feat(a1z): a1z learning workflow

[TomCC7]

Contribution path

• Stacked follow-up to feat: add experimental LeRobot policy runtime #3315, porting the remaining applicable pieces of Adventure x #3138.
• Base branch: feat/experimental-lerobot-policy

Problem

PR #3315 ports the shared LeRobot policy infrastructure, but the A1Z-specific teach, replay, dataset preparation, and policy-running workflow from #3138 is still missing.

Solution

• Add safe A1Z episode loading, validation, smoothing, time scaling, and replay trajectory generation.
• Add A1Z teach, replay, and policy blueprint factories with lazy optional LeRobot imports.
• Add the teach, replay, and run-policy CLI commands.
• Add episode undo support and a Galaxea A1Z data-preparation profile.
• Publish camera metadata immediately and make recorder shutdown deterministic with rate-limited warnings.
• Document the A1Z learning workflow and add focused unit coverage.

How to test

• Focused A1Z, camera, episode monitor, data-preparation, and recorder tests pass.
• Blueprint validation and generation checks pass.
• Documentation link checks pass.
• Mypy passes for the changed implementation.
• Pre-commit passes.

AI assistance

Implemented and reviewed with OpenAI Codex.

Checklist

• CLA signed, if required
• Tests added or updated
• Documentation added or updated

[codecov]

Codecov Report

❌ Patch coverage is 63.43085% with 275 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
dimos/robot/manipulators/a1z/cli.py	10.18%	237 Missing and 1 partial ⚠️
dimos/robot/manipulators/a1z/teach_replay.py	81.41%	15 Missing and 14 partials ⚠️
dimos/memory2/module.py	80.95%	3 Missing and 1 partial ⚠️
dimos/imitation/dataprep/core.py	60.00%	0 Missing and 2 partials ⚠️
...imos/robot/manipulators/a1z/blueprints/learning.py	94.44%	2 Missing ⚠️

@@                         Coverage Diff                          @@
##           feat/experimental-lerobot-policy    #3318      +/-   ##
====================================================================
- Coverage                             75.23%   75.19%   -0.05%     
====================================================================
  Files                                  1146     1152       +6     
  Lines                                110632   111379     +747     
  Branches                               9936    10006      +70     
====================================================================
+ Hits                                  83237    83753     +516     
- Misses                                24544    24750     +206     
- Partials                               2851     2876      +25     

Flag	Coverage Δ
OS-ubuntu-24.04-arm	69.26% <63.43%> (+0.14%)	⬆️
OS-ubuntu-latest	71.17% <63.43%> (-0.01%)	⬇️
Py-3.10	71.17% <63.43%> (-0.01%)	⬇️
Py-3.11	71.17% <63.43%> (-0.02%)	⬇️
Py-3.12	71.16% <63.43%> (-0.02%)	⬇️
Py-3.13	71.17% <63.43%> (-0.01%)	⬇️
Py-3.14	71.17% <63.43%> (-0.02%)	⬇️
Py-3.14t	71.17% <63.43%> (-0.02%)	⬇️
SelfHosted-Large	29.05% <22.87%> (-0.04%)	⬇️
SelfHosted-Linux	35.77% <22.87%> (-0.09%)	⬇️

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

Files with missing lines	Coverage Δ
dimos/hardware/sensors/camera/module.py	82.60% <100.00%> (+22.31%)	⬆️
dimos/hardware/sensors/camera/test_module.py	100.00% <100.00%> (ø)
dimos/imitation/collection/episode_monitor.py	90.08% <100.00%> (+2.23%)	⬆️
dimos/imitation/collection/test_episode_monitor.py	99.15% <100.00%> (+0.27%)	⬆️
dimos/imitation/dataprep/test_config_profiles.py	100.00% <100.00%> (ø)
dimos/imitation/dataprep/test_core.py	99.39% <100.00%> (+0.01%)	⬆️
dimos/memory2/test_recorder.py	100.00% <100.00%> (ø)
dimos/robot/manipulators/a1z/test_setup.py	100.00% <100.00%> (ø)
dimos/robot/manipulators/a1z/test_teach_replay.py	100.00% <100.00%> (ø)
dimos/imitation/dataprep/core.py	86.59% <60.00%> (-0.71%)	⬇️
... and 4 more

... and 4 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.