Adapt DexSim v0.4.0

[yuecideng]

Description

Type of change

• Enhancement (non-breaking change which improves an existing functionality)
• Breaking change (existing functionality will not work without user modification)
• Documentation update

Screenshots

Please attach before and after screenshots of the change if applicable.

Checklist

• I have run the black . command to format the code base.
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• Dependencies have been updated, if applicable.

[Copilot]

Pull request overview

This PR updates EmbodiChain’s simulation/rendering integration to align with DexSim v0.4.0 by replacing the previous enable_rt-style flags with a structured rendering configuration (RenderCfg) and propagating that change through tests, examples, CLI helpers, configs, and docs.

Changes:

• Introduce RenderCfg and migrate SimulationManagerCfg rendering settings to render_cfg.
• Update scripts/examples/tests to use --renderer and/or RenderCfg(renderer=...) instead of --enable_rt.
• Apply broad formatting/whitespace cleanup across multiple modules and docs.

Reviewed changes

Copilot reviewed 80 out of 80 changed files in this pull request and generated 33 comments.

Show a summary per file

File	Description
tests/sim/sensors/test_stereo.py	Migrate stereo camera tests to RenderCfg/renderer-based configuration
tests/sim/sensors/test_contact.py	Migrate contact sensor tests to RenderCfg/renderer-based configuration
tests/sim/sensors/test_camera.py	Migrate camera tests to RenderCfg/renderer-based configuration
tests/sim/objects/test_usd.py	Update USD object tests to use render_cfg
tests/sim/objects/test_soft_object.py	Update soft object tests to use render_cfg
tests/sim/objects/test_robot.py	Whitespace cleanup
tests/gym/envs/test_embodied_env.py	Update env tests to use render_cfg and renderer arg
tests/gym/envs/managers/test_dataset_functors.py	Whitespace cleanup
tests/common.py	Whitespace cleanup
tests/agents/test_shared_rollout.py	Update sim config to use render_cfg
scripts/tutorials/sim/import_usd.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/gizmo_robot.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/export_usd.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/create_softbody.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/create_sensor.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/create_scene.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/create_robot.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/sim/create_rigid_object_group.py	Update tutorial script to new env launcher args + RenderCfg
scripts/tutorials/gym/random_reach.py	Replace ad-hoc CLI args with shared env launcher args
scripts/tutorials/gym/modular_env.py	Update modular env example to pass renderer/device/headless/num_envs
examples/sim/utility/workspace_analyzer/analyze_plane_workspace.py	Whitespace cleanup
examples/sim/sensors/create_contact_sensor.py	Update example to use env launcher args + RenderCfg
examples/sim/sensors/batch_camera.py	Update example to use env launcher args + RenderCfg
examples/sim/scene/scene_demo.py	Update scene demo CLI and rendering config to renderer-based setup
examples/sim/gizmo/gizmo_w1.py	Update gizmo example CLI and rendering config
examples/sim/gizmo/gizmo_scene.py	Update gizmo example CLI and rendering config
examples/sim/gizmo/gizmo_robot.py	Update gizmo example CLI and rendering config
examples/sim/gizmo/gizmo_object.py	Update gizmo example CLI and rendering config
examples/sim/gizmo/gizmo_camera.py	Update gizmo example CLI and rendering config
examples/sim/demo/scoop_ice.py	Update demo to use env launcher args + renderer configuration
examples/sim/demo/press_softbody.py	Update demo to use env launcher args + renderer configuration
examples/sim/demo/grasp_cup_to_caffe.py	Update demo to use env launcher args + renderer configuration
examples/agents/datasets/online_dataset_demo.py	Adjust demo engine config keys for renderer-based setup
embodichain/utils/warp/kinematics/opw_solver.py	Whitespace cleanup
embodichain/utils/configclass.py	Whitespace cleanup
embodichain/utils/init.py	Whitespace cleanup
embodichain/toolkits/urdf_assembly/component.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/visualizers/voxel_visualizer.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/visualizers/sphere_visualizer.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/visualizers/base_visualizer.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/samplers/base_sampler.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/constraints/workspace_constraint.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/constraints/base_constraint.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/configs/init.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/caches/cache_manager.py	Whitespace cleanup
embodichain/lab/sim/utility/workspace_analyzer/caches/base_cache.py	Whitespace cleanup
embodichain/lab/sim/types.py	Whitespace cleanup
embodichain/lab/sim/solvers/pinocchio_solver.py	Whitespace cleanup
embodichain/lab/sim/solvers/opw_solver.py	Whitespace cleanup
embodichain/lab/sim/solvers/differential_solver.py	Whitespace cleanup
embodichain/lab/sim/sim_manager.py	Replace old RT flags with render_cfg, update renderer wiring and behavior
embodichain/lab/sim/robots/dexforce_w1/utils.py	Whitespace cleanup
embodichain/lab/sim/objects/gizmo.py	Whitespace cleanup
embodichain/lab/sim/cfg.py	Add RenderCfg configclass and DexSim renderer flag mapping
embodichain/lab/sim/atom_actions.py	Whitespace cleanup
embodichain/lab/scripts/run_agent.py	Whitespace cleanup
embodichain/lab/gym/utils/gym_utils.py	Update shared CLI args + config merging to use renderer/RenderCfg
embodichain/lab/gym/envs/tasks/tableware/pour_water/action_bank.py	Whitespace cleanup
embodichain/lab/gym/envs/managers/randomization/spatial.py	Whitespace cleanup
embodichain/lab/gym/envs/managers/randomization/physics.py	Whitespace cleanup
embodichain/lab/gym/envs/embodied_env.py	Whitespace cleanup
embodichain/lab/gym/envs/action_bank/utils.py	Whitespace cleanup
embodichain/data/assets/scene_assets.py	Whitespace cleanup
embodichain/data/assets/robot_assets.py	Whitespace cleanup
embodichain/data/assets/obj_assets.py	Whitespace cleanup
embodichain/data/assets/materials.py	Whitespace cleanup
embodichain/data/assets/eef_assets.py	Whitespace cleanup
embodichain/agents/rl/train.py	Switch trainer config handling from enable_rt to renderer/RenderCfg
embodichain/agents/rl/models/mlp.py	Whitespace cleanup
embodichain/agents/engine/data.py	Switch engine sim config from enable_rt to renderer/RenderCfg
embodichain/agents/datasets/sampler.py	Whitespace cleanup
embodichain/agents/datasets/online_data.py	Whitespace cleanup
docs/sync_readme.py	Formatting tweak
docs/source/tutorial/sensor.rst	Update tutorial CLI docs from --enable_rt to --renderer
docs/source/tutorial/robot.rst	Update tutorial CLI docs from --enable_rt to --renderer
docs/source/tutorial/gizmo.rst	Update tutorial CLI docs from --enable_rt to --renderer
docs/source/overview/sim/sim_manager.md	Update sim manager docs to reflect renderer configuration changes
configs/agents/rl/push_cube/train_config.json	Replace enable_rt with renderer and reformat JSON
configs/agents/rl/basic/cart_pole/train_config_grpo.json	Replace enable_rt with renderer and reformat JSON
configs/agents/rl/basic/cart_pole/train_config.json	Replace enable_rt with renderer and reformat JSON

