id stringclasses 3 values | formula stringclasses 3 values | system stringclasses 1 value | bandgap_eV float64 3.2 5.6 | structure_path stringclasses 3 values |
|---|---|---|---|---|
ABO3_0001 | SrTiO3 | perovskite | 3.2 | data/structures/ABO3_0001.xyz |
ABO3_0002 | BaZrO3 | perovskite | 5 | data/structures/ABO3_0002.xyz |
ABO3_0003 | LaAlO3 | perovskite | 5.6 | data/structures/ABO3_0003.xyz |
AI4Materials Demo FAIR Perovskites
This is a teaching dataset demonstrating F.A.I.R. hosting on the Hugging Face Hub.
It contains a small table of oxide perovskites with band gaps and toy EXTXYZ structures.
Contents
data/table.csv— main tabular datadata/records.jsonl— line-delimited JSON mirrordata/structures/*.xyz— example structures (EXTXYZ)metadata/schema.json— JSON Schema for validationCITATION.cff,LICENSE— citation & reuse terms
Provenance
Synthetic examples generated for classroom demonstration on {datetime.utcnow().date()}.
How to load
from datasets import load_dataset
ds = load_dataset("cparidaAI/fair_dataset_demo", data_files={"train": "data/table.csv"})
print(ds)
Or download a structure file:
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="cparidaAI/fair_dataset_demo", filename="data/structures/ABO3_0001.xyz")
License
MIT. Please cite using CITATION.cff.
F.A.I.R. checklist
- Findable: metadata, tags, README, (add DOI later via Zenodo)
- Accessible: public repo, open formats
- Interoperable: CSV/JSON/EXTXYZ, schema describes fields/units
- Reusable: license, clear citation, validation, examples
- Downloads last month
- 46