[Draft] Cable joint to support minimal eval fk, ik - #3200
Conversation
Signed-off-by: JC <jumyungc@nvidia.com>
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yml Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthrough
ChangesCABLE kinematics and dynamics integration
Estimated code review effort: 3 (Moderate) | ~25 minutes Sequence Diagram(s)sequenceDiagram
participant ModelBuilder
participant JointState
participant eval_fk
participant eval_ik
participant MotionSubspace
ModelBuilder->>JointState: Initialize CABLE relative pose and twist coordinates
JointState->>eval_fk: Provide joint_q and joint_qd
eval_fk->>JointState: Reconstruct cable body transforms and velocities
JointState->>eval_ik: Provide cable body transforms and velocities
eval_ik->>JointState: Recover relative pose and twist coordinates
MotionSubspace->>MotionSubspace: Build six CABLE motion-subspace columns
Possibly related PRs
Suggested labels: Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
🧹 Nitpick comments (2)
newton/_src/sim/enums.py (1)
169-171: ⚡ Quick winTrim redundant inline narration in the CABLE branch.
This comment restates what the docstring and assignments already convey. Please shorten or remove it to keep inline comments focused on non-obvious constraints only.
As per coding guidelines, “Code comments should be brief and only for non-obvious code… not what the code already shows.”Proposed edit
- # DOF entries stay as VBD stretch + bend/twist slots, while position - # coordinates hold the full relative anchor pose (3 translation + 4 - # quaternion) so forward kinematics can reconstruct the rod. + # Keep CABLE velocity DoFs axis-driven while storing full anchor pose for FK reconstruction.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/_src/sim/enums.py` around lines 169 - 171, The inline comment spanning lines 169-171 in the CABLE branch is redundant as it restates information already conveyed by the docstring and variable assignments above it. Trim this comment to remove the narration about DOF entries, VBD stretch, bend/twist slots, and the full relative anchor pose details, keeping only any non-obvious constraints or implementation notes that are not already apparent from the code itself. If nothing non-obvious remains, remove the comment entirely.Source: Coding guidelines
newton/tests/test_cable.py (1)
3555-3563: ⚡ Quick winAdd an explicit assertion that
eval_ikpreserves CABLEjoint_qd.Line 3559 validates
joint_qrecovery, but this test does not currently lock in the paired contract that CABLEjoint_qdmust remain untouched.Suggested test addition
# Wipe joint_q so the round-trip can't lean on the builder's seeding. state.joint_q.zero_() + joint_qd_before = state.joint_qd.numpy().copy() + # eval_ik: recover joint_q (relative anchor pose) from the built body_q. newton.eval_ik(model, state, state.joint_q, state.joint_qd) jq = state.joint_q.numpy() test.assertTrue(np.any(np.abs(jq) > 1.0e-6), msg="eval_ik should populate CABLE joint_q") + np.testing.assert_allclose( + state.joint_qd.numpy(), + joint_qd_before, + rtol=0.0, + atol=0.0, + err_msg="eval_ik should leave CABLE joint_qd unchanged", + )🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/tests/test_cable.py` around lines 3555 - 3563, After the eval_ik function call in the test, capture the current state of joint_qd by converting it to a numpy array similar to how joint_q is captured, then add an assertion to verify that joint_qd remains unchanged (all values equal to zero or the original values) after the eval_ik call, ensuring that eval_ik does not modify joint_qd as part of the contract for CABLE joint handling.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@newton/_src/sim/enums.py`:
- Around line 169-171: The inline comment spanning lines 169-171 in the CABLE
branch is redundant as it restates information already conveyed by the docstring
and variable assignments above it. Trim this comment to remove the narration
about DOF entries, VBD stretch, bend/twist slots, and the full relative anchor
pose details, keeping only any non-obvious constraints or implementation notes
that are not already apparent from the code itself. If nothing non-obvious
remains, remove the comment entirely.