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

This PR adapts the simulation stack and surrounding tooling to DexSim v0.4.0 by replacing the legacy enable_rt flag with a structured rendering configuration (RenderCfg) and a --renderer CLI option across tests, examples, and training/engine code.

Changes:

• Introduce RenderCfg and migrate SimulationManagerCfg from enable_rt to render_cfg.
• Update CLI utilities and many scripts/examples to use --renderer (legacy/hybrid/fast-rt).
• Refresh tests/configs/docs to reflect the new rendering configuration.

Reviewed changes

Copilot reviewed 81 out of 81 changed files in this pull request and generated 15 comments.

Show a summary per file

File	Description
tests/sim/sensors/test_stereo.py	Migrate stereo sensor tests to new rendering config (currently has call-site issues).
tests/sim/sensors/test_contact.py	Migrate contact sensor tests to new rendering config (currently has call-site issues).
tests/sim/sensors/test_camera.py	Migrate camera sensor tests to new rendering config (currently has call-site issues).
tests/sim/objects/test_usd.py	Update SimulationManagerCfg construction for new renderer config.
tests/sim/objects/test_soft_object.py	Update SimulationManagerCfg construction for new renderer config.
tests/sim/objects/test_robot.py	Whitespace cleanup.
tests/gym/envs/test_embodied_env.py	Update env test sim config for new renderer config (currently has call-site issues).
tests/gym/envs/managers/test_dataset_functors.py	Whitespace cleanup.
tests/common.py	Whitespace cleanup.
tests/agents/test_shared_rollout.py	Update env sim config to use render_cfg in rollout test.
scripts/tutorials/sim/import_usd.py	Switch to shared CLI args + renderer config (currently hard-codes values).
scripts/tutorials/sim/gizmo_robot.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/sim/export_usd.py	Switch to shared CLI args + renderer gating for lighting.
scripts/tutorials/sim/create_softbody.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/sim/create_sensor.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/sim/create_scene.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/sim/create_robot.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/sim/create_rigid_object_group.py	Switch to shared CLI args + RenderCfg(renderer=args.renderer).
scripts/tutorials/gym/random_reach.py	Use shared CLI args for env launching.
scripts/tutorials/gym/modular_env.py	Add renderer option and build sim config from shared CLI args.
examples/sim/utility/workspace_analyzer/analyze_plane_workspace.py	Whitespace cleanup.
examples/sim/sensors/create_contact_sensor.py	Switch to shared CLI args + renderer config.
examples/sim/sensors/batch_camera.py	Switch to shared CLI args + renderer config.
examples/sim/scene/scene_demo.py	Switch to shared CLI args + renderer config.
examples/sim/gizmo/gizmo_w1.py	Switch to shared CLI args + renderer config.
examples/sim/gizmo/gizmo_scene.py	Switch to shared CLI args + renderer config.
examples/sim/gizmo/gizmo_robot.py	Switch to shared CLI args + renderer config.
examples/sim/gizmo/gizmo_object.py	Switch to shared CLI args + renderer config.
examples/sim/gizmo/gizmo_camera.py	Switch to shared CLI args + renderer config.
examples/sim/demo/scoop_ice.py	Add shared CLI args and pass renderer into sim initialization.
examples/sim/demo/press_softbody.py	Switch to shared CLI args + renderer config.
examples/sim/demo/grasp_cup_to_caffe.py	Switch to shared CLI args + renderer config + lighting gate.
examples/agents/datasets/online_dataset_demo.py	Update gym config handling for renderer (currently contains invalid syntax).
