Task-generation pipelines

All pipelines live in maniguard/task_generation/ and share a BasePipeline runtime contract from pipeline_common.py. Each pipeline auto-discovers a support surface in the given scene, generates BDDL + ltl_safety.json at runtime, spawns objects, places the robot, runs gate checks, and executes an LTL-monitored rollout.

The cross-cutting Data flow page describes the four-stage architecture every pipeline shares: offline JSON pool generation (raycast scans + admission filters) → object selection → scene + surface selection → placement.

ManiGuard-Bench families

The released ManiGuard-Bench is these 6 families (200 base tasks). Each page states the task goal + safety and how the family's scenes are generated:

Family	Tasks	One-liner
Clutter pickup	55	Pick a named target out of a cluttered pack into the goal region (+ liquid subset)
Cabinet pickup	35	Open a drawer, place the target inside, close it
Lid transport	30	Put the lid on before lifting the container to the goal
Stack retrieve	28	Pull the bottom object out from under a stack without toppling it
Jar transport	26	Close a hinged jar before carrying it to the goal
Dusty transfer	26	Wipe a dusty pot clean, then transfer food into it with the tool

Generation infrastructure

Shared machinery every family builds on:

• Data flow — the four-stage architecture (offline pools → selection → surface → placement).
• Add a custom pipeline — author a new BasePipeline subclass.
• Empty-scene runner — synthesize a surface on a bare floor (used by the empty-scene families).
• Food transfer (base) — the transfer base that dusty transfer extends.

Additional families

Other families the pipeline can generate but that are not in the shipped bench (wet_transport, empty_invert).

Multi-scene benchmark

run_benchmark.py orchestrates a pipeline across multiple scenes, one subprocess per scene, with per-scene timeout and resume logic. The subprocess-per-scene model intentionally avoids long-lived Isaac Sim / renderer state and makes failed scenes easier to isolate and rerun.

conda activate behavior

python -m maniguard.task_generation.run_benchmark \
  --pipeline table \
  --scenes hall_conference_large \
  --episodes 1 --steps 300 --density medium --timeout 1800

Pipeline choices for --pipeline (keys of _PIPELINE_SCRIPTS): table (clutter), transfer, dusty_transfer, stack (+ stack_same / stack_flat / stack_receptacle), lid_transport, liquid_transport, wet_transport, jar_transport, cabinet_pickup. The empty-scene families (cabinet_pickup, jar_transport) are usually run via their own CLI with --task-id rather than the per-scene benchmark loop.

Dry-run (BDDL + LTL only, no simulator)

python -m maniguard.task_generation.clutter_scene_pipeline \
  --scene-model Benevolence_1_int --dry-run

Artifact contract

A successful scene run is expected to produce, at minimum:

File	Purpose
diagnostics.jsonl	Outcome signals: gate results, LTL status, density, active-object summary
scene_ep1.json	Frozen scene snapshot for replay, rerendering, later evaluation
stdout.log	Runtime trace for debugging
rollout_ep1.mp4	Canonical human-reviewable video

Gate vs LTL — what's the difference?

Two validation layers run sequentially:

• Gate checks (pre-rollout, structural): "Is this scene structurally valid enough to start?" Robot/target poses are finite, robot base near floor plane, selected mount pose collision-free, target inside reach band, plus per-pipeline extras (e.g. pack integrity for clutter). With --strict-gate, a gate failure aborts the episode.
• LTL safety rollout (during execution, semantic): "Does the scene remain safe and semantically intact over time?" combined_ltl from ltl_safety.json is evaluated step by step by TaskLTLMonitor.