In `@newton/tests/test_cable.py`:
- Around line 3555-3563: After the eval_ik function call in the test, capture
the current state of joint_qd by converting it to a numpy array similar to how
joint_q is captured, then add an assertion to verify that joint_qd remains
unchanged (all values equal to zero or the original values) after the eval_ik
call, ensuring that eval_ik does not modify joint_qd as part of the contract for
CABLE joint handling.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 5ad934ac-b29b-44ef-9eb8-38fc869d938c
📒 Files selected for processing (5)
CHANGELOG.mdnewton/_src/sim/articulation.pynewton/_src/sim/builder.pynewton/_src/sim/enums.pynewton/tests/test_cable.py
Signed-off-by: JC <jumyungc@nvidia.com>
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
Signed-off-by: JC <jumyungc@nvidia.com>
Signed-off-by: JC <jumyungc@nvidia.com>
Signed-off-by: JC <jumyungc@nvidia.com>
Signed-off-by: JC <jumyungc@nvidia.com>
Signed-off-by: JC <jumyungc@nvidia.com>
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@newton/_src/sim/articulation.py`:
- Around line 1382-1549: Update eval_inverse_dynamics_force() to reject models
containing JointType.CABLE, matching the existing guard in
Model.inverse_dynamics(). Perform this validation before launching
eval_articulation_inverse_dynamics_force_kernel so CABLE models cannot produce
an invalid tau.
In `@newton/_src/sim/builder.py`:
- Around line 4414-4424: Update the xform remapping logic in
ModelBuilder.add_builder() so world-root CABLE joints are handled alongside FREE
joints. Recompute their 7-DOF joint_q pose using the transformed parent/child
world poses, keeping joint_q consistent with joint_X_p for add_world() and
replicate() copies.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 9f4dc073-6b63-4141-8d96-49b7e543fccc
📒 Files selected for processing (7)
CHANGELOG.mddocs/concepts/articulations.rstnewton/_src/sim/articulation.pynewton/_src/sim/builder.pynewton/_src/sim/enums.pynewton/_src/sim/model.pynewton/_src/solvers/vbd/solver_vbd.py
🚧 Files skipped from review as they are similar to previous changes (3)
- docs/concepts/articulations.rst
- newton/_src/sim/enums.py
- CHANGELOG.md
There was a problem hiding this comment.
Caution
Inline review comments failed to post. This is likely due to GitHub's internal server error or limits when posting large numbers of comments. If you are seeing this consistently it is likely a permissions issue. Please check "Moderation" -> "Code review limits" under your organization settings.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@newton/_src/sim/articulation.py`:
- Around line 1382-1549: Update eval_inverse_dynamics_force() to reject models
containing JointType.CABLE, matching the existing guard in
Model.inverse_dynamics(). Perform this validation before launching
eval_articulation_inverse_dynamics_force_kernel so CABLE models cannot produce
an invalid tau.
In `@newton/_src/sim/builder.py`:
- Around line 4414-4424: Update the xform remapping logic in
ModelBuilder.add_builder() so world-root CABLE joints are handled alongside FREE
joints. Recompute their 7-DOF joint_q pose using the transformed parent/child
world poses, keeping joint_q consistent with joint_X_p for add_world() and
replicate() copies.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 9f4dc073-6b63-4141-8d96-49b7e543fccc
📒 Files selected for processing (7)
CHANGELOG.mddocs/concepts/articulations.rstnewton/_src/sim/articulation.pynewton/_src/sim/builder.pynewton/_src/sim/enums.pynewton/_src/sim/model.pynewton/_src/solvers/vbd/solver_vbd.py
🚧 Files skipped from review as they are similar to previous changes (3)
- docs/concepts/articulations.rst
- newton/_src/sim/enums.py
- CHANGELOG.md
🛑 Comments failed to post (2)
newton/_src/sim/articulation.py (1)
1382-1549: 🗄️ Data Integrity & Integration | 🟡 Minor | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash set -euo pipefail # Map relevant symbols and locations. rg -n "CABLE|inverse_dynamics_force|eval_inverse_dynamics_force|inverse_dynamics\(" newton/_src/sim/articulation.py # Show surrounding context for the relevant sections, capped to reasonable line ranges. nl -ba newton/_src/sim/articulation.py | sed -n '1,260p' nl -ba newton/_src/sim/articulation.py | sed -n '1320,1605p'Repository: newton-physics/newton
