[pre-commit.ci] pre-commit autoupdate

.pre-commit-config.yaml

@@ -13,7 +13,7 @@ repos:
     -   id: check-yaml
     -   id: check-toml
 -   repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.14.5
+    rev: v0.16.9
     hooks:
     -   id: ruff
         args: [ --fix ]

docs/source/FAQ/index.md

@@ -122,8 +122,8 @@ from metasim.sim.sapien_handler import SapienHandler
 from metasim.sim.isaacsim_handler import IsaacSimHandler
 
 # Same scenario works with any simulator
-handler = MujocoHandler(scenario)    # or
-handler = SapienHandler(scenario)    # or
+handler = MujocoHandler(scenario)  # or
+handler = SapienHandler(scenario)  # or
 handler = IsaacSimHandler(scenario)
 ```
 

docs/source/dataset_benchmark/benchmark/usage.md

@@ -43,16 +43,8 @@ class RandomizationCfg:
 Without the need for knowing the details of randomized evaluation, you can simply define the level of randomization with numbers 0-3, and use the randomization config inside of a scenario config:
 
 ```python
-randomization = RandomizationCfg(
-    camera=False, light=False, ground=False, reflection=False
-)
-scenario = ScenarioCfg(
-    task=task,
-    robot=robot,
-    cameras=[camera],
-    randomization=randomization,
-    try_add_table=True
-)
+randomization = RandomizationCfg(camera=False, light=False, ground=False, reflection=False)
+scenario = ScenarioCfg(task=task, robot=robot, cameras=[camera], randomization=randomization, try_add_table=True)
 ```
 
 And the instantiated environment will be automatically randomized!

docs/source/dataset_benchmark/dataset/multiagent.md

@@ -16,7 +16,7 @@ demos; each demo carries `init_state`, `actions`, and optional `states`:
 
 ```python
 {
-    "franka_left":  [{"init_state": {...}, "actions": [...], "states": None}, ...],
+    "franka_left": [{"init_state": {...}, "actions": [...], "states": None}, ...],
     "franka_right": [{"init_state": {...}, "actions": [...], "states": None}, ...],
     "metadata": {"num_agents": 2, "agents": ["franka_left", "franka_right"]},
 }

docs/source/dataset_benchmark/integrations/libero.md

@@ -70,13 +70,13 @@ user-set value or the base `~/.libero`). No manual config editing needed.
 ## Usage
 
 ```python
-import roboverse_pack.tasks.libero          # auto-registers Libero/<suite>__<task>
-import roboverse_pack.tasks.libero_plus     # auto-registers LiberoPlus/<suite>__<task>
+import roboverse_pack.tasks.libero  # auto-registers Libero/<suite>__<task>
+import roboverse_pack.tasks.libero_plus  # auto-registers LiberoPlus/<suite>__<task>
 from roboverse_pack.tasks.libero_plus import make_liberoplus_env
 
 # build any of the 10,120 perturbation tasks by (suite, task index)
 env = make_liberoplus_env("libero_object", 7, seed=0)
-obs, reward, done, info = env.step([0.0] * 7)   # native legacy-gym 4-tuple (kept for fidelity)
+obs, reward, done, info = env.step([0.0] * 7)  # native legacy-gym 4-tuple (kept for fidelity)
 ```
 
 ## Native MetaSim tasks — run (and delete) LIBERO
@@ -93,9 +93,9 @@ other RoboVerse task and run with `libero`/`robosuite` uninstalled:
 from metasim.task.registry import get_task_class
 
