luhr/biofaudry-digital-twin

BioFaudry laboratory digital twin assets

Private asset delivery for MagicSim. Includes all 796 asset types, 941 authored joints, and the complete laboratory: 1,110 placed objects plus the room shell (759 asset types in the scene). Source authoring examples and the skeleton/navigation scene variants are also included; the 796-asset acceptance count refers to the production models.

Content	Location
Native USD, textures, mass/collision/joint annotations, device profiles	Assets/DigitalTwin/BioFaudry/<group>/<asset_id>/
Complete USD laboratory	Assets/DigitalTwin/BioFaudry/Scene/lab_scene.usda
796 native Blender models	outputs/digital_twin/blender/assets/
Complete Blender laboratory, packed model libraries and textures	outputs/digital_twin/blender/biofaudry_lab.blend
796 browser-compatible GLB models	outputs/digital_twin/glb/models/
1,592 asset preview images	outputs/digital_twin/gallery/img/
Asset acceptance receipts	acceptance/<asset_id>/
Whole-scene, device and matter acceptance	outputs/digital_twin/ JSON reports and renders
Single-download asset bundle	BioFaudry_Assets.zip
SHA-256 for each repository payload file	manifest.json

USD is authoritative for physical mass, collisions and joints. Open the native scene in Isaac Sim 5.1, or the packed .blend in Blender 4.2 or later. GLB files provide visual geometry; use USD and annotations for simulation.

Download

Log in to the Hugging Face account that has access to this private dataset:

python -m pip install huggingface_hub
hf auth login

from huggingface_hub import snapshot_download
snapshot_download(
    repo_id="luhr/biofaudry-digital-twin",
    repo_type="dataset",
    local_dir="biofaudry-assets",
    ignore_patterns=["BioFaudry_Assets.zip", "BioFaudry_Assets.zip.sha256"],
)

To run MagicSim, check out its main branch, then copy the downloaded Assets/ and outputs/ directories into that checkout, preserving the directory layout. The runtime configuration is src/magicsim/Env/Conf/biofaudry_lab_env.yaml. Source code is distributed through GitHub. This HF repository contains the asset delivery and its evidence; the website project is stored separately.

Validation and scope

The original native delivery passed content-bound acceptance for all 796 assets. Its whole-scene simulation, device behavior, matter tests and rendered evidence are included in outputs/digital_twin/delivery_report.json and supporting reports. Repository payloads preserve the validated USD/Blender bytes. Acceptance receipt paths are relative so that they work after downloading to another machine.

Dimensions and placements include photo/video estimates with recorded confidence. Cloth, thin film and packaging use rigid approximations. Joint sweeps use a holding fixture and a separate unrestrained settle check. These files do not claim measured calibration of every real laboratory instrument.