Improve AhmedML dataset card

README.md

@@ -1,65 +1,84 @@
 ---
+pretty_name: AhmedML
 license: cc-by-sa-4.0
-viewer: false 
+viewer: false
+tags:
+- computational-fluid-dynamics
+- scientific-machine-learning
+- automotive-aerodynamics
+- high-fidelity-simulation
+- surrogate-modeling
+- ai-for-science
+- open-data
 ---
-AhmedML: High-Fidelity Computational Fluid Dynamics dataset for incompressible, low-speed bluff body aerodynamics
--------
 
-Contact: 
-----------
-Neil Ashton (NVIDIA) - contact@caemldatasets.org
+# AhmedML: High-Fidelity Computational Fluid Dynamics Dataset for Incompressible, Low-Speed Bluff Body Aerodynamics
 
-website:
-----------
-https://caemldatasets.org
+[Project website](https://caemldatasets.org) · [Paper](https://arxiv.org/abs/2407.20801) · [Hugging Face paper page](https://huggingface.co/papers/2407.20801) · [Official splits](splits/)
 
-Summary:
--------
+## At a Glance
 
-This dataset contains 500 different geometric variations of the Ahmed Car Body - a simplified car-like shape that exhibits many of the flow topologies that are present on bluff bodies such as road vehicles. The dataset contains a wide range of geometries that exhibit fundamental flow physics such as geometry and pressure-induced flow separation of flows as well as 3D vortical structures. Each variation of the Ahmed car body were run using a time-accurate hybrid Reynolds-Averaged Navier-Stokes (RANS) - Large-Eddy Simulation (LES) turbulence modelling approach using the open-source CFD code OpenFOAM. The dataset contains both surface boundary, 3D volume, geometry STL and forces/moments in open-source formats (.vtu,.vtp). 
+| | |
+|---|---|
+| Domain | Incompressible, low-speed automotive bluff-body aerodynamics |
+| Scale | 500 geometric variants of the Ahmed body |
+| CFD method | Time-accurate hybrid RANS–LES using OpenFOAM v2212 |
+| Simulation size | Approximately 20 million cells and 80 convective time units per case |
+| Published data | Surface and volume fields, geometry, slices, images, and force/moment coefficients |
+| License | CC BY-SA 4.0 |
 
-CFD Solver:
-----------
-All cases were run using the open-source finite-volume code OpenFOAM v2212. Each case was run transiently for approximately 80 convective time units (CTU) on meshes of approximately 20M cells.  Please see the paper for full details on the code and validation:
+## Summary
 
-How to cite this dataset:
-----------------
-In order to cite the use of this dataset please cite the paper below which contains full details on the dataset. It can be found here: https://arxiv.org/abs/2407.20801
+AhmedML contains 500 geometric variations of the Ahmed body, a simplified car-like shape that exhibits many of the flow topologies found on bluff bodies such as road vehicles. The variants cover fundamental flow physics including geometry- and pressure-induced separation and three-dimensional vortical structures.
 
