[KrishnaH96]

Why

#4326 drives the bimanual OpenArm from two Feetech leaders with an OpenArm-specific module. #5177 made that module generic. This PR is the OpenArm half: the robot folder carries only the mapping, limits, velocity profile and blueprints. Supersedes #4326 together with #5177.

What

dimos/robot/manipulators/openarm/blueprints/

• mini_teleop.py: MAPPINGS (leader joint N drives arm/openarm_<side>_jointN inside the URDF limits, trigger servo closes the jaw), velocity profile 3/3/3/3/6/6/6 rad/s, 6 Hz leader filter, 10 ms leader tick. Blueprints teleop-openarm-mini (one machine), -leader (operator laptop, zenoh client) and -follower (robot, dials the router). --simulation swaps in the mock connection.
• mini_collection.py: openarm-mini-collection and openarm-mini-follower-collection add the recorder (coordinator_joint_state, arm_position_command, 16 joints each) and three 640x480 webcams.
• Tests for the mapping against the loaded URDF, task wiring, port selection, split deployment and the collection schema.
• Docs: docs/capabilities/manipulation/openarm_mini_teleop.md, linked from the OpenArm integration page.

No change under dimos/teleop/leader/ or dimos/control/.

How to test

uv sync --extra leader-teleop --extra manipulation
uv run pytest dimos/robot/manipulators/openarm/blueprints dimos/teleop/leader
dimos --simulation run teleop-openarm-mini --leaderteleopmodule.reader mock

With the leaders and the OpenArm (calibration files in ~/.local/state/dimos/teleop/leader/<side>/, trigger included):

dimos hardware feetech-leader joint-tui --side left --port /dev/serial/by-id/<left>
dimos run teleop-openarm-mini --leaderteleopmodule.port-left /dev/serial/by-id/<left> \
  --leaderteleopmodule.port-right /dev/serial/by-id/<right> --left-can-port can1 --right-can-port can2

Checks

• pytest (openarm blueprints + leader): 81 passed; codebase checks: 167 passed; ruff, mypy, deptry clean.
• Simulation: four modules deploy, joint_trajectory executes the mock leader at 50 Hz, both gripper tasks bound.
• Hardware, Alfred 2026-11-10, leaders on a laptop and follower on the Jetson: both arms and jaws follow; three episodes recorded with openarm-mini-follower-collection and prepared into a LeRobot v3 dataset (2,432 frames at 30 Hz, 16-dim state and action, three videos).
• Smoothness, from command-vs-state analysis of the recordings: the coordinator sharing a Python process with cameras and Viser lost a quarter of its 100 Hz ticks; with its own worker (feat(teleop): generic Feetech leader teleop module and coordinator port #5177) it runs at 100.0/s, wrist roughness halves and tracking error drops a third. The 6 Hz filter halves wrist roughness again for ~10 ms of lag; the 10 ms tick takes ~10 ms off wrist lag. cc's fix(damiao): consume the latest target on one SDK-owning thread #5180 was A/B tested on this rig and is independent of this PR.
• Known: the recorder's strict quality mode rejects an episode on a single dropped camera frame; the dataset above was prepared with the schema's quality mode set to fill. Raised with cc.

Checklist

• I have read and approved the CLA.

[github-actions]

go2 replay realtime (arm64)

Details

Benchmark suite	Current: 269b2df	Previous: 14c62ab	Ratio
peak memory	1794.664 MB	1762.355 MB	1.02
peak threads	368 threads	385 threads	0.96
network (transport)	2731.097 MB	2731.013 MB	1.00
disk write	1.84 MB	1.605 MB	1.15
instructions	138.728 G	138.791 G	1.00

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

[codecov]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
8173	1	8172	55

View the top 1 failed test(s) by shortest run time

experimental.gateway.tests.test_config::test_every_registered_blueprint_is_annotated

Stack Traces | 0.162s run time

def test_every_registered_blueprint_is_annotated() -> None:
        source = annotation_file.load()
        annotated = {name for robot in source["robots"].values() for name in robot["blueprints"]}
        missing = sorted(
            name
            for name in all_blueprints
            if name not in annotated and not annotation_file.HIDDEN_NAME.search(name)
        )
>       assert not missing, f"add these to experimental/gateway/annotations.yaml: {', '.join(missing)}"
E       AssertionError: add these to experimental/gateway/annotations.yaml: openarm-mini-collection, openarm-mini-follower-collection, teleop-openarm-mini, teleop-openarm-mini-follower, teleop-openarm-mini-leader
E       assert not ['openarm-mini-collection', 'openarm-mini-follower-collection', 'teleop-openarm-mini', 'teleop-openarm-mini-follower', 'teleop-openarm-mini-leader']

annotated  = {'a1z-planner-coordinator', 'alfred-keyboard-teleop', 'alfred-mls-nav', 'alfred-mls-nav-lidar', 'alfred-nav', 'coordinator-a1z', ...}
missing    = ['openarm-mini-collection', 'openarm-mini-follower-collection', 'teleop-openarm-mini', 'teleop-openarm-mini-follower', 'teleop-openarm-mini-leader']
source     = {'robots': {'a1z': {'blueprints': {'a1z-planner-coordinator': {'description': 'Motion planning and execution for a Gal...': {'label': 'Arms'}, 'dog': {'label': 'Dogs'}, 'drone': {'label': 'Drones'}, 'humanoid': {'label': 'Humanoids'}, ...}}

.../gateway/tests/test_config.py:255: AssertionError

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}