 Task = get_task_class("libero_native.10__KITCHEN_SCENE3_turn_on_the_stove_and_put_the_moka_pot_on_it")
-env = Task()                                   # loads the vendored scene; no libero
+env = Task()  # loads the vendored scene; no libero
 obs, info = env.reset()
-obs, reward, terminated, time_out, info = env.step(action)   # 7-D OSC delta + gripper
+obs, reward, terminated, time_out, info = env.step(action)  # 7-D OSC delta + gripper
 ```
 
 Each task is a first-class `BaseTaskEnv` (`NativeLiberoEnv`); a static bundle under

docs/source/dataset_benchmark/integrations/mjlab.md

@@ -84,13 +84,15 @@ paths:
 
    ```python
    ScenarioCfg(
-       robots=[RobotCfg(
-           name="g1",
-           num_joints=29,
-           fix_base_link=False,
-           mjcf_path=mjlab_asset("asset_zoo/robots/unitree_g1/xmls/g1.xml"),
-           enabled_self_collisions=False,
-       )],
+       robots=[
+           RobotCfg(
+               name="g1",
+               num_joints=29,
+               fix_base_link=False,
+               mjcf_path=mjlab_asset("asset_zoo/robots/unitree_g1/xmls/g1.xml"),
+               enabled_self_collisions=False,
+           )
+       ],
        simulator="mujoco",
        add_default_ground=False,  # raw rollout doesn't add a ground either
    )

docs/source/dataset_benchmark/integrations/simpler_env.md

@@ -69,20 +69,23 @@ rendering works, no sm_120 wall.
 ## Usage
 
 ```python
-import roboverse_pack.tasks.simpler_env          # auto-registers SimplerEnv/<task> + simpler.<task>
+import roboverse_pack.tasks.simpler_env  # auto-registers SimplerEnv/<task> + simpler.<task>
 
 # (1) MetaSim-native via gym
 import gymnasium as gym
+
 env = gym.make("SimplerEnv/google_robot_pick_coke_can")
 obs, info = env.reset(seed=0)
 obs, reward, terminated, truncated, info = env.step(env.action_space.sample())
 
 # (2) MetaSim-native via the MetaSim task registry
 from metasim.task.registry import get_task_class
+
 task = get_task_class("simpler.widowx_stack_cube")()
 
 # (3) optional upstream passthrough (requires the SimplerEnv clone)
 from roboverse_pack.tasks.simpler_env import register_simpler_env_passthrough
+
 register_simpler_env_passthrough(prefix="SimplerEnvPassthrough/")
 env = gym.make("SimplerEnvPassthrough/google_robot_pick_coke_can")
 ```

docs/source/roboverse_learn/imitation_learning/contributing.md

@@ -136,6 +136,7 @@ from roboverse_learn.base.base_runner import BaseRunner
 abs_config_path = str(pathlib.Path(__file__).resolve().parent.joinpath("configs").absolute())
 OmegaConf.register_new_resolver("eval", eval, replace=True)
 
+
 @hydra.main(config_path=abs_config_path, version_base=None)
 def main(cfg):
     OmegaConf.resolve(cfg)
@@ -145,6 +146,7 @@ def main(cfg):
     runner: BaseRunner = cls(cfg)
     runner.run()
 
+
 if __name__ == "__main__":
     main()
 ```

