feat(manipulation): add heuristic grasp provider

[ruthwikdasyam]

Contribution path

• Small, focused change extracted from the pick-and-place exploration work.

Problem

Pick-and-place needs a lightweight grasp provider that does not require the optional GraspGenX runtime.

Solution

• Add HeuristicGraspModule, implementing the existing GraspGenSpec.
• Produce one deterministic top-down parallel-jaw proposal from a segmented object point cloud.
• Validate point-cloud metadata and geometry.
• Register the module and add focused unit coverage.

How to Test

TMPDIR=/dev/shm uv run --no-sync pytest dimos/manipulation/grasping/test_heuristic_grasp.py dimos/robot/test_all_blueprints_generation.py -q
TMPDIR=/dev/shm uv run --no-sync mypy dimos/manipulation/grasping/heuristic_grasp.py
TMPDIR=/dev/shm uv run --no-sync ruff format --check dimos/manipulation/grasping/heuristic_grasp.py dimos/manipulation/grasping/test_heuristic_grasp.py
TMPDIR=/dev/shm uv run --no-sync ruff check dimos/manipulation/grasping/heuristic_grasp.py dimos/manipulation/grasping/test_heuristic_grasp.py

AI assistance

OpenCode with gpt-5.6-terra assisted with implementation and validation. The author will review the public method descriptions before they are added.

Checklist

• I have read and approved the CLA.

[codecov]

❌ 2 Tests Failed:

Tests completed	Failed	Passed	Skipped
4817	2	4815	87

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

dimos.e2e_tests.test_manipulation_planning_groups::test_dual_arm_plans_and_dispatches_both_arms_through_control_coordinator

Stack Traces | 120s run time

lcm_spy = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x758352ac7f20>
start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350717100>

    def test_dual_arm_plans_and_dispatches_both_arms_through_control_coordinator(
        lcm_spy: LcmSpy,
        start_blueprint: Callable[..., DimosCliCall],
    ) -> None:
        """Plan one generated plan over both arms and dispatch through one trajectory task."""
>       _start_openarm_mock_planner(start_blueprint, lcm_spy)

lcm_spy    = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x758352ac7f20>
start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350717100>

dimos/e2e_tests/test_manipulation_planning_groups.py:190: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
dimos/e2e_tests/test_manipulation_planning_groups.py:151: in _start_openarm_mock_planner
    lcm_spy.wait_for_saved_topic(JOINT_STATE_TOPIC, timeout=120.0)
        lcm_spy    = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x758352ac7f20>
        start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350717100>
dimos/e2e_tests/lcm_spy.py:97: in wait_for_saved_topic
    wait_until(
        condition  = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350717a60>
        self       = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x758352ac7f20>
        timeout    = 120.0
        topic      = '/coordinator_joint_state#sensor_msgs.JointState'
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

predicate = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350717a60>

    def wait_until(
        predicate: Callable[[], bool],
        *,
        timeout: float,
        interval: float = 0.1,
        message: str | None = None,
    ) -> None:
        """Poll ``predicate`` until it returns truthy or ``timeout`` elapses."""
        deadline = time.monotonic() + timeout
        while time.monotonic() < deadline:
            if predicate():
                return
            time.sleep(interval)
>       raise TimeoutError(message or f"Timed out after {timeout}s waiting for condition")
E       TimeoutError: Timeout waiting for topic /coordinator_joint_state#sensor_msgs.JointState

deadline   = 4268088.689404055
interval   = 0.1
message    = 'Timeout waiting for topic /coordinator_joint_state#sensor_msgs.JointState'
predicate  = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350717a60>
timeout    = 120.0

.../utils/testing/waiting.py:35: TimeoutError

View the full list of 1 ❄️ flaky test(s)

dimos.e2e_tests.test_manipulation_planning_groups::test_single_arm_plans_and_executes_through_control_coordinator

Flake rate in main: 7.69% (Passed 12 times, Failed 1 times)

Stack Traces | 120s run time

lcm_spy = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x7583526115e0>
start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350716b60>

    def test_single_arm_plans_and_executes_through_control_coordinator(
        lcm_spy: LcmSpy,
        start_blueprint: Callable[..., DimosCliCall],
    ) -> None:
        """Plan with one arm and execute through its trajectory task."""
>       _start_openarm_mock_planner(start_blueprint, lcm_spy)

lcm_spy    = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x7583526115e0>
start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350716b60>

dimos/e2e_tests/test_manipulation_planning_groups.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
dimos/e2e_tests/test_manipulation_planning_groups.py:151: in _start_openarm_mock_planner
    lcm_spy.wait_for_saved_topic(JOINT_STATE_TOPIC, timeout=120.0)
        lcm_spy    = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x7583526115e0>
        start_blueprint = <function start_blueprint.<locals>.set_name_and_start at 0x758350716b60>
dimos/e2e_tests/lcm_spy.py:97: in wait_for_saved_topic
    wait_until(
        condition  = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350715260>
        self       = <dimos.e2e_tests.lcm_spy.LcmSpy object at 0x7583526115e0>
        timeout    = 120.0
        topic      = '/coordinator_joint_state#sensor_msgs.JointState'
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

predicate = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350715260>

    def wait_until(
        predicate: Callable[[], bool],
        *,
        timeout: float,
        interval: float = 0.1,
        message: str | None = None,
    ) -> None:
        """Poll ``predicate`` until it returns truthy or ``timeout`` elapses."""
        deadline = time.monotonic() + timeout
        while time.monotonic() < deadline:
            if predicate():
                return
            time.sleep(interval)
>       raise TimeoutError(message or f"Timed out after {timeout}s waiting for condition")
E       TimeoutError: Timeout waiting for topic /coordinator_joint_state#sensor_msgs.JointState

deadline   = 4267968.487287269
interval   = 0.1
message    = 'Timeout waiting for topic /coordinator_joint_state#sensor_msgs.JointState'
predicate  = <function LcmSpy.wait_for_saved_topic.<locals>.condition at 0x758350715260>
timeout    = 120.0

.../utils/testing/waiting.py:35: TimeoutError

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

[greptile-apps]

Greptile Summary

This change adds a deterministic, PCA-based heuristic grasp provider, focused proposal tests, and generated catalog registration. The rectangular-fixture test checks that the gripper approaches from above but does not check its yaw, so a rotation that places the jaws across the fixture's wide dimension can remain undetected.

T-Rex validation blocked

The focused test could not start because the Python package tqdm is missing. A narrowly scoped yaw-regression harness then could not import the grasp provider because the Python package zenoh is missing.

Confidence Score: 4/5

The provider's yaw behavior should receive explicit regression coverage before it is relied on for rectangular-object grasp planning.

The remaining finding is limited to test coverage: the current assertion constrains the downward approach direction but not the yaw-dependent jaw direction.

Files Needing Attention: dimos/manipulation/grasping/test_heuristic_grasp.py needs a jaw-axis assertion for the non-square rectangular fixture.

T-Rex Logs

What T-Rex did

• T-Rex completed the requested verification, but its local artifact references were not uploaded.
• T-Rex produced proof for a posted P2 finding, and the reviewer comment contains the finding details.
• Validation was blocked before the relevant behavior could execute due to missing tqdm and zenoh during imports, and the attempt record is preserved.

_{Ran code and verified through T-Rex}

_{Reviews (1): Last reviewed commit: "feat(manipulation): add heuristic grasp ..." | Re-trigger Greptile}