.ac_targ.critic(obs_next, act_targ)
             backup = rew + self.cfgs.gamma * (1 - done) * q_targ
         # MSE loss against Bellman backup
-        loss_q = ((q - backup) ** 2).mean()
+        loss_q = ((q_value - backup) ** 2).mean()
         # Useful info for logging
-        q_info = dict(Q1Vals=q.detach().numpy())
+        q_info = dict(Q1Vals=q_value.detach().numpy())
         return loss_q, q_info
 
     def compute_loss_c(self, data):
@@ -249,17 +249,17 @@ def compute_loss_c(self, data):
             data['obs_next'],
             data['done'],
         )
-        qc = self.actor_critic.cost_critic(obs, act)
+        cost_q_value = self.actor_critic.cost_critic(obs, act)
 
         # Bellman backup for Q function
         with torch.no_grad():
             action, _ = self.ac_targ.pi.predict(obs_next, deterministic=True)
             qc_targ = self.ac_targ.c(obs_next, action)
             backup = cost + self.cfgs.gamma * (1 - done) * qc_targ
         # MSE loss against Bellman backup
-        loss_qc = ((qc - backup) ** 2).mean()
+        loss_qc = ((cost_q_value - backup) ** 2).mean()
         # Useful info for logging
-        qc_info = dict(QCosts=qc.detach().numpy())
+        qc_info = dict(QCosts=cost_q_value.detach().numpy())
 
         return loss_qc, qc_info
 

omnisafe/algorithms/on_policy/cup.py

@@ -152,7 +152,6 @@ def slice_data(self, data) -> dict:
                     'adv': adv[i * batch_size : (i + 1) * batch_size],
                     'discounted_ret': discounted_ret[i * batch_size : (i + 1) * batch_size],
                     'cost_adv': cost_adv[i * batch_size : (i + 1) * batch_size],
-                    'target_v': target_v[i * batch_size : (i + 1) * batch_size],
                 }
             )
 

omnisafe/common/lagrange.py

@@ -22,6 +22,7 @@
 class Lagrange(abc.ABC):
     """Abstract base class for Lagrangian-base Algorithms."""
 
+    # pylint: disable-next=too-many-arguments
     def __init__(
         self,
         cost_limit: float,

tests/test_policy.py

@@ -32,6 +32,7 @@
         'PCPO',
         'FOCOPS',
         'CPPOPid',
+        'CUP',
     ]
 )
 def test_on_policy(algo):