Adapters

Adapters download or generate raw biomolecular data and produce the initial Batch stream, Split, and Assets objects consumed by datasets. Summaries below are taken from each class docstring.

Download protein structures from the AlphaFold Protein Structure Database.

class bioverse.adapters.alpha_fold.AlphaFoldAdapter[source]

Bases: Adapter

Download protein structures from the AlphaFold Protein Structure Database.

classmethod download(name='swissprot_pdb', version='v4')[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download AlphaFold structures with Foldseek-based homology exclusions for inverse folding benchmarks.

class bioverse.adapters.alpha_fold_exclusion.AlphaFoldExclusionAdapter[source]

Bases: Adapter

Download AlphaFold structures with Foldseek-based homology exclusions for inverse folding benchmarks.

classmethod download(name: str = 'swissprot_pdb', version: str = 'v4')[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download AlphaFold structures curated for the inverse folding benchmark suite.

class bioverse.adapters.alpha_fold_inv_bench.AlphaFoldInvBenchAdapter[source]

Bases: Adapter

Download AlphaFold structures curated for the inverse folding benchmark suite.

classmethod download(af_name: str = 'swissprot_pdb', af_version: str = 'v4', use_api_fallback: bool = True, api_workers: int | None = None)[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download protein-ligand complexes from the ARES docking benchmark.

class bioverse.adapters.ares.AresAdapter[source]

Bases: Adapter

Download protein-ligand complexes from the ARES docking benchmark.

classmethod download()[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download deep mutational scanning benchmarks from ProteinGym.

class bioverse.adapters.protein_gym.ProteinGymAdapter[source]

Bases: Adapter

Download deep mutational scanning benchmarks from ProteinGym.

classmethod download(assays: list[str] | None = None)[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download sequences and structures for the ProteinInvBench inverse folding suite.

class bioverse.adapters.protein_inv_bench.ProteinInvBenchAdapter[source]

Bases: Adapter

Download sequences and structures for the ProteinInvBench inverse folding suite.

classmethod download()[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download biomolecular structure datasets from ProteinShake.

class bioverse.adapters.proteinshake.ProteinShakeAdapter[source]

Bases: Adapter

Download biomolecular structure datasets from ProteinShake.

classmethod download(dataset: str)[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download small-molecule properties from the QM9 (Quantum Machines) dataset.

class