+Each variant was simulated using a time-accurate hybrid Reynolds-Averaged Navier–Stokes (RANS) and Large-Eddy Simulation (LES) turbulence-modelling approach in the open-source CFD code OpenFOAM. The dataset provides surface boundary data, three-dimensional volume data, geometry STL files, slices, images, and forces/moments in open formats including VTP and VTU.
+
+## CFD Solver
+
+All cases were run with the open-source finite-volume code OpenFOAM v2212. Each transient simulation covered approximately 80 convective time units (CTU) on a mesh of approximately 20 million cells. See the [paper](https://arxiv.org/abs/2407.20801) for full solver and validation details.
+
+## How to Cite This Dataset
+
+Please cite the corresponding paper:
+
+```bibtex
 @article{ashton2024ahmed,
-    title = {{AhmedML: High-Fidelity Computational Fluid Dynamics dataset for incompressible, low-speed bluff body aerodynamics}},
-    year = {2024},
-    journal = {arxiv.org},
-    author = {Ashton, Neil and Maddix, Danielle and Gundry, Samuel and Shabestari, Parisa}
+  title   = {{AhmedML: High-Fidelity Computational Fluid Dynamics Dataset for Incompressible, Low-Speed Bluff Body Aerodynamics}},
+  author  = {Ashton, Neil and Maddix, Danielle and Gundry, Samuel and Shabestari, Parisa},
+  journal = {arXiv preprint arXiv:2407.20801},
+  year    = {2024},
+  url     = {https://arxiv.org/abs/2407.20801}
 }
+```
+
+## Dataset Structure and Files
+
+Each `run_i` folder corresponds to a different geometry and contains the following files, where `i` is the run number:
+
+* `ahmed_i.stl`: Geometry STL (approximately 5 MB).
+* `geo_parameters_i.csv`: Parameters defining the geometry (`run_500` is missing this per-run file).
+* `boundary_i.vtp`: Surface boundary data (approximately 500 MB).
+* `boundary_cell_area_i.npy`: Native polygon areas in `m^2`, in the exact `CellData` tuple order of `boundary_i.vtp`.
+* `volume_i.vtu`: Volume field data (approximately 5 GB).
+* `force_mom_i.csv`: Time-averaged force coefficients (`Cd`, `Cl`) using a constant reference area.
+* `force_mom_varref_i.csv`: Time-averaged force coefficients (`Cd`, `Cl`) using a geometry-dependent reference area.
+* `slices/`: VTP slices in the x, y, and z directions containing flow-field variables.
+* `images/`: Images of variables including `CpT` and `UxMean` on x-, y-, and z-oriented slices.
 
-Files:
--------
-Each folder (e.g run_1,run_2...run_"i" etc) corresponds to a different geometry that contains the following files where "i" is the run number: 
-* ahmed_i.stl : geometry stl (~5mb):
-* geo_parameters_1.csv (missing run 500): parameters that define the geometry
-* boundary_i.vtp : Boundary VTP (~500mb)
-* `boundary_cell_area_i.npy`: Native polygon areas (`m^2`) for `boundary_i.vtp`, in exact `CellData` tuple order.
-* volume_i.vtu : Volume field VTU (~5GB)
-* force_mom_i.csv : forces (Cd,Cl) time-averaged with constant reference area
-* force_mom_varref_i.csv : forces (Cd,Cl) time-averaged with varying reference area
-* slices : folder containing .vtp slices in x,y,z that contain flow-field variables
-* images : (folder) that contains images of the following variables (CpT, UxMean) for slices of the domain in the X,Y & Z locations.
-In addition we provide:
-* force_mom_all.csv : run, cd,cl for all runs in a single file
-* force_mom_varref_all.csv : run, cd,cl for all runs in a single file with varying reference area
-* geo_parameters_all.csv : all the geometry parameters for each run inside a single file
-* ahmedml.slvs : SolveSpace input file to create the parametric geometries
-* stl : folder containing stl files that were used as inputs to the OpenFOAM process
-* openfoam-casesetup.tgz : complete OpenFOAM setup that can be used to extend or reproduce the dataset
-* validation : folder containing full outputs from all four mesh levels that were used to validate the methodology
-* [`splits/`](splits/): deterministic benchmark manifests, methods documentation, derived metrics, diagnostic figures, and generation code
-
-## Recommended dataset splits
-
-AhmedML provides eight deterministic train/validation/test split families in
-[`splits/manifest.json`](splits/manifest.json). Identifiers correspond to the
-top-level `run_N` directories.
+The repository root also contains:
+
+* `force_mom_all.csv`: `run`, `cd`, and `cl` for all runs using a constant reference area.
+* `force_mom_varref_all.csv`: `run`, `cd`, and `cl` for all runs using a geometry-dependent reference area.
+* `geo_parameters_all.csv`: Geometry parameters for all runs in one file.
+* `ahmedml.slvs`: SolveSpace input used to create the parametric geometries.
+* `stl/`: STL files used as inputs to the OpenFOAM process.
+* `openfoam-casesetup.tgz`: Complete OpenFOAM setup for extending or reproducing the dataset.
+* `validation/`: Full outputs from all four mesh levels used to validate the methodology.
+* [`splits/`](splits/): Deterministic benchmark manifests, methods documentation, derived metrics, diagnostic figures, and generation code.
+
+## Recommended Dataset Splits
+
+AhmedML provides eight deterministic train/validation/test split families in [`splits/manifest.json`](splits/manifest.json). Identifiers correspond to the top-level `run_N` directories.
 