Length of output: 1123
🏁 Script executed:
#!/bin/bash set -euo pipefail # Inspect any model/container-level guards for CABLE joints. rg -n "CABLE|container|guard|reject|unsupported" newton/_src/sim -g '*.py' # Read the inverse_dynamics path in the same file if present. nl -ba newton/_src/sim/articulation.py | sed -n '900,1180p'Repository: newton-physics/newton
Length of output: 6321
🏁 Script executed:
#!/bin/bash set -euo pipefail rg -n "CABLE|inverse_dynamics_force|eval_inverse_dynamics_force|inverse_dynamics\(" newton/_src/sim/articulation.py nl -ba newton/_src/sim/articulation.py | sed -n '1320,1605p' nl -ba newton/_src/sim/articulation.py | sed -n '900,1180p'Repository: newton-physics/newton
Length of output: 1123
🏁 Script executed:
#!/bin/bash set -euo pipefail # Inspect the model guard and call sites for eval_inverse_dynamics_force. sed -n '2178,2218p' newton/_src/sim/model.py printf '\n---\n' rg -n "eval_inverse_dynamics_force\(" newton/_src/sim -g '*.py'Repository: newton-physics/newton
Length of output: 2196
🏁 Script executed:
#!/bin/bash set -euo pipefail rg -n "eval_inverse_dynamics_force" newton -g '*.py' printf '\n--- exports ---\n' sed -n '1,220p' newton/__init__.pyRepository: newton-physics/newton
Length of output: 6478
Guard
eval_inverse_dynamics_force()against CABLE models.Model.inverse_dynamics()blocks CABLE, but this exported helper can still be called directly on a CABLE model and will leave the 6-DOF root wrench in the parent frame, producing a wrongtau. Add the sameJointType.CABLEcheck here.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/_src/sim/articulation.py` around lines 1382 - 1549, Update eval_inverse_dynamics_force() to reject models containing JointType.CABLE, matching the existing guard in Model.inverse_dynamics(). Perform this validation before launching eval_articulation_inverse_dynamics_force_kernel so CABLE models cannot produce an invalid tau.newton/_src/sim/builder.py (1)
4414-4424: 🗄️ Data Integrity & Integration | 🟡 Minor | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash set -euo pipefail # Inspect the relevant builder and articulation code paths. git grep -n "JointType.CABLE\|JointType.FREE\|add_builder\|replicate\|add_world\|joint_q_start\|joint_X_p" -- newton/_src/sim/builder.py newton/_src/sim # Show the surrounding code for the reported line range and the xform-offset branch. sed -n '4388,4445p' newton/_src/sim/builder.py sed -n '3535,3605p' newton/_src/sim/builder.py # Find cable-related helpers/tests to see how world-anchored cable joints are used. git grep -n "add_joint_cable\|cable" -- newton testsRepository: newton-physics/newton
Length of output: 50378
🏁 Script executed:
#!/bin/bash set -euo pipefail # Read only the relevant slices around add_builder/add_joint_cable and the FK docs. sed -n '3450,3615p' newton/_src/sim/builder.py sed -n '4947,5175p' newton/_src/sim/builder.py sed -n '4800,4845p' newton/_src/sim/builder.py sed -n '500,530p' newton/_src/sim/articulation.py # Check whether any tests or helpers explicitly cover world-anchored cable replication. git grep -n "add_joint_cable(parent=-1\|replicate(.*cable\|add_world(.*cable\|world-anchored cable\|CABLE" -- tests newton/_src | head -n 200Repository: newton-physics/newton
Length of output: 28101
Extend the xform remap to world-root
CABLEjoints inModelBuilder.add_builder()
CABLEjoints use the same 7-DOFjoint_qpose layout asFREE, so copying a world-root cable with a non-identityxformleavesjoint_qstale whilejoint_X_pis offset. That breakseval_fkforadd_world()/replicate()copies. Extend theFREEbranch to coverCABLEas well, or document that world-root cables cannot be replicated with transforms.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/_src/sim/builder.py` around lines 4414 - 4424, Update the xform remapping logic in ModelBuilder.add_builder() so world-root CABLE joints are handled alongside FREE joints. Recompute their 7-DOF joint_q pose using the transformed parent/child world poses, keeping joint_q consistent with joint_X_p for add_world() and replicate() copies.
Signed-off-by: JC <jumyungc@nvidia.com>
There was a problem hiding this comment.
🧹 Nitpick comments (2)
newton/_src/solvers/vbd/solver_vbd.py (1)
1307-1312: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winUse prefix-first names for the stiffness locals.
