newton/utils.py (3)

112-119: API removal looks good; consider a gentle compat shim for clearer errors.

Removing parse_{usd,urdf,mjcf} from newton.utils aligns with the PR objective. To reduce friction for downstream users importing these names, consider adding a module-level __getattr__ shim that raises a targeted error with migration guidance. This preserves the removal while yielding a crisp, actionable message on import.

Example addition (place near the end of the module):

# Compat shim for removed importer helpers (intentional: names are gone from __all__)
def __getattr__(name: str):
    removed = {
        "parse_usd": "ModelBuilder.add_usd(...)",
        "parse_urdf": "ModelBuilder.add_urdf(...)",
        "parse_mjcf": "ModelBuilder.add_mjcf(...)",
    }
    if name in removed:
        # Keep AttributeError so 'from newton.utils import parse_usd' fails,
        # but with a helpful, migration-oriented message.
        raise AttributeError(
            f"newton.utils.{name} was removed. Use {removed[name]} instead. "
            "See docs/migration.rst for examples."
        )
    raise AttributeError(name)

---

112-119: Update migration guide with before/after snippets.

Per the PR description, ensure docs/migration.rst includes concise examples mapping the old helpers to ModelBuilder.add_*. Keep it minimal and copy-pasteable.

Proposed outline:

• Old: from newton.utils import parse_usd → New: ModelBuilder.add_usd(path, ...)
• Old: parse_urdf(path, ...) → New: ModelBuilder.add_urdf(path, ...)
• Old: parse_mjcf(path, ...) → New: ModelBuilder.add_mjcf(path, ...)
• Note about the removal date and a one-line rationale.

---

24-48: Optional: freeze all at the end to avoid accidental mutation.

Small hygiene improvement: convert __all__ to a tuple at the end of the module. This prevents accidental in-place mutations elsewhere.

Apply at EOF:

__all__ = tuple(__all__)

Also applies to: 67-110, 117-129

Configuration used: Path: .coderabbit.yml

Review profile: CHILL

Plan: Pro

💡 Knowledge Base configuration:

• MCP integration is disabled by default for public repositories
• Jira integration is disabled by default for public repositories
• Linear integration is disabled by default for public repositories

You can enable these sources in your CodeRabbit configuration.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/utils.py:30-33
Timestamp: 2025-08-12T18:07:17.267Z
Learning: In newton/_src/utils/import_usd.py, the PXR/USD modules are lazily loaded using try-except blocks within the functions rather than at module import time, making it safe to import parse_usd at the module level without causing import-time failures for users without USD dependencies.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/utils.py:30-33
Timestamp: 2025-08-12T18:07:17.267Z
Learning: In newton/_src/utils/import_usd.py, the PXR/USD modules are lazily loaded using try-except blocks within the functions rather than at module import time, making it safe to import parse_usd at the module level without causing import-time failures for users without USD dependencies.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/utils.py:30-33
Timestamp: 2025-08-12T18:07:17.267Z
Learning: In newton/_src/utils/import_usd.py, the PXR/USD modules are lazily loaded using try-except blocks within the functions (lines 109-111 and 1289-1291) rather than at module import time, making it safe to import parse_usd at the module level without causing import-time failures for users without USD dependencies.

newton/_src/core/spatial.py (1)

243-245: Unify skew-matrix notation ([p]× vs [p]_x).

Use the same [p]_{×} symbol in the prose as in the equation to avoid confusion.

-    where *R* and *p* are the rotation and translation components of ``t`` and
-    ``[p]_x`` is the skew-symmetric matrix of *p*.
+    where *R* and *p* are the rotation and translation components of ``t`` and
+    ``[p]_{\\times}`` is the skew-symmetric matrix of *p*.

newton/utils.py (1)

111-118: Optional: add a helpful deprecation shim for removed symbols.

Provide a targeted AttributeError with migration guidance to reduce user friction.

+def __getattr__(name: str):
+    mapping = {
+        "parse_usd": "ModelBuilder.add_usd(...)",
+        "parse_urdf": "ModelBuilder.add_urdf(...)",
+        "parse_mjcf": "ModelBuilder.add_mjcf(...)",
+        "transform_inertia": "newton.geometry.transform_inertia(...)",
+    }
+    if name in mapping:
+        raise AttributeError(
+            f"newton.utils.{name} was removed. Use {mapping[name]} instead. "
+            "See docs/migration.rst for details."
+        )
+    raise AttributeError(f"module 'newton.utils' has no attribute '{name}'")

newton/_src/solvers/featherstone/kernels.py (1)

61-96: Semantics are correct; simplify matrix construction.

Compute R directly from the quaternion to reduce ops and match style used in geometry/inertia.

