Pick the fiducial cube
======================

The office world's 50 mm cube carries six ``DICT_4X4_50`` markers, IDs 0–5.
Each marker's 36 mm outer black edge is 25 mm from the cube centre. This
example detects the visible faces with a calibrated wrist camera, prefers the
top face when available, and aligns the tool with the detected tag using the
nearest of four equivalent cube-face yaws. It approaches along the tag's ``-z``
axis to grasp the cube.
The nearby ChArUco board uses a different
marker dictionary, so its IDs cannot be confused with the cube's.
Short pauses after arm and gripper actions let the camera and grasp settle
before the next step.

.. tab-set::
    :sync-group: sdk-language

    .. tab-item:: Python
        :sync: python

        .. literalinclude:: scripts/vision-pick-fiducial-cube.py
           :language: python

    .. tab-item:: TypeScript
        :sync: typescript

        .. literalinclude:: scripts/vision-pick-fiducial-cube.ts
           :language: typescript