Rename
ke_stretchandke_bendtostretch_keandbend_keso the semantic qualifier comes first.As per coding guidelines, use prefix-first naming for autocomplete.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/_src/solvers/vbd/solver_vbd.py` around lines 1307 - 1312, Rename the local stiffness variables in the joint initialization block from ke_stretch and ke_bend to stretch_ke and bend_ke, and update all references in the assignments to joint_k_max_np and joint_k_init_np accordingly.Source: Coding guidelines
newton/tests/test_import_usd_deformable_cable.py (1)
242-246: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick winAssert the complete six-DOF material block.
This test checks only the first linear and angular entries, so incorrect values in the other four entries would go unnoticed. Assert all three linear and all three angular entries to match the documented replication contract.
Proposed test adjustment
- self.assertAlmostEqual(ke[dof0], expected_stretch, delta=expected_stretch * 1e-3) - self.assertAlmostEqual(ke[dof0 + 3], expected_bend, delta=expected_bend * 1e-3) + for i in range(3): + self.assertAlmostEqual(ke[dof0 + i], expected_stretch, delta=expected_stretch * 1e-3) + self.assertAlmostEqual(ke[dof0 + 3 + i], expected_bend, delta=expected_bend * 1e-3)🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/tests/test_import_usd_deformable_cable.py` around lines 242 - 246, Extend the assertions in the material-check section of the cable test around builder.joint_target_ke so all six entries for j0 are validated: the three linear DOFs must match expected_stretch and the three angular DOFs must match expected_bend, using the existing tolerance convention.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@newton/_src/solvers/vbd/solver_vbd.py`:
- Around line 1307-1312: Rename the local stiffness variables in the joint
initialization block from ke_stretch and ke_bend to stretch_ke and bend_ke, and
update all references in the assignments to joint_k_max_np and joint_k_init_np
accordingly.
In `@newton/tests/test_import_usd_deformable_cable.py`:
- Around line 242-246: Extend the assertions in the material-check section of
the cable test around builder.joint_target_ke so all six entries for j0 are
validated: the three linear DOFs must match expected_stretch and the three
angular DOFs must match expected_bend, using the existing tolerance convention.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 1c2a9b11-e5e3-436b-b46c-7ae1b61315bf
📒 Files selected for processing (10)
CHANGELOG.mddocs/concepts/articulations.rstnewton/_src/sim/articulation.pynewton/_src/sim/builder.pynewton/_src/sim/enums.pynewton/_src/sim/model.pynewton/_src/solvers/vbd/solver_vbd.pynewton/tests/test_cable.pynewton/tests/test_import_usd_deformable_cable.pynewton/tests/test_inverse_dynamics.py
🚧 Files skipped from review as they are similar to previous changes (2)
- newton/_src/sim/enums.py
- newton/_src/sim/model.py
Signed-off-by: JC <jumyungc@nvidia.com>
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
newton/_src/sim/builder.py (1)
4380-4410: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick winCABLE target_pos values for linear/angular axes are silently dropped.
For
BALL/FREE/DISTANCE,add_jointcopies each axis'sdim.target_posintojoint_target_q(linear axes directly, angular axes either via quaternion composition or raw copy) before setting the quat-w slot. The newelif joint_type == JointType.CABLE:branch only sets the quat-w padding to1.0and never writes any of the 6 linear/angulardim.target_posvalues intojoint_target_q. Any caller using the genericadd_joint(JointType.CABLE, ...)API with non-zerotarget_poson the linear/angular axes will have those targets silently discarded instead of raising an error or being honored, unlike every other multi-axis joint type.add_joint_cableitself never setstarget_poson the generatedJointDofConfigs, so this doesn't manifest through that helper today, but it's a latent correctness gap in the generaladd_jointcontract for CABLE.🐛 Proposed fix: fold CABLE into the FREE/DISTANCE branch