docs/source/roboverse_learn/reinforcement_learning/fast_td3.md

@@ -13,45 +13,39 @@ The script uses a `CONFIG` dictionary for all parameters. Key options include:
 ```python
 CONFIG = {
     # Environment
-    "sim": "mjx",                    # Simulator backend
-    "robots": ["h1"],               # Robot type
-    "task": "humanoid.run",         # Task name
-    "num_envs": 1024,              # Number of parallel environments
-    "decimation": 10,              # Control decimation
-
+    "sim": "mjx",  # Simulator backend
+    "robots": ["h1"],  # Robot type
+    "task": "humanoid.run",  # Task name
+    "num_envs": 1024,  # Number of parallel environments
+    "decimation": 10,  # Control decimation
     # Training
-    "total_timesteps": 1500,       # Total training steps
-    "batch_size": 32768,           # Batch size for updates
-    "buffer_size": 20480,          # Replay buffer size
-
+    "total_timesteps": 1500,  # Total training steps
+    "batch_size": 32768,  # Batch size for updates
+    "buffer_size": 20480,  # Replay buffer size
     # Algorithm
-    "gamma": 0.99,                 # Discount factor
-    "tau": 0.1,                    # Target network update rate
-    "policy_frequency": 2,         # Policy update frequency
-    "num_updates": 12,             # Updates per step
-
+    "gamma": 0.99,  # Discount factor
+    "tau": 0.1,  # Target network update rate
+    "policy_frequency": 2,  # Policy update frequency
+    "num_updates": 12,  # Updates per step
     # Networks
     "critic_learning_rate": 0.0003,
     "actor_learning_rate": 0.0003,
     "critic_hidden_dim": 1024,
     "actor_hidden_dim": 512,
-
     # Distributional Q-learning
     "num_atoms": 101,
     "v_min": -250.0,
     "v_max": 250.0,
-
     # Optimizations
-    "use_cdq": True,               # Clipped Double Q-learning
-    "compile": True,               # PyTorch compilation
-    "obs_normalization": True,     # Observation normalization
-    "amp": True,                   # Automatic mixed precision
-    "amp_dtype": "fp16",          # Precision type
-
+    "use_cdq": True,  # Clipped Double Q-learning
+    "compile": True,  # PyTorch compilation
+    "obs_normalization": True,  # Observation normalization
+    "amp": True,  # Automatic mixed precision
+    "amp_dtype": "fp16",  # Precision type
     # Logging
-    "use_wandb": False,           # Weights & Biases integration
-    "eval_interval": 700,         # Evaluation frequency
-    "save_interval": 700,         # Model saving frequency
+    "use_wandb": False,  # Weights & Biases integration
+    "eval_interval": 700,  # Evaluation frequency
+    "save_interval": 700,  # Model saving frequency
 }
 ```
 

docs/source/roboverse_learn/reinforcement_learning/humanoid.md

@@ -264,8 +264,8 @@ ground_cfg = GroundCfg(
     dynamic_friction=1.0,
     elements={
         "slope": [SlopeCfg(origin=[0, 0], size=[2.0, 2.0], slope=0.3)],
-        "stair": [StairCfg(origin=[5, 0], size=[2.0, 2.0], step_height=0.1)]
-    }
+        "stair": [StairCfg(origin=[5, 0], size=[2.0, 2.0], step_height=0.1)],
+    },
 )
 ```
 

examples/obj_layout/README.md

@@ -119,14 +119,14 @@ class PutBananaTask(Task):
                 name="banana",
                 usd_path="...",
                 pos=(0.5, 0.2, 0.15),  # ← From saved poses
-                rot=(0, 0, 0, 1),      # ← From saved poses
+                rot=(0, 0, 0, 1),  # ← From saved poses
             ),
         ],
         robots=[
             RobotCfg(
                 name="franka",
-                pos=(0.0, 0.0, 0.0),   # ← From saved poses
-                rot=(0, 0, 0, 1),      # ← From saved poses
+                pos=(0.0, 0.0, 0.0),  # ← From saved poses
+                rot=(0, 0, 0, 1),  # ← From saved poses
             ),
         ],
     )

examples/rl/fast_td3/fttd3_module.py

@@ -431,7 +431,8 @@ def projection(
         next_dist = F.softmax(self.forward(obs, actions), dim=1)
         proj_dist = torch.zeros_like(next_dist)
         offset = (
-            torch.linspace(0, (batch_size - 1) * self.num_atoms, batch_size, device=device)
+            torch
+            .linspace(0, (batch_size - 1) * self.num_atoms, batch_size, device=device)
             .unsqueeze(1)
             .expand(batch_size, self.num_atoms)
             .long()

examples/viser/VISER_USAGE_GUIDE.md

@@ -185,8 +185,8 @@ visualizer.enable_trajectory_playback()
 visualizer.setup_ik_solver(robot_name, robot_config, env_handler)
 
 # Load and set trajectory
-visualizer.load_trajectory('trajectory.pkl.gz')
-visualizer.set_current_trajectory('franka', 0)
+visualizer.load_trajectory("trajectory.pkl.gz")
+visualizer.set_current_trajectory("franka", 0)
 ```
 