-    q = wp.transform_get_rotation(t_inv)
-    p = wp.transform_get_translation(t_inv)
-
-    r1 = wp.quat_rotate(q, wp.vec3(1.0, 0.0, 0.0))
-    r2 = wp.quat_rotate(q, wp.vec3(0.0, 1.0, 0.0))
-    r3 = wp.quat_rotate(q, wp.vec3(0.0, 0.0, 1.0))
-
-    R = wp.matrix_from_cols(r1, r2, r3)
+    q = wp.transform_get_rotation(t_inv)
+    p = wp.transform_get_translation(t_inv)
+    R = wp.quat_to_matrix(q)
     S = wp.skew(p) @ R
     T = wp.spatial_adjoint(R, S)
     return wp.mul(wp.mul(wp.transpose(T), I), T)

docs/api/newton_geometry.rst (1)

22-25: Docs entry additions look good; consider adding a versionchange note.

Add a short directive to signal the move from newton.utils to newton.geometry for transform_inertia.

Example (adjust version as appropriate):

.. versionchanged:: 0.x
   ``transform_inertia`` moved from ``newton.utils`` to ``newton.geometry``.

newton/_src/geometry/inertia.py (2)

421-423: Prefer wp.matmul to ensure device-side matmul portability

Some Warp builds don’t support the Python “@” operator in device code. Using wp.matmul avoids ambiguity.

Apply:

-    return R @ I @ wp.transpose(R) + m * (
+    return wp.matmul(wp.matmul(R, I), wp.transpose(R)) + m * (
         wp.dot(p, p) * wp.mat33(1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0) - wp.outer(p, p)
     )

---

387-416: Add minimal unit tests for transform_inertia

Cover identity rotation/shift, pure shift, and a simple rotation case to lock behavior.

I can draft tests exercising:

• q = identity, p = 0 → I’ == I
• q = identity, p = (a,0,0) → I’ adds m a^2 to yy/zz, zero to xx
• q = 90° about z → I’ == R I R^T
  Want me to open a small test PR?

Configuration used: Path: .coderabbit.yml

Review profile: CHILL

Plan: Pro

💡 Knowledge Base configuration:

• MCP integration is disabled by default for public repositories
• Jira integration is disabled by default for public repositories
• Linear integration is disabled by default for public repositories

You can enable these sources in your CodeRabbit configuration.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/geometry.py:16-22
Timestamp: 2025-08-12T17:51:37.474Z
Learning: In the Newton public API refactor, common geometry symbols like ParticleFlags and ShapeFlags are now exposed at the top level in newton/__init__.py rather than through newton.geometry. The newton/geometry.py module is intentionally focused only on broad-phase collision detection classes (BroadPhaseAllPairs, BroadPhaseExplicit, BroadPhaseSAP).

Learnt from: shi-eric
PR: newton-physics/newton#521
File: newton/examples/example_cloth_hanging.py:36-36
Timestamp: 2025-08-12T05:17:34.423Z
Learning: The Newton migration guide (docs/migration.rst) is specifically for documenting how to migrate existing warp.sim functionality to Newton equivalents. New Newton-only features that didn't exist in warp.sim do not need migration documentation.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/geometry.py:16-22
Timestamp: 2025-08-12T17:51:37.474Z
Learning: In the Newton public API refactor, common geometry symbols like ParticleFlags and ShapeFlags are now exposed at the top level in newton/__init__.py rather than through newton.geometry. The newton/geometry.py module is intentionally focused only on broad-phase collision detection classes (BroadPhaseAllPairs, BroadPhaseExplicit, BroadPhaseSAP).

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/_src/solvers/featherstone/kernels.py:18-20
Timestamp: 2025-08-12T17:58:00.815Z
Learning: In newton/_src/core/__init__.py, the function transform_twist is properly re-exported from the spatial module via "from .spatial import (..., transform_twist, ...)" and included in __all__, making it available for import as "from ...core import transform_twist" from other modules.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/_src/solvers/featherstone/kernels.py:18-20
Timestamp: 2025-08-12T17:58:00.815Z
Learning: In newton/_src/core/__init__.py, the function transform_twist is re-exported from the spatial module via "from .spatial import transform_twist", making it available for import as "from ...core import transform_twist" from other modules.

Learnt from: nvlukasz
PR: newton-physics/newton#519
File: newton/__init__.py:25-29
Timestamp: 2025-08-14T17:38:36.106Z
Learning: The Newton project prefers incremental __all__ building using __all__ += [...] pattern to group exports with their related imports, rather than a single consolidated __all__ at the end of the file.

Learnt from: shi-eric
PR: newton-physics/newton#521
File: newton/examples/example_cloth_hanging.py:36-36
Timestamp: 2025-08-12T05:17:34.423Z
Learning: The Newton migration guide (docs/migration.rst) is specifically for documenting how to migrate existing warp.sim functionality to Newton equivalents. New Newton-only features that didn't exist in warp.sim do not need migration documentation.