- if joint_type == JointType.BALL or joint_type == JointType.FREE or joint_type == JointType.DISTANCE: + if ( + joint_type == JointType.BALL + or joint_type == JointType.FREE + or joint_type == JointType.DISTANCE + or joint_type == JointType.CABLE + ): if joint_type == JointType.BALL: quat_offset = target_q_offset else: for i, dim in enumerate(linear_axes): self.joint_target_q[target_q_offset + i] = dim.target_pos quat_offset = target_q_offset + 3 import newton # noqa: PLC0415 if newton.use_coord_layout_targets: qx, qy, qz, qw = self._quat_from_axis_targets( angular_axes[0].target_pos, angular_axes[1].target_pos, angular_axes[2].target_pos, ) self.joint_target_q[quat_offset + 0] = qx self.joint_target_q[quat_offset + 1] = qy self.joint_target_q[quat_offset + 2] = qz self.joint_target_q[quat_offset + 3] = qw else: for i, dim in enumerate(angular_axes): self.joint_target_q[quat_offset + i] = dim.target_pos self.joint_target_q[quat_offset + 3] = 1.0 - elif joint_type == JointType.CABLE: - self.joint_target_q[target_q_offset + 6] = 1.0 elif joint_type != JointType.FIXED:🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/_src/sim/builder.py` around lines 4380 - 4410, Update the add_joint branch handling JointType.CABLE so it follows the same target_pos copying behavior as the FREE/DISTANCE multi-axis path: write all linear and angular axis target_pos values into joint_target_q, then set the quaternion padding component to 1.0. Preserve the existing BALL quaternion handling and CABLE target layout.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@newton/_src/sim/builder.py`:
- Around line 4380-4410: Update the add_joint branch handling JointType.CABLE so
it follows the same target_pos copying behavior as the FREE/DISTANCE multi-axis
path: write all linear and angular axis target_pos values into joint_target_q,
then set the quaternion padding component to 1.0. Preserve the existing BALL
quaternion handling and CABLE target layout.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 412640d4-49d0-4bff-9895-53de82f3753c
📒 Files selected for processing (6)
CHANGELOG.mddocs/concepts/articulations.rstnewton/_src/sim/builder.pynewton/_src/sim/enums.pynewton/_src/solvers/vbd/solver_vbd.pynewton/tests/test_cable.py
🚧 Files skipped from review as they are similar to previous changes (4)
- CHANGELOG.md
- docs/concepts/articulations.rst
- newton/_src/sim/enums.py
- newton/_src/solvers/vbd/solver_vbd.py
Signed-off-by: JC <jumyungc@nvidia.com>
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
newton/tests/test_model.py (1)
1663-1688: 📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick winAlign the docstring with
ModelBuilder._quat_from_axis_targets
newton/tests/test_model.py:1666-1688still references:meth:\ModelBuilder._quat_from_euler_zyx`, but the test and builder useModelBuilder._quat_from_axis_targets`. Update the docstring so it matches the helper the assertion actually exercises.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@newton/tests/test_model.py` around lines 1663 - 1688, Update the test docstring describing the Flag=True coordinate conversion to reference ModelBuilder._quat_from_axis_targets instead of ModelBuilder._quat_from_euler_zyx, matching the helper used by expected_quat and the exercised builder behavior.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@newton/tests/test_model.py`:
- Around line 1663-1688: Update the test docstring describing the Flag=True
coordinate conversion to reference ModelBuilder._quat_from_axis_targets instead
of ModelBuilder._quat_from_euler_zyx, matching the helper used by expected_quat
and the exercised builder behavior.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 63790f77-9392-46aa-9dd8-4e883912f7b5
📒 Files selected for processing (6)
newton/_src/sim/articulation.pynewton/_src/sim/builder.pynewton/_src/solvers/vbd/solver_vbd.pynewton/tests/test_import_usd_deformable_cable.pynewton/tests/test_inverse_dynamics.pynewton/tests/test_model.py
🚧 Files skipped from review as they are similar to previous changes (3)
- newton/tests/test_import_usd_deformable_cable.py
- newton/_src/solvers/vbd/solver_vbd.py
- newton/_src/sim/builder.py
Signed-off-by: JC <jumyungc@nvidia.com>
1 participant
Description
This PR gives JointType.CABLE a standard kinematic state layout:
Cable joints now use the existing relative-pose FK, IK, and Jacobian paths, making them easier to consume from general kinematic and reduced-coordinate
workflows.
Details
SolverVBD continues to advance maximal body_q/body_qd state. When joint state is needed after VBD stepping, eval_ik synchronizes it before eval_fk or other
reduced-coordinate consumers use it.
Checklist
CHANGELOG.mdhas been updated (if user-facing change)Test plan
Bug fix
Steps to reproduce:
Minimal reproduction:
New feature / API change
Summary by CodeRabbit
Summary
New Features
joint_q) and 6-DoF relative twist (joint_qd).Bug Fixes
Tests
Documentation