 | Split | Type | Train | Validation | Test | Intended evaluation |
 |---|---:|---:|---:|---:|---|
@@ -72,95 +91,95 @@ top-level `run_N` directories.
 | `low_drag` | OOD | 350 | 50 | 100 | Low-drag extrapolation |
 | `image_wake` | OOD | 350 | 50 | 100 | Image-derived wake extrapolation |
 
-The `full` assignment follows the established seed-42 AhmedML baseline used by
-Noether. The reduced-data training sets are strictly nested and share the same
-validation and test cases. For the OOD families, validation is selected from
-the training-side population, while the held-out extreme is reserved for final
-testing.
+The `full` assignment follows the established seed-42 AhmedML baseline used by Noether. The reduced-data training sets are strictly nested and share the same validation and test cases. For the OOD families, validation is selected from the training-side population, while the held-out extreme is reserved for final testing.
 
-Download only the split package with:
+Complete definitions, construction methods, missing-data treatment, diagnostic figures, and reproducibility instructions are provided in [`splits/README.md`](splits/README.md).
 
-```bash
-hf download neashton/ahmedml \
-  --type dataset \
-  --include "splits/**" \
-  --local-dir ./ahmedml
-```
+## Native Boundary Polygon Areas
 
-Complete definitions, construction methods, missing-data treatment,
-diagnostic figures, and reproducibility instructions are provided in
-[`splits/README.md`](splits/README.md).
+Each `run_N/boundary_cell_area_N.npy` is a one-dimensional little-endian float32 array containing one area in `m^2` per native polygon—and therefore per `CellData` tuple—of `run_N/boundary_N.vtp`, in its original cell order. The array is valid only for that exact raw VTP. The definition, generator, hashes, and complete 500-run manifest are in [`surface_cell_areas/`](surface_cell_areas/).
 
-## Native boundary polygon areas
+## How to Download
 
-Each `run_N/boundary_cell_area_N.npy` is a one-dimensional little-endian float32 array containing one area in `m^2` per native polygon—and therefore per `CellData` tuple—of `run_N/boundary_N.vtp`, in its original cell order. The array is valid only for that exact raw VTP. The definition, generator, hashes, and complete 500-run manifest are in [`surface_cell_areas/`](surface_cell_areas/).
+AhmedML is publicly available on Hugging Face and is approximately 2 TB in full. We recommend the Hugging Face CLI for selective, resumable downloads and a dry-run preview before transferring large files.
 
-Downloads
---------------
+### Prerequisites
 
-The dataset is now available on HuggingFace. Below are some examples of how to download all or selected parts of the dataset. Please refer to the HuggingFace documentation for other ways to accessing the dataset and building workflows.
+```bash
+pip install -U huggingface_hub hf_xet
+```
 
-Example 1: Download all files (~2TB)
---------
-Please note you'll need to have git lfs installed first, then you can run the following command:
+Authentication is optional for this public dataset, but recommended for large downloads to avoid unauthenticated rate limits:
 
+```bash
+hf auth login
 ```
-git clone git@hf.co:datasets/neashton/ahmedml
+
+### Preview or Download the Full Dataset
+
+Preview the full download:
+
+```bash
+hf download neashton/ahmedml \
+  --repo-type dataset \
+  --local-dir ./ahmedml_data \
+  --dry-run
 ```
 
