Repository navigation
Rename UrdfManager to UsdManager - #5111
jeff-hykin wants to merge 100 commits into
Conversation
USD is the description format dimos is moving to, and this module is the thing that answers for a robot's description, so it takes the name of where it is going rather than of the file format it currently reads. Mechanical, with the docstrings rewritten to be honest about the half-way state: the model it loads is still authored as URDF and its tf and file serving are still URDF-shaped, while get_usd / get_scene are the USD side, converting on the way out. dimos/robot/urdf_manager.py -> usd_manager.py dimos/robot/urdf_manager_point_clearing.py -> usd_manager_point_clearing.py dimos/robot/test_urdf_manager.py -> test_usd_manager.py UrdfManager / UrdfManagerConfig / UrdfManagerPointClearing[Config] -> Usd* all_blueprints keys: urdf-manager[-point-clearing] -> usd-manager[-point-clearing] Nothing outside the repo referenced the registry keys, and the lowercased blueprint-config key (urdfmanager) appeared nowhere, so the only user-visible breaks are the registry keys themselves and the RPC topic, which move together with the deployment. No alias is left behind: a rename with a compatibility alias is not a rename. Note this will conflict with the open PRs that touch urdf_manager.py (#5032, #5020, #5002).
There was a problem hiding this comment.
go2 replay realtime (arm64)
Details
| Benchmark suite | Current: a71a821 | Previous: 5bc57a0 | Ratio |
|---|---|---|---|
peak memory |
1745.973 MB |
1776.91 MB |
0.98 |
peak threads |
376 threads |
376 threads |
1 |
network (transport) |
2733.761 MB |
2730.972 MB |
1.00 |
disk write |
1.875 MB |
1.59 MB |
1.18 |
instructions |
137.074 G |
138.52 G |
0.99 |
This comment was automatically generated by workflow using github-action-benchmark.
Codecov Report❌ Patch coverage is
@@ Coverage Diff @@
## jeff/feat/usd_manager_livox #5111 +/- ##
===============================================================
- Coverage 82.47% 82.46% -0.01%
===============================================================
Files 1872 1872
Lines 184719 184720 +1
Branches 15505 15505
===============================================================
- Hits 152341 152338 -3
- Misses 28392 28396 +4
Partials 3986 3986
Flags with carried forward coverage won't be shown. Click here to find out more.
... and 5 files with indirect coverage changes 🚀 New features to boost your workflow:
|
Resolved in favour of the incoming standard-USD content (displayName, mesh references, no custom attributes) with this branch's UsdManager naming applied over it.
Resolved in favour of the incoming [asset_root, usd] answer and the removed get_scene deadline, with this branch's UsdManager naming applied over it.
…d_manager_livox Straightforward merge; the UsdManager naming is applied over the test lines the merge brought in under their old spelling.
Resolved in favour of the incoming tests, with the UsdManager naming applied over them.
…ger_livox Resolved in favour of the incoming versions, with the UsdManager naming applied over them.
assets/README.md conflicted only because this branch had renamed mentions the file no longer makes; it goes back to main's version, which names no manager at all.
Resolved in favour of the incoming tests with the UsdManager naming applied over them.
Took the trimmed versions and re-applied the UsdManager naming over them.
[Medium impact] The rename reaches module selection and robot blueprint imports across several robot workflows. The PR appears safe to merge, with non-blocking fixes for stale test types and removed documentation. FindingsSummaryRenames
Reviews (1) · Last reviewed commit: "Merge branch 'jeff/feat/usd_manager_livo..." · Reviewed by Greptile |
| CircularMountError, | ||
| Mount, | ||
| UrdfManager, | ||
| UsdManager, |
There was a problem hiding this comment.
| A TwistStamped on ``ee_twist_command`` moves the gripper tip at that | ||
| velocity (m/s, rad/s, world frame); a zero twist, or none for 0.3 s, stops it. | ||
| """ | ||
| pass |
There was a problem hiding this comment.
Blueprint operating instructions disappear
Replacing HybridTeleopCoordinator's description with pass removes the jog frame, units, and stop timeout. The same cleanup removes the hold protocol and the reason for the larger TF tolerances in manipulation_mapping. These details help teammates use and change the blueprints correctly. Restore the descriptions and comments with updated manager names.
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!
…_rename # Conflicts: # dimos/robot/test_usd_manager.py
This PR has been flagged CONFLICTING for twelve hours and it is not. Measured against its
own base, jeff/feat/usd_manager_livox:
base is an ancestor of head yes
ahead / behind 92 / 0
git merge-tree conflicts 0
A fast-forward cannot conflict. The flag is GitHub's, and the cost of it is real: GitHub
will not build a test merge for a PR it believes is dirty, so this PR and the one based on
it have never had a pull_request CI run -- their only green came from a workflow_dispatch,
where the self-hosted suite is skipped.
Nothing here changes a byte of the tree. It exists to make GitHub recompute the merge state.
main grew hybrid_teleop_sim's collection session while this branch deleted that file's docstrings; the code is main's and the docstrings stay deleted.
main's Piper teleop blueprint imports dimos.robot.urdf_manager, which this branch renames, so the merge left a blueprint that cannot be imported at all -- and the webxr blueprint that pulls it in, and the a1z test that pulls that in. The manager's own test still spelled the old name in two annotations, which only survived because they are never evaluated.
Every conflict was the rename against a restored docstring: kept the UsdManager names and the docstring main wrote, and pointed MULTI_DOF_JOINT_TYPES's import at its new home in dimos.robot.joint_tf.
Pim, on #5102: "don't delete prexisting comments/docstrings". The rename touched thirteen files that already existed on main, and this branch's stripping pass took their comments with it -- 29 docstrings and 24 comments, none of them about UrdfManager or UsdManager. The xarm teleop ones are the costly kind: which task drives which joint, what an app's twist timeout is for, that drive_joint is open at 0 rad. Restored from main by qualified name, so only the ones whose code is still here came back. One comment the tool could not place was checked by hand and is not lost: the "return every sample" note moved with max_candidates into grasp.py.
The only conflict was test_usd.py, renamed here and deleted there; took the deletion. The rename travels with the file onto #4911, which is where it lands.
…ve up Jeff asked for #5109 to be empty, with the lidar work in a later PR. This is the PR that reads it: `dimos/sim2/livox.py` here imports `mid360_usd`, and without these two files this branch does not import at all (`ImportError: cannot import name 'mid360_usd'`). The files arrive renamed, since this branch sits above #5111: the mount-contract test now imports `dimos.robot.usd_manager` and is named for it.
Renames
UrdfManagertoUsdManager, since the module now serves USD.