run-data-collection

name	run-data-collection
description	Launch a data_collection automated trajectory-collection task on a GPU host using the `geniesim autocollect run` CLI verb (which wraps scripts/run_data_collection.sh: docker run -d + in-container server+client). Trigger: when the user asks to "采集数据", "跑数据采集", "run data collection", "collect a task", "生产轨迹", "launch a tasks/geniesim_2025/<...>.json", or wants to produce agibot-format episodes from a data_collection task template.
license	MPL-2.0
metadata	{"author":"genie-sim","version":"1.0"}

When to Use

• User wants to produce trajectory episodes from a data_collection task template on a workstation with Docker + an NVIDIA GPU.
• User references a task under source/data_collection/tasks/.

Do not use for:

• Running a benchmark/evaluation task → run-benchmark.
• Just listing/inspecting tasks → geniesim autocollect list directly.

Critical Patterns

1. run is host-orchestrated, not an in-container exec. It shells out to scripts/run_data_collection.sh, which does docker run -d against geniesim3-data-collection:latest and the entrypoint launches two processes (Isaac Sim server + task client). Don't treat it like .