-Example 2: only download select files (STL,images & force and moments):
----------
-Create the following bash script that could be adapted to loop through only select runs or to change to download different files e.g boundary/volume.
+Remove `--dry-run` to start the download.
+
+### Preview Selected Metadata and Benchmark Files
+
+```bash
+hf download neashton/ahmedml \
+  --repo-type dataset \
+  --local-dir ./ahmedml_data \
+  --include "force_mom_all.csv" \
+  --include "force_mom_varref_all.csv" \
+  --include "geo_parameters_all.csv" \
+  --include "splits/**" \
+  --dry-run
 ```
-#!/bin/bash
 
-# Set the paths
-HF_OWNER="neashton"
-HF_PREFIX="ahmedml"
+Remove `--dry-run` to download those files.
 
-# Set the local directory to download the files
-LOCAL_DIR="./ahmed_data"
+### Preview a Small Per-Run Subset
 
-# Create the local directory if it doesn't exist
-mkdir -p "$LOCAL_DIR"
+This example selects only the geometry and force file from one run:
 
-# Loop through the run folders from 1 to 500
-for i in $(seq 1 500); do
-    RUN_DIR="run_$i"
-    RUN_LOCAL_DIR="$LOCAL_DIR/$RUN_DIR"
+```bash
+hf download neashton/ahmedml \
+  --repo-type dataset \
+  --local-dir ./ahmedml_data \
+  --include "run_1/ahmed_1.stl" \
+  --include "run_1/force_mom_1.csv" \
+  --dry-run
+```
 
-    # Create the run directory if it doesn't exist
-    mkdir -p "$RUN_LOCAL_DIR"
+Remove `--dry-run` to download the selected files. Keep glob patterns in quotes so the shell does not expand them locally.
 
-    # Download the ahmed_i.stl file
-    wget "https://huggingface.co/datasets/${HF_OWNER}/${HF_PREFIX}/resolve/main/$RUN_DIR/ahmed_$i.stl" -O "$RUN_LOCAL_DIR/ahmed_$i.stl"
+## Acknowledgements
 
-    # Download the force_mom_i.csv file
-    wget "https://huggingface.co/datasets/${HF_OWNER}/${HF_PREFIX}/resolve/main/$RUN_DIR/force_mom_$i.csv" -O "$RUN_LOCAL_DIR/force_mom_$i.csv"
+* OpenFOAM solver and workflow development by Neil Ashton (Amazon Web Services, now NVIDIA).
+* Geometry parameterization by Samuel Gundry and Parisa Shabestari (Amazon Web Services).
+* Guidance on dataset preparation for ML by Danielle Maddix (Amazon Web Services).
+* Simulation runs, HPC setup, and dataset preparation by Neil Ashton (Amazon Web Services, now NVIDIA).
 
-done
-```
+## License
 
+This dataset is provided under the CC BY-SA 4.0 license. See `LICENSE.txt` for the full license text.
 
-Acknowledgements
------------
-* OpenFOAM solver and workflow development by Neil Ashton (Amazon Web Services, now NVIDIA)
-* Geometry parameterization by Samuel Gundry (Amazon Web Services) and Parisa Shabestari (Amazon Web Services)
-* Guidance on dataset preparation for ML by Danielle Madix (Amazon Web Services)
-* Simulation runs, HPC setup and dataset preparation by Neil Ashton (Amazon Web Services, now NVIDIA)
+## Contact
 
-License
-----
-This dataset is provided under the CC BY SA 4.0 license, please see LICENSE.txt for full license text.
+Dataset questions and corrections: Neil Ashton at [contact@caemldatasets.org](mailto:contact@caemldatasets.org).
 
-version history:
----------------
+## Version History
 
 * 17/08/2026 - Added native boundary polygon-area arrays for all 500 runs, with reproducibility metadata.
 * 17/08/2026 - Added deterministic benchmark train/validation/test splits, including nested data-efficiency and out-of-distribution evaluation protocols.
-* 15/02/2025 - files uploaded to HuggingFace
-* 12/11/2024 - added validation folder that contains the full output from all four mesh levels that were used to validate the methodology used.
-* 04/08/2024 - updates to the file description and arxiv paper
-* 05/06/2024 - global forces/geo added for all runs
-* 01/05/2024 - force/moments corrected (prior version had incorrect Cs data)
-* 18/04/2024 - draft version produced 
+* 15/02/2025 - Files uploaded to Hugging Face.
+* 12/11/2024 - Added the `validation/` folder containing full output from all four mesh levels used to validate the methodology.
+* 04/08/2024 - Updated the file description and arXiv paper.
+* 05/06/2024 - Added global force and geometry data for all runs.
+* 01/05/2024 - Corrected forces/moments; the prior version contained incorrect `Cs` data.
+* 18/04/2024 - Draft version produced.