[TomCC7]

Contribution path

• PRD: Manipulation SDK preview, discussion #3923.
• Stacked on feat(imitation): add native OpenYAM collection #3854.

Problem

• OpenYAM needs a runnable policy graph with optional Quest takeover.
• Robot composition should use the generic policy/control contract from its parent instead of introducing another learning-profile abstraction.

Solution

• Add openyam-lerobot-rollout for Quest-free execution and openyam-lerobot-quest-rollout for optional takeover.
• Compose the LeRobot module, wrist RGB camera, and a named low-priority trajectory task with canonical OpenYAM joints.
• Reuse the existing Quest tasks for the takeover graph.
• Keep robot-specific joint ordering and camera defaults in this Blueprint; remove the obsolete OpenYamLearningProfile dependency.
• Own the dual-arm regression assertions that require hardware-aware trajectory acceptance/rejection; the underlying canonical hardware-limit setup lives in fix(openyam): align learning stack with canonical model #3853.
• Inherit chunk execution, preflight, cancellation, button edges, and limits from refactor(imitation): isolate LeRobot policy runtime #3315. This PR only adds robot wiring, tests, registry entries, and usage docs.

API shape

openyam-lerobot-rollout
  policy + wrist camera + ControlCoordinator

openyam-lerobot-quest-rollout
  policy + wrist camera + Quest teleop + TeleopControlCoordinator

• Configure either graph through normal dimos run module options.
• Python clients find RolloutControlSpec and explicitly call preflight/start/stop.
• The attached rollout panel is added later in feat(imitation): add CLI learning workflow #3931. Stopping a client is not a policy-stop request.

How to Test

With a compatible checkpoint and configured arm/camera:

dimos --can-port follower_l run openyam-lerobot-rollout --policy.policy-path CHECKPOINT_DIR --policy.task "pick up the cube" --wristcamera.hardware.camera-index 0

• Start through RolloutControlSpec only after verifying preflight readiness and clearing the workspace.
• Use openyam-lerobot-quest-rollout to test Quest takeover.
• In the previous cleanup, 18 focused rollout/OpenYAM tests and scoped mypy passed; the generated registry is current.
• Current review cleanup: the dual-arm trajectory acceptance/rejection tests pass in the final 224-test focused suite. This layer adds no new rollout behavior during the cleanup.
• Bounded-motion rollout, takeover, and shutdown require a hardware retest. ABC and dual-arm rollout are out of scope.

AI assistance

• Review cleanup updated by OpenAI Codex (agent).
• OpenAI Codex (GPT-5) assisted extensively with implementation, tests, documentation, and stack cleanup.

Checklist

• I have read and approved the CLA.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

@@                        Coverage Diff                         @@
##           feat/openyam-native-collection    #3855      +/-   ##
==================================================================
+ Coverage                           76.76%   79.94%   +3.17%     
==================================================================
  Files                                1648     1580      -68     
  Lines                              159551   149594    -9957     
  Branches                            14197    13166    -1031     
==================================================================
- Hits                               122486   119599    -2887     
+ Misses                              33539    26463    -7076     
- Partials                             3526     3532       +6     

Components	Coverage Δ
Tests	95.13% <100.00%> (+3.36%)	⬆️

Flag	Coverage Δ
OS-macos-latest	75.50% <100.00%> (+0.01%)	⬆️
OS-ubuntu-24.04-arm	75.42% <100.00%> (+0.01%)	⬆️
OS-ubuntu-latest	76.07% <100.00%> (+0.01%)	⬆️
Py-3.10	76.07% <100.00%> (+0.01%)	⬆️
Py-3.11	75.47% <100.00%> (-0.58%)	⬇️
Py-3.12	76.07% <100.00%> (+0.01%)	⬆️
SelfHosted-Large	32.18% <58.75%> (+1.50%)	⬆️
SelfHosted-Linux	36.84% <58.75%> (+0.12%)	⬆️
SelfHosted-macOS	37.20% <58.75%> (+0.20%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
dimos/robot/all_blueprints.py	100.00% <ø> (ø)
...anipulators/openyam/blueprints/learning_rollout.py	100.00% <100.00%> (ø)
...os/robot/manipulators/openyam/blueprints/teleop.py	100.00% <100.00%> (ø)
...lators/openyam/blueprints/test_learning_rollout.py	100.00% <100.00%> (ø)

... and 187 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[github-actions]

go2 replay realtime

Details

Benchmark suite	Current: 9872000	Previous: 58acbee	Ratio
first frame wall	6.62 s	9.623 s	0.69
first frame cpu	13.616 s	21.264 s	0.64
run cpu	19.224 s	33.16 s	0.58
run cpu (user)	14.157 s	24.525 s	0.58
run cpu (system)	5.067 s	8.635 s	0.59
peak memory	2662.824 MB	2701.34 MB	0.99
peak threads	405 threads	408 threads	0.99
disk read	29.77 MB	19.965 MB	1.49
disk write	11.055 MB	11.285 MB	0.98
network (transport)	2732.816 MB	2725.519 MB	1.00
network (external rx)	0.114 MB	0.069 MB	1.65
network (external tx)	0.156 MB	0.346 MB	0.45

This comment was automatically generated by workflow using github-action-benchmark.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'go2 replay realtime'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.08.

Benchmark suite	Current: 9872000	Previous: 58acbee	Ratio
disk read	29.77 MB	19.965 MB	1.49
network (external rx)	0.114 MB	0.069 MB	1.65

This comment was automatically generated by workflow using github-action-benchmark.

[TomCC7]

Replaced by #4344 on a compliant branch at the identical head commit 9872000f7cbb394f531ce1bd1db89b23a9c8217d. The seven layer diffs and downstream PR IDs were verified unchanged in native stack #4345: #4343 → #4344 → #3921 → #3931 → #3942 → #4088 → #4156.

Closing this PR only for the user-approved branch naming migration. This branch and all review comments remain available; unresolved findings and existing main/CI blockers are not marked resolved. No behavior changes were included in this migration.