 ### Custom Scene Setup

packages/metasim/docs/source/concept/config.md

@@ -33,10 +33,8 @@ class ScenarioCfg:
     # Runtime
     render: RenderCfg = RenderCfg()
     sim_params: SimParamCfg = SimParamCfg()
-    simulator: Literal["isaaclab","isaacgym","sapien2","sapien3",
-                       "genesis","pybullet","mujoco"] | None = None
-    renderer:  Literal["isaaclab","isaacgym","sapien2","sapien3",
-                       "genesis","pybullet","mujoco"] | None = None
+    simulator: Literal["isaaclab", "isaacgym", "sapien2", "sapien3", "genesis", "pybullet", "mujoco"] | None = None
+    renderer: Literal["isaaclab", "isaacgym", "sapien2", "sapien3", "genesis", "pybullet", "mujoco"] | None = None
 
     # Misc
     num_envs: int = 1
@@ -153,13 +151,12 @@ The class provides mechanisms for asset management and dynamic updates:
 
 ```python
 RobotCfg(
-  name="robot_template",
-  num_joints=2,
-  urdf_path="roboverse_data/robots/your_robot/urdf/your_robot.urdf",
-  fix_base_link=True,
-  enabled_gravity=True,
-  control_type={"joint1": "position", "joint2": "effort"},
-  actuators={"joint1": BaseActuatorCfg(stiffness=500, damping=10),
-             "joint2": BaseActuatorCfg(effort_limit_sim=50)}
+    name="robot_template",
+    num_joints=2,
+    urdf_path="roboverse_data/robots/your_robot/urdf/your_robot.urdf",
+    fix_base_link=True,
+    enabled_gravity=True,
+    control_type={"joint1": "position", "joint2": "effort"},
+    actuators={"joint1": BaseActuatorCfg(stiffness=500, damping=10), "joint2": BaseActuatorCfg(effort_limit_sim=50)},
 )
 ```

packages/metasim/docs/source/concept/get_extras.md

@@ -53,8 +53,8 @@ Implement the `_extra_spec` method inside your task definition:
 ```python
 def _extra_spec(self) -> dict:
     return {
-        "imu_pos": SitePos("imu"),   # World position of IMU sensor site
-        "head_vel": BodyVel("head"), # Velocity of the robot's head
+        "imu_pos": SitePos("imu"),  # World position of IMU sensor site
+        "head_vel": BodyVel("head"),  # Velocity of the robot's head
     }
 ```
 
@@ -90,10 +90,7 @@ extras = handler.get_extra()
 This returns a dictionary containing your custom observations:
 
 ```python
-{
-    "imu_pos": tensor([[0.1, 0.2, 0.3], [...]]),
-    "head_vel": tensor([[0.01, 0.02, 0.03], [...]])
-}
+{"imu_pos": tensor([[0.1, 0.2, 0.3], [...]]), "head_vel": tensor([[0.01, 0.02, 0.03], [...]])}
 ```
 
 You can directly use this data for:
@@ -134,6 +131,7 @@ Example:
 from metasim.queries.base import BaseQueryType
 import torch
 
+
 class BodyMassQuery(BaseQueryType):
     def bind_handler(self, handler):
         # Store the handler for later use
@@ -158,7 +156,7 @@ After defining your custom query class, simply instantiate and use it directly i
 ```python
 def _extra_spec(self) -> dict:
     return {
-        "body_mass": BodyMassQuery(), # Your custom query
+        "body_mass": BodyMassQuery(),  # Your custom query
         "imu_pos": SitePos("imu"),
     }
 ```