[Fix] Register Warp task variants in benchmark and environment scripts#5784
Conversation
Three benchmark scripts (benchmark_rsl_rl, benchmark_non_rl,
benchmark_rlgames) and four environment scripts (list_envs,
random_agent, zero_agent, export_IODescriptors) import
isaaclab_tasks but not isaaclab_tasks_experimental, so the -Warp-v0
task variants never register in gymnasium. Running any of them with
e.g. --task Isaac-Cartpole-Direct-Warp-v0 fails with
gymnasium.error.NameNotFound.
Add the same conditional import the RL launch scripts already use:
import isaaclab_tasks
with contextlib.suppress(ImportError):
import isaaclab_tasks_experimental
The suppress block keeps the scripts working when the experimental
extension isn't installed.
Refs NVBug 6224916
There was a problem hiding this comment.
Review Summary
This is a clean and well-documented bug fix. The PR correctly addresses the missing task registration for -Warp-v0 variants in benchmark and environment scripts.
✅ What looks good
-
Follows established patterns: The conditional import pattern using
contextlib.suppress(ImportError)matches exactly what the RL training scripts (e.g.,scripts/reinforcement_learning/rsl_rl/train.py:56-58) already use. Good consistency. -
Graceful degradation: When
isaaclab_tasks_experimentalis not installed, the scripts continue to work for non-Warp tasks. This maintains backward compatibility with minimal Isaac Lab installs. -
Import placement: The experimental import is correctly placed after
import isaaclab_tasksand before anyfrom isaaclab_tasks.utils import ...statements, ensuring task registration happens before any task lookups. -
Documentation: The commit message and PR description are thorough — root cause, fix rationale, verification steps, and cherry-pick guidance are all clearly documented.
-
Changelog fragment: Properly added with correct RST formatting.
📝 Minor observations
-
PLACEHOLDER comment placement inconsistency: In
scripts/environments/export_IODescriptors.py, the# PLACEHOLDER: Extension templatecomment appears after the new import block (line 37), while in the benchmark scripts it appears before (e.g.,benchmark_non_rl.pyline 88). This is cosmetic but worth noting for future consistency. -
Broken Links CI check: The CI shows a failing "Check for Broken Links" job. This appears unrelated to the changes in this PR (no documentation links were modified), but worth confirming it's a pre-existing issue.
Verdict
The fix is correct and follows established conventions. The code is ready to merge once CI completes (assuming the broken links issue is unrelated).
Greptile SummaryAdds the missing
Confidence Score: 4/5The fix is straightforward and functionally correct; the only gap is a cosmetic comment placement inconsistency in three of the seven changed files. All seven scripts now correctly acquire the experimental task registry when available, with the ImportError safely suppressed for minimal installs. Three environment scripts omit the PLACEHOLDER comment that the canonical pattern and the four other changed files include immediately above the suppress block; the comment is left orphaned further down each file. scripts/environments/random_agent.py, scripts/environments/zero_agent.py, scripts/environments/export_IODescriptors.py — the PLACEHOLDER comment needs to be moved above the new import block to match the canonical pattern. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Script Startup] --> B[import isaaclab_tasks]
B --> C{contextlib.suppress block}
C -->|ImportError silenced| D[isaaclab_tasks_experimental unavailable\nMinimal install — Warp tasks not registered]
C -->|Success| E[isaaclab_tasks_experimental imported\nWarp -Warp-v0 task variants registered]
D --> F[gym.make task]
E --> F
F -->|Warp task + no experimental| G[NameNotFound error — FIXED by this PR]
F -->|Standard task OR Warp + experimental| H[Environment created successfully]
Reviews (1): Last reviewed commit: "[Fix] Register Warp task variants in ben..." | Re-trigger Greptile |
| import isaaclab_tasks # noqa: F401 | ||
|
|
||
| with contextlib.suppress(ImportError): | ||
| import isaaclab_tasks_experimental # noqa: F401 | ||
| from isaaclab_tasks.utils import ( |
There was a problem hiding this comment.
# PLACEHOLDER: Extension template (do not remove this comment) prefix that the canonical pattern requires (see scripts/reinforcement_learning/rsl_rl/train.py:56-58). The comment was already present further down the file (line 39) and now sits orphaned after the import it was meant to mark. The same issue exists in zero_agent.py and export_IODescriptors.py.
| import isaaclab_tasks # noqa: F401 | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import ( | |
| import isaaclab_tasks # noqa: F401 | |
| # PLACEHOLDER: Extension template (do not remove this comment) | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import ( |
| import isaaclab_tasks # noqa: F401 | ||
|
|
||
| with contextlib.suppress(ImportError): | ||
| import isaaclab_tasks_experimental # noqa: F401 | ||
| from isaaclab_tasks.utils import ( |
There was a problem hiding this comment.
# PLACEHOLDER: Extension template (do not remove this comment) prefix as in random_agent.py. The comment currently lives orphaned at line 39 after the block it should introduce.
| import isaaclab_tasks # noqa: F401 | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import ( | |
| import isaaclab_tasks # noqa: F401 | |
| # PLACEHOLDER: Extension template (do not remove this comment) | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import ( |
| import isaaclab_tasks # noqa: F401 | ||
|
|
||
| with contextlib.suppress(ImportError): | ||
| import isaaclab_tasks_experimental # noqa: F401 | ||
| from isaaclab_tasks.utils import parse_env_cfg |
There was a problem hiding this comment.
# PLACEHOLDER: Extension template (do not remove this comment) prefix. The comment remains at its original position after parse_env_cfg, separated from the block it's meant to annotate.
| import isaaclab_tasks # noqa: F401 | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import parse_env_cfg | |
| import isaaclab_tasks # noqa: F401 | |
| # PLACEHOLDER: Extension template (do not remove this comment) | |
| with contextlib.suppress(ImportError): | |
| import isaaclab_tasks_experimental # noqa: F401 | |
| from isaaclab_tasks.utils import parse_env_cfg |
isaac-sim#5784) ## 1. Summary - Benchmark scripts and standalone environment helpers failed with ``gymnasium.error.NameNotFound: Environment 'Isaac-Cartpole-Direct-Warp' doesn't exist`` when invoked with any ``-Warp-v0`` task. Docs at ``source/overview/core-concepts/physical-backends/newton/warp-environments.rst`` advertise the very command that fails. - Root cause: the affected scripts import ``isaaclab_tasks`` only; the ``-Warp-v0`` variants are registered in the optional ``isaaclab_tasks_experimental`` extension. The RL training/play scripts already conditionally import it; these did not. - Fix: add the same ``with contextlib.suppress(ImportError): import isaaclab_tasks_experimental`` block after the existing ``import isaaclab_tasks``. Seven scripts, 39 insertions total. ## 2. Scripts changed | File | Before | After | |---|---|---| | ``scripts/benchmarks/benchmark_rsl_rl.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/benchmarks/benchmark_non_rl.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/benchmarks/benchmark_rlgames.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/environments/list_envs.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/environments/random_agent.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/environments/zero_agent.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | | ``scripts/environments/export_IODescriptors.py`` | imports ``isaaclab_tasks`` only | + conditional ``isaaclab_tasks_experimental`` | All eight ``scripts/reinforcement_learning/{rl_games,rsl_rl,sb3,skrl}/{train,play}.py`` scripts already follow this pattern; the affected scripts are the remaining gap. ## 3. Verification - Pre-commit clean on all 8 changed files (ruff, ruff-format, codespell, RST hooks). - Pattern mirrors ``scripts/reinforcement_learning/rsl_rl/train.py:56-58``, which is the canonical placement. - ``contextlib.suppress(ImportError)`` keeps the scripts working when the experimental extension isn't installed (e.g. minimal IsaacLab installs). ## 4. Cherry-pick target The customer-visible bug is on ``release/3.0.0-beta2`` per the NVBug. After this lands on ``develop``, cherry-pick onto that release branch so QA can verify on the same build. Refs NVBug 6224916
… locomanip quaternion, deformable demo, IK, benchmark scripts (#5816) # Description Cherry pick following PRs from develop: - #5779 - #5760 - #5791 - #5810 - #5811 - #5812 - #5754 - #5802 - #5803 - #5782 - #5804 - #5744 - #5775 - #5784 --------- Signed-off-by: Kelly Guo <kellyg@nvidia.com> Co-authored-by: John <jaybdub@users.noreply.github.com> Co-authored-by: myurasov-nv <168484206+myurasov-nv@users.noreply.github.com> Co-authored-by: Antoine RICHARD <antoiner@nvidia.com> Co-authored-by: Mike Yan Michelis <46975745+mmichelis@users.noreply.github.com> Co-authored-by: hujc <jichuanh@nvidia.com>
1. Summary
gymnasium.error.NameNotFound: Environment 'Isaac-Cartpole-Direct-Warp' doesn't existwhen invoked with any-Warp-v0task. Docs atsource/overview/core-concepts/physical-backends/newton/warp-environments.rstadvertise the very command that fails.isaaclab_tasksonly; the-Warp-v0variants are registered in the optionalisaaclab_tasks_experimentalextension. The RL training/play scripts already conditionally import it; these did not.with contextlib.suppress(ImportError): import isaaclab_tasks_experimentalblock after the existingimport isaaclab_tasks. Seven scripts, 39 insertions total.2. Scripts changed
scripts/benchmarks/benchmark_rsl_rl.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/benchmarks/benchmark_non_rl.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/benchmarks/benchmark_rlgames.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/environments/list_envs.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/environments/random_agent.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/environments/zero_agent.pyisaaclab_tasksonlyisaaclab_tasks_experimentalscripts/environments/export_IODescriptors.pyisaaclab_tasksonlyisaaclab_tasks_experimentalAll eight
scripts/reinforcement_learning/{rl_games,rsl_rl,sb3,skrl}/{train,play}.pyscripts already follow this pattern; the affected scripts are the remaining gap.3. Verification
scripts/reinforcement_learning/rsl_rl/train.py:56-58, which is the canonical placement.contextlib.suppress(ImportError)keeps the scripts working when the experimental extension isn't installed (e.g. minimal IsaacLab installs).4. Cherry-pick target
The customer-visible bug is on
release/3.0.0-beta2per the NVBug. After this lands ondevelop, cherry-pick onto that release branch so QA can verify on the same build.Refs NVBug 6224916