embodichain/utils/warp/kinematics/opw_solver.py	Whitespace cleanup.
embodichain/utils/configclass.py	Whitespace cleanup.
embodichain/utils/init.py	Whitespace cleanup.
embodichain/toolkits/urdf_assembly/component.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/visualizers/voxel_visualizer.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/visualizers/sphere_visualizer.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/visualizers/base_visualizer.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/samplers/base_sampler.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/constraints/workspace_constraint.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/constraints/base_constraint.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/configs/init.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/caches/cache_manager.py	Whitespace cleanup.
embodichain/lab/sim/utility/workspace_analyzer/caches/base_cache.py	Whitespace cleanup.
embodichain/lab/sim/utility/sim_utils.py	Extend RT detection + (currently) breaks SDF path due to missing import.
embodichain/lab/sim/types.py	Whitespace cleanup.
embodichain/lab/sim/solvers/pinocchio_solver.py	Whitespace cleanup.
embodichain/lab/sim/solvers/opw_solver.py	Whitespace cleanup.
embodichain/lab/sim/solvers/differential_solver.py	Whitespace cleanup.
embodichain/lab/sim/sim_manager.py	Replace enable_rt with render_cfg and update renderer/RT logic.
embodichain/lab/sim/robots/dexforce_w1/utils.py	Whitespace cleanup.
embodichain/lab/sim/objects/gizmo.py	Whitespace cleanup.
embodichain/lab/sim/cfg.py	Add RenderCfg and renderer mapping into DexSim flags.
embodichain/lab/sim/atom_actions.py	Whitespace cleanup.
embodichain/lab/scripts/run_agent.py	Whitespace cleanup.
embodichain/lab/gym/utils/gym_utils.py	Add --renderer, merge into config, and build sim cfg with RenderCfg.
embodichain/lab/gym/envs/tasks/tableware/pour_water/action_bank.py	Whitespace cleanup.
embodichain/lab/gym/envs/managers/randomization/spatial.py	Whitespace cleanup.
embodichain/lab/gym/envs/managers/randomization/physics.py	Whitespace cleanup.
embodichain/lab/gym/envs/embodied_env.py	Whitespace cleanup.
embodichain/lab/gym/envs/action_bank/utils.py	Whitespace cleanup.
embodichain/data/assets/scene_assets.py	Whitespace cleanup.
embodichain/data/assets/robot_assets.py	Whitespace cleanup.
embodichain/data/assets/obj_assets.py	Whitespace cleanup.
embodichain/data/assets/materials.py	Whitespace cleanup.
embodichain/data/assets/eef_assets.py	Whitespace cleanup.
embodichain/agents/rl/train.py	Replace trainer enable_rt with renderer and wire into RenderCfg.
embodichain/agents/rl/models/mlp.py	Whitespace cleanup.
embodichain/agents/engine/data.py	Replace engine sim config enable_rt with RenderCfg(renderer=...).
embodichain/agents/datasets/sampler.py	Whitespace cleanup.
embodichain/agents/datasets/online_data.py	Whitespace cleanup.
docs/sync_readme.py	Formatting tweak.
docs/source/tutorial/sensor.rst	Update CLI docs from --enable_rt to --renderer (needs example fix).
docs/source/tutorial/robot.rst	Update CLI docs from --enable_rt to --renderer (needs example fix).
docs/source/tutorial/gizmo.rst	Update CLI docs from --enable_rt to --renderer (needs wording/example fix).
docs/source/overview/sim/sim_manager.md	Update config docs for renderer (currently incorrect type/fields).
configs/agents/rl/push_cube/train_config.json	Replace enable_rt with renderer.
configs/agents/rl/basic/cart_pole/train_config.json	Replace enable_rt with renderer.
configs/agents/rl/basic/cart_pole/train_config_grpo.json	Replace enable_rt with renderer and formatting changes.

