Bundled datasets

GraphChem ships five curated fuel-property CSVs under graphchem/datasets/static/. Load them with the graphchem.datasets helpers. These tables are for examples and reproducible experiments in this package; they are not a live external database.

Common fields

  • Format: CSV with a SMILES column and one property column.

  • Loader return type: tuple[list[str], torch.Tensor] where the tensor has shape (n, 1) and dtype float32.

  • Package license: MIT (code). Dataset provenance is separate — see each card.

  • Limitations: Bundled snapshot only; labels are not revalidated here; invalid or unusual SMILES may still be present as historically shipped.

Provenance for all five sets is Unknown pending maintainer confirmation (design Q5). Do not treat the numeric labels as citable literature values without an external reference.

Cetane number (CN)

  • Identifier: cn

  • Loader: graphchem.datasets.load_cn()

  • Package data: graphchem/datasets/static/cn.csv

  • Columns: SMILES, CN

  • Rows: 460

  • Property: Cetane number

  • Units: Unknown (values as stored in the CSV)

  • Provenance / primary literature: Unknown — omitted for now

  • Access / license notes: Distributed with the GraphChem package; confirm redistribution terms with maintainers before republishing the table alone

Lower heating value (LHV)

  • Identifier: lhv

  • Loader: graphchem.datasets.load_lhv()

  • Package data: graphchem/datasets/static/lhv.csv

  • Columns: SMILES, LHV

  • Rows: 388

  • Property: Lower heating value

  • Units: Unknown (values as stored in the CSV)

  • Provenance / primary literature: Unknown — omitted for now

  • Access / license notes: Distributed with the GraphChem package; confirm redistribution terms with maintainers before republishing the table alone

Motor octane number (MON)

  • Identifier: mon

  • Loader: graphchem.datasets.load_mon()

  • Package data: graphchem/datasets/static/mon.csv

  • Columns: SMILES, MON

  • Rows: 308

  • Property: Motor octane number

  • Units: Unknown (octane number scale as stored)

  • Provenance / primary literature: Unknown — omitted for now

  • Access / license notes: Distributed with the GraphChem package; confirm redistribution terms with maintainers before republishing the table alone

Research octane number (RON)

  • Identifier: ron

  • Loader: graphchem.datasets.load_ron()

  • Package data: graphchem/datasets/static/ron.csv

  • Columns: SMILES, RON

  • Rows: 308

  • Property: Research octane number

  • Units: Unknown (octane number scale as stored)

  • Provenance / primary literature: Unknown — omitted for now

  • Access / license notes: Distributed with the GraphChem package; confirm redistribution terms with maintainers before republishing the table alone

Yield sooting index (YSI)

  • Identifier: ysi

  • Loader: graphchem.datasets.load_ysi()

  • Package data: graphchem/datasets/static/ysi.csv

  • Columns: SMILES, YSI

  • Rows: 558

  • Property: Yield sooting index

  • Units: Unknown (YSI scale as stored; includes negative values)

  • Provenance / primary literature: Unknown — omitted for now

  • Access / license notes: Distributed with the GraphChem package; confirm redistribution terms with maintainers before republishing the table alone