`airsspy.volume_minsep_data`#

Dataset helpers for volume and minsep predictors.

Module Contents#

`VolumeAggregate`	Aggregated volume target for one reduced formula.
`PairAggregate`	Aggregated minsep target for one formula and canonical pair.

`canonical_pair`	Return a deterministic canonical pair tuple.
`canonical_pair_key`	Return a canonical pair key formatted as `A-B`.
`composition_dict_from_formula`	Parse a reduced formula into a composition dictionary.
`formula_num_atoms`	Return the number of atoms in one reduced formula unit.
`formula_species`	Return species present in a reduced formula.
`enumerate_formula_pairs`	Enumerate all canonical species pairs implied by a formula.
`load_dataset_records`	Load volume/minsep training records from disk.
`volume_per_atom_from_record`	Normalize a structure’s cell volume by the number of atoms in the cell.
`canonical_minsep_map`	Collapse mirrored minsep entries like `A-B` and `B-A` into one key.
`aggregate_training_targets`	Aggregate structure-level records into formula-level model targets.
`volume_rows_to_dicts`	Serialize volume aggregate rows to dictionaries.
`pair_rows_to_dicts`	Serialize pair aggregate rows to dictionaries.
`get_minsep_dict`	Return the minimum observed distance for each ordered species pair.
`build_dataset_rows`	Build raw minsep/volume rows from Materials Project-like documents.
`build_minsep_volume_dataset`	Build a raw minsep/volume dataset from a Materials Project document dump.
`curate_minsep_volume_dataset`	Select the lowest-energy-above-hull row for each reduced formula.

airsspy.volume_minsep_data.canonical_pair(symbol_a: str, symbol_b: str) → tuple[str, str][source]#: Return a deterministic canonical pair tuple.

airsspy.volume_minsep_data.canonical_pair_key(symbol_a: str, symbol_b: str) → str[source]#: Return a canonical pair key formatted as A-B.

airsspy.volume_minsep_data.composition_dict_from_formula(formula: str) → dict[str, float][source]#: Parse a reduced formula into a composition dictionary.

airsspy.volume_minsep_data.formula_num_atoms(formula: str) → float[source]#: Return the number of atoms in one reduced formula unit.

airsspy.volume_minsep_data.formula_species(formula: str) → tuple[str, ...][source]#: Return species present in a reduced formula.

airsspy.volume_minsep_data.enumerate_formula_pairs(formula: str) → list[tuple[str, str]][source]#: Enumerate all canonical species pairs implied by a formula.

airsspy.volume_minsep_data.load_dataset_records(path: str | pathlib.Path) → list[dict][source]#: Load volume/minsep training records from disk.

airsspy.volume_minsep_data.volume_per_atom_from_record(record: dict) → float[source]#: Normalize a structure’s cell volume by the number of atoms in the cell.

airsspy.volume_minsep_data.canonical_minsep_map(record: dict) → dict[str, float][source]#: Collapse mirrored minsep entries like A-B and B-A into one key.

class airsspy.volume_minsep_data.VolumeAggregate[source]#

Aggregated volume target for one reduced formula.

class airsspy.volume_minsep_data.PairAggregate[source]#

Aggregated minsep target for one formula and canonical pair.

airsspy.volume_minsep_data.aggregate_training_targets(records: collections.abc.Iterable[dict]) → tuple[list[airsspy.volume_minsep_data.VolumeAggregate], list[airsspy.volume_minsep_data.PairAggregate], dict][source]#: Aggregate structure-level records into formula-level model targets.

airsspy.volume_minsep_data.volume_rows_to_dicts(rows: collections.abc.Iterable[airsspy.volume_minsep_data.VolumeAggregate]) → list[dict][source]#: Serialize volume aggregate rows to dictionaries.

airsspy.volume_minsep_data.pair_rows_to_dicts(rows: collections.abc.Iterable[airsspy.volume_minsep_data.PairAggregate]) → list[dict][source]#: Serialize pair aggregate rows to dictionaries.

airsspy.volume_minsep_data.get_minsep_dict(structure: pymatgen.core.Structure, cutoff: float = 5.0) → dict[str, float][source]#: Return the minimum observed distance for each ordered species pair.

airsspy.volume_minsep_data.build_dataset_rows(docs: collections.abc.Iterable[dict[str, Any]], *, show_progress: bool = True) → list[dict[str, Any]][source]#: Build raw minsep/volume rows from Materials Project-like documents.

airsspy.volume_minsep_data.build_minsep_volume_dataset(*, mp_docs_path: str | pathlib.Path, output_path: str | pathlib.Path, show_progress: bool = True) → dict[str, Any][source]#: Build a raw minsep/volume dataset from a Materials Project document dump.

airsspy.volume_minsep_data.curate_minsep_volume_dataset(*, minsep_dataset_path: str | pathlib.Path, mp_docs_path: str | pathlib.Path, output_path: str | pathlib.Path) → dict[source]#: Select the lowest-energy-above-hull row for each reduced formula.