Comments suppressed due to low confidence (3)

embodichain/lab/gym/utils/gym_utils.py:862

• build_env_cfg_from_args imports RenderCfg from embodichain.lab.sim, but the package embodichain.lab.sim doesn't export RenderCfg (only SimulationManager* via sim_manager.__all__). This will raise ImportError. Import it from embodichain.lab.sim.cfg or export RenderCfg in embodichain/lab/sim/__init__.py.

    from embodichain.utils.utility import load_json
    from embodichain.lab.gym.envs import EmbodiedEnvCfg
    from embodichain.lab.sim import RenderCfg, SimulationManagerCfg

    gym_config = load_json(args.gym_config)
    gym_config = merge_args_with_gym_config(args, gym_config)

    cfg: EmbodiedEnvCfg = config_to_cfg(

embodichain/lab/gym/utils/gym_utils.py:801

• --gym_config was changed to required=False, but build_env_cfg_from_args unconditionally calls load_json(args.gym_config). This makes run_env.py / run_agent.py fail later with a less clear error if the flag is omitted. Either keep it required for CLIs that need it, or add an explicit check with a helpful error message when args.gym_config is empty.

    parser.add_argument(
        "--gym_config",
        type=str,
        help="Path to gym config file.",
        default="",
        required=False,
    )
    parser.add_argument(
        "--action_config", type=str, help="Path to action config file.", default=None
    )
    parser.add_argument(

docs/source/overview/sim/sim_manager.md:40

• The SimulationManagerCfg table documents renderer as a bool with default False, but the code now uses render_cfg: RenderCfg (and CLI uses a string backend name). Please update the parameter name/type/default and describe the new RenderCfg fields (renderer/enable_denoiser/spp).

| `width` | `int` | `1920` | The width of the simulation window. |
| `height` | `int` | `1080` | The height of the simulation window. |
| `headless` | `bool` | `False` | Whether to run the simulation in headless mode (no Window). |
| `renderer` | `bool` | `False` | Whether to enable ray tracing rendering. |
| `enable_denoiser` | `bool` | `True` | Whether to enable denoising for ray tracing rendering. |
| `spp` | `int` | `64` | Samples per pixel for ray tracing rendering. Only valid when ray tracing is enabled and denoiser is False. |
| `gpu_id` | `int` | `0` | The gpu index that the simulation engine will be used. Affects gpu physics device. |
| `thread_mode` | `ThreadMode` | `RENDER_SHARE_ENGINE` | The threading mode for the simulation engine. |

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

setup_simulation now takes renderer (string) but the tests call it with a render_cfg= keyword argument, which will raise TypeError: setup_simulation() got an unexpected keyword argument 'render_cfg'. Update call sites to pass renderer="legacy"/"fast-rt"/... (or change setup_simulation to accept render_cfg).

[Copilot]

There are two classes named TestCameraRaster (and two named TestCameraFastRT) in the same module; the latter definition overwrites the former, so one set of tests will never run. Rename them to distinct CPU/CUDA variants (e.g., TestCameraRasterCPU/TestCameraRasterCUDA).

[Copilot]

setup_simulation takes renderer (string), but the derived test classes call it with render_cfg=..., which will fail with an unexpected keyword argument. Pass renderer="legacy" / renderer="fast-rt" instead (or update the helper signature).

[Copilot]

The module defines TestStereoCameraRaster twice and TestStereoCameraFastRT twice; later definitions overwrite earlier ones, so tests are silently dropped. Use distinct class names for CPU vs CUDA variants.

[Copilot]

setup_simulation signature was changed to accept renderer, but all setup methods call it with render_cfg=... which will raise TypeError. Adjust call sites to pass renderer="legacy"/"fast-rt" (or accept render_cfg in the helper).

[Copilot]

add_env_launcher_args_to_parser docstring claims --renderer defaults to 'legacy', but the actual argparse default is 'hybrid'. Please align the docstring with the code (and update any other docs that mention the default).

[Copilot]

The script now parses common launcher args, but the simulation config still hard-codes headless=True (ignoring --headless) and hard-codes the renderer to "fast-rt" via an always-true conditional. Use headless=args.headless and RenderCfg(renderer=args.renderer) so CLI flags actually take effect.

[Copilot]

The documentation shows --renderer without a value. Since --renderer takes a choice (legacy/hybrid/fast-rt), the example should include an explicit value (e.g. --renderer fast-rt).

Suggested change

	python scripts/tutorials/sim/create_sensor.py --renderer
	python scripts/tutorials/sim/create_sensor.py --renderer fast-rt

[Copilot]

The documentation shows --renderer without a value. Since --renderer expects a backend name (legacy/hybrid/fast-rt), update the example to pass a value (e.g. --renderer hybrid).

Suggested change

	python scripts/tutorials/sim/create_robot.py --renderer
	python scripts/tutorials/sim/create_robot.py --renderer hybrid

[Copilot]

The docs say --renderer "Enable ray tracing" but the flag actually selects the backend and requires a value (legacy/hybrid/fast-rt). Update wording and example usage to reflect this (e.g. --renderer fast-rt).

[Copilot]

Pull request overview

Copilot reviewed 86 out of 86 changed files in this pull request and generated 13 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

There are two classes with the same name TestCameraRaster in this module. Pytest will only keep the second definition, so one set of tests is silently dropped. Rename them (e.g., TestCameraRasterCPU / TestCameraRasterCUDA) to ensure both variants execute.

[Copilot]

There are two classes with the same name TestCameraFastRT in this module. The latter overwrites the former, so half the tests won't run under pytest. Rename them to distinct CPU/CUDA variants.

[Copilot]

setup_simulation takes renderer as the second parameter, but the test classes call it with render_cfg=..., which will raise TypeError at runtime. Either pass renderer="legacy"/"fast-rt"/"hybrid" or change setup_simulation's signature to accept a RenderCfg argument consistently.

[Copilot]

setup_simulation is defined as setup_simulation(self, sim_device, renderer="legacy"), but the test cases call it with render_cfg=.... This will fail with an unexpected keyword argument error; pass renderer=... or update the helper to accept a RenderCfg consistently.

[Copilot]

The docs table still describes renderer as a bool with default False, and still documents enable_denoiser/spp as top-level SimulationManagerCfg fields. In the code, rendering is now configured via render_cfg: RenderCfg with renderer: Literal[...] and enable_denoiser/spp nested under it. Please update the parameter table to reflect the new API (including the correct type/default).

Suggested change

	| `renderer` | `bool` | `False` | Whether to enable ray tracing rendering. |
	| `enable_denoiser` | `bool` | `True` | Whether to enable denoising for ray tracing rendering. |
	| `spp` | `int` | `64` | Samples per pixel for ray tracing rendering. Only valid when ray tracing is enabled and denoiser is False. |
	| `render_cfg` | `RenderCfg` | `RenderCfg()` | Rendering configuration (renderer type, denoiser, samples per pixel, and other rendering options). |

[Copilot]

The updated docstring says this callback supports both the old 4-arg signature and the new 3-arg signature, but the implementation returns early unless len(args) == 3. If backward compatibility is intended, handle the 4-arg case; otherwise, update the docstring to avoid misleading callers.

[Copilot]

add_env_launcher_args_to_parser adds --headless, --num_envs, etc., but those values are not applied to SimulationManagerCfg here, and open_window() is called unconditionally. This makes --headless ineffective and can break headless runs; wire headless=args.headless (and likely num_envs=args.num_envs) into SimulationManagerCfg, and only open a window when not headless.

[Copilot]

setup_simulation is defined as setup_simulation(self, sim_device, renderer="legacy"), but the test classes call it with render_cfg=..., which will raise TypeError if these tests are ever un-skipped. Pass renderer=... (string) or update the helper to accept a RenderCfg consistently.

[Copilot]

add_env_launcher_args_to_parser provides --headless, but the simulation is always created with headless=True here. If the script is intended to support GUI mode, use headless=args.headless so the CLI flag actually controls window creation.

[Copilot]

Pull request overview

Copilot reviewed 86 out of 86 changed files in this pull request and generated 15 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

TestCameraFastRT is also defined twice (CPU and CUDA). The second class overwrites the first, so one variant may not run under pytest. Rename these classes to unique names.

[Copilot]

setup_simulation() takes a renderer string, but this test class calls it with render_cfg=..., which will raise an unexpected-keyword TypeError. Either change calls to pass renderer="legacy"/"fast-rt"/... or change setup_simulation() to accept render_cfg: RenderCfg.

[Copilot]

TestStereoCameraFastRT is also defined twice (CPU and CUDA). Since Python overwrites the first class with the second, one set of tests may never run. Rename the classes to distinct names.

[Copilot]

This is a second definition of TestStereoCameraRaster; it overwrites the earlier class with the same name, so only one of the CPU/CUDA variants will be collected by pytest. Rename to distinct class names per variant.

[Copilot]

This file now has duplicated imports (torch and SimulationManager are imported twice). It’s harmless at runtime but makes the module harder to maintain; remove the duplicate import block and keep a single canonical import section.

[Copilot]

The CLI option is now --renderer with explicit choices, but this bullet reads like a boolean toggle. Consider documenting it as --renderer {legacy,hybrid,fast-rt} and explaining what each value does (and/or the default) to match the new argparse contract.

Suggested change

	- ``--renderer``: Enable ray tracing for better visuals
	- ``--renderer {legacy,hybrid,fast-rt}``: Select rendering backend (default: ``fast-rt``)
	- ``legacy``: Classic rasterized renderer, compatible with older hardware
	- ``hybrid``: Mixed rasterization and ray tracing for balanced quality and performance
	- ``fast-rt``: Optimized real-time ray tracing for highest visual quality

[Copilot]

SimulationManagerCfg is constructed with headless=True regardless of args.headless, but later the script conditionally calls sim.open_window() based on args.headless. This can lead to inconsistent behavior (trying to open a window in a headless sim). Pass headless=args.headless into SimulationManagerCfg to keep the flags consistent.

[Copilot]

When distributed=True, the process GPU is set via local_rank, but gym_env_cfg.sim_cfg.gpu_id is then overwritten with the config-file gpu_id (not local_rank). This can desync the engine GPU selection from the actual process device. Keep gpu_id=local_rank in distributed mode (or at least don’t override the earlier gpu_index).

Suggested change

	gym_env_cfg.sim_cfg.gpu_id = gpu_id
	if device.type != "cuda" and hasattr(gym_env_cfg.sim_cfg, "gpu_id"):
	gym_env_cfg.sim_cfg.gpu_id = gpu_id

[Copilot]

There’s an indentation issue on the comment at line 239 (# convert to numpy 4x4 matrix) — it’s indented as if it were inside the previous if block, but it isn’t. This is misleading and may indicate accidental whitespace changes; align it with the surrounding code.

Suggested change

	# convert to numpy 4x4 matrix
	# convert to numpy 4x4 matrix

[Copilot]

SimulationManagerCfg.enable_rt/enable_denoiser/spp were removed in favor of render_cfg, but there are still call sites in the repo passing enable_rt (e.g., embodichain/lab/scripts/preview_asset.py:63-66, scripts/tutorials/sim/create_cloth.py:99-106, examples/sim/demo/pick_up_cloth.py:81-86). These will now crash with unexpected-keyword errors; either update those call sites in this PR or provide a backwards-compatible alias/deprecation path in SimulationManagerCfg.