Author fixed-root-link world joint with pure USD

[rebeccazhang0707]

Description

modify_articulation_root_properties created the fixed-root-link world joint through omni.physx.scripts.utils.createJoint, so spawning a fixed-base articulation (fix_root_link=True) raised ModuleNotFoundError: omni.physx on kitless backends (e.g. Newton) where Kit and PhysX extensions are not loaded.

This PR replaces that call with a pure-USD _create_world_fixed_joint helper that reproduces the behavior of omni.physx.createJoint for the single-body (world-attached) case: it targets only body1 (the root link), anchors localPos0/localRot0 at the root link's world pose with scale/shear removed, sets effectively unbreakable break force/torque, walks up to the first writable ancestor when the root link lives inside an instanceable/prototype subtree, and uniquifies the joint name on collision. The same code path now runs on every backend, so Kit/PhysX behavior is unchanged.

Regression tests run on an in-memory USD stage without launching Isaac Sim / Kit, covering the world-anchored joint attributes and the instanceable-root fallback. Since this helper is the single fixed-root-link authoring path for all backends, passing without Kit also proves the kitless case.

Type of change

• Bug fix (non-breaking change which fixes an issue)

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have added a changelog fragment under source/<pkg>/changelog.d/ for every touched package (do not edit CHANGELOG.rst or bump extension.toml — CI handles that)
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[greptile-apps]

Greptile Summary

This PR replaces the omni.physx.scripts.utils.createJoint call in modify_articulation_root_properties with a self-contained _create_world_fixed_joint helper that authors the same UsdPhysics.FixedJoint using only public USD APIs, removing the hard omni.physx import that caused ModuleNotFoundError on kitless backends such as Newton.

• _create_world_fixed_joint replicates the original behavior faithfully: body0 is left unset (world frame), body1 targets the articulation root, localPos0/localRot0 are derived from the root's world pose, and break thresholds are set to FLT_MAX. Instanceable/prototype ancestor traversal mirrors the original createJoint logic.
• Two new in-memory-stage regression tests verify the world-anchored joint attributes and the instanceable-root fallback without requiring a running Isaac Sim/Kit process.

Confidence Score: 4/5

The production code in schemas.py is correct and safe to merge; the new helper faithfully reproduces the omni.physx joint-authoring logic using only public USD APIs.

The core fix is straightforward: every joint attribute matches the original createJoint call, the instanceable-ancestor walk-up is correct, and break thresholds replicate FLT_MAX. The only gaps are in the test file — the localRot0 extraction path is never exercised with a non-identity rotation, and the uniquifier deduplication loop has no test coverage. Neither gap affects the production code path.

source/isaaclab/test/sim/test_schemas_kitless_fixed_joint.py — missing coverage for the rotation extraction and uniquifier branches.

Important Files Changed

Filename	Overview
source/isaaclab/isaaclab/sim/schemas/schemas.py	Replaces omni.physx.scripts.utils.createJoint with a pure-USD _create_world_fixed_joint helper; import of Gf added; all joint attributes match the original call's intent.
source/isaaclab/test/sim/test_schemas_kitless_fixed_joint.py	New kitless-safe regression tests on an in-memory USD stage; covers world-anchored joint attributes and instanceable-root fallback, but the localRot0 path and uniquifier logic are not exercised.
source/isaaclab/changelog.d/schemas-kitless-fixed-joint.rst	Changelog fragment correctly describes the fix and references the affected public function.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[modify_articulation_root_properties] --> B{fix_root_link set?}
    B -- No --> Z[return True]
    B -- Yes --> C{existing fixed joint?}
    C -- Yes --> D[enable/disable existing joint]
    C -- No, fix_root_link=True --> E{prim has RigidBodyAPI?}
    E -- No --> F[raise NotImplementedError]
    E -- Yes --> G[_create_world_fixed_joint]
    G --> H{prim instanceable or in prototype?}
    H -- Yes --> I[walk up to first writable ancestor]
    H -- No --> J[use articulation prim as base]
    I --> K{base == pseudo-root?}
    K -- Yes --> L[joint_base_path = empty]
    K -- No --> M[joint_base_path = ancestor path]
    J --> N[joint_base_path = articulation prim path]
    L & M & N --> O{FixedJoint name taken?}
    O -- No --> P[UsdPhysics.FixedJoint.Define]
    O -- Yes --> Q[increment uniquifier until free]
    Q --> P
    P --> R[Set body1, localPos0, localRot0, localPos1, localRot1, breakForce, breakTorque]
    R --> S[move ArticulationRootAPI to parent prim]
    S --> Z

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
    A[modify_articulation_root_properties] --> B{fix_root_link set?}
    B -- No --> Z[return True]
    B -- Yes --> C{existing fixed joint?}
    C -- Yes --> D[enable/disable existing joint]
    C -- No, fix_root_link=True --> E{prim has RigidBodyAPI?}
    E -- No --> F[raise NotImplementedError]
    E -- Yes --> G[_create_world_fixed_joint]
    G --> H{prim instanceable or in prototype?}
    H -- Yes --> I[walk up to first writable ancestor]
    H -- No --> J[use articulation prim as base]
    I --> K{base == pseudo-root?}
    K -- Yes --> L[joint_base_path = empty]
    K -- No --> M[joint_base_path = ancestor path]
    J --> N[joint_base_path = articulation prim path]
    L & M & N --> O{FixedJoint name taken?}
    O -- No --> P[UsdPhysics.FixedJoint.Define]
    O -- Yes --> Q[increment uniquifier until free]
    Q --> P
    P --> R[Set body1, localPos0, localRot0, localPos1, localRot1, breakForce, breakTorque]
    R --> S[move ArticulationRootAPI to parent prim]
    S --> Z

Loading

_{Reviews (1): Last reviewed commit: "Author fixed-root-link world joint with ..." | Re-trigger Greptile}

[greptile-apps]

localRot0 code path is untested

Both test functions use a prim with a zero or identity rotation, so joint.GetLocalRot0Attr().Get() is always the identity quaternion and the Gf.Quatf(world_pose.ExtractRotationQuat()) branch is never exercised. If the rotation-extraction logic regresses (e.g. wrong matrix decomposition order, inadvertent RemoveScaleShear stripping the rotation for scaled parents), no test would catch it. Adding a third test that translates and rotates the prim, then asserts the authored localRot0 matches the expected quaternion, would close this gap.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

[greptile-apps]

Uniquifier logic not tested

The name-deduplication loop inside _create_world_fixed_joint (FixedJoint → FixedJoint0 → …) is never exercised. A test that calls _create_world_fixed_joint twice on the same parent prim and asserts that two distinct joint prims exist with unique paths would cover this branch and protect against off-by-one regressions in the uniquifier counter.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!