unisim.backend.mujoco.state_commit

Prepared writes consumed by the single MuJoCo native state submitter.

These adapter-local values describe an already validated transaction. They do not inspect assets, call the engine, or impose an entity layout on model-file models with arbitrary valid MuJoCo topology and transmissions.

Functions

selected_rows(value, num_envs)

state_values(value, shape, dtype, name)

Classes

MocapWrite

MocapWrite(index: 'int', poses: 'np.ndarray')

ModelWrite

ModelWrite(field: 'str', values: 'np.ndarray', column: 'int | None' = None)

StateCommitPlan

Owned prepared rows; public callers never submit this internal value.
class unisim.backend.mujoco.state_commit.ModelWrite[source]

Bases: object

ModelWrite(field: ‘str’, values: ‘np.ndarray’, column: ‘int | None’ = None)

Parameters:
field: str
values: ndarray
column: int | None = None
__init__(field, values, column=None)
Parameters:
class unisim.backend.mujoco.state_commit.MocapWrite[source]

Bases: object

MocapWrite(index: ‘int’, poses: ‘np.ndarray’)

Parameters:
index: int
poses: ndarray
__init__(index, poses)
Parameters:
class unisim.backend.mujoco.state_commit.StateCommitPlan[source]

Bases: object

Owned prepared rows; public callers never submit this internal value.

reset_world preserves the historical full-reset lifecycle; false preserves every unspecified entity/environment channel. defaults is used only by the explicit reset-to-initial-state entry point.

Parameters:
rows: ndarray
qpos: ndarray
qvel: ndarray
qpos_columns: tuple[int, ...]
qvel_columns: tuple[int, ...]
reset_world: bool = False
mocap: tuple[MocapWrite, ...] = ()
control_clear: tuple[int, ...] = ()
activation_clear: tuple[int, ...] = ()
force_dof_clear: tuple[int, ...] = ()
force_body_clear: tuple[int, ...] = ()
model_writes: tuple[ModelWrite, ...] = ()
defaults: tuple[tuple[str, ndarray], ...] = ()
control_values: ndarray | None = None
activation_values: ndarray | None = None
__init__(rows, qpos, qvel, qpos_columns, qvel_columns, reset_world=False, mocap=(), control_clear=(), activation_clear=(), force_dof_clear=(), force_body_clear=(), model_writes=(), defaults=(), control_values=None, activation_values=None)
Parameters:
unisim.backend.mujoco.state_commit.selected_rows(value, num_envs)[source]
Parameters:
Return type:

ndarray

unisim.backend.mujoco.state_commit.state_values(value, shape, dtype, name)[source]
Parameters:
Return type:

ndarray