Task batching

Group adjacent moveto and end_effector tasks for one MTC plan while preserving task order. Every other task type forms a single unit.

beambot.core.task_batching · Source

Task execution · Guide · Symbol index

How it is used

After parsing, the orchestrator passes the ordered task list to group_into_batches(). Adjacent moveto and end_effector tasks become one MTC planning unit; a tool exchange, pause, vision task, or other task separates those runs.

The function only groups dictionaries. It does not validate, plan, or execute them. A group can be marked batched even when it contains just one eligible task; disabling batching makes every task a single unit.

Start here

Callable	Purpose
group_into_batches	Group adjacent motion and gripper tasks while preserving order. Other task types remain single units; even one eligible task forms a `batched` unit.

Example

For grouping purposes, only task_type is inspected. These abbreviated dictionaries illustrate the order; a complete executable task document needs each operation’s required fields.

from beambot.core.task_batching import group_into_batches

tasks = [
    {"task_type": "moveto"},
    {"task_type": "end_effector"},
    {"task_type": "pause"},
    {"task_type": "moveto"},
]
groups = group_into_batches(tasks)
assert [(kind, len(items)) for kind, items in groups] == [
    ("batched", 2),  # arm move and gripper state share one plan
    ("single", 1),   # pause remains separate
    ("batched", 1),  # the next motion starts a new planning unit
]

Functions

`group_into_batches()`

Group adjacent motion and gripper tasks while preserving order. Other task types remain single units; even one eligible task forms a batched unit.

Entry point · Source

group_into_batches(
    tasks: list[dict[str, Any]],
    enabled: bool = True,
) -> list[tuple[str, list[dict[str, Any]]]]

Parameter	Meaning
`tasks`	Ordered task dictionaries; returned groups retain the original task objects.
`enabled`	Enable grouping; false returns one single unit per task.

Returns: Ordered list of (batched_or_single, task_list) tuples; empty input produces an empty list.

Errors: Malformed non-mapping task elements raise during .get access; task fields are otherwise not validated.

Side effects: None; input task dictionaries are not changed.

Referenced in source (1)

MTCOrchestratorServer._execute — beambot.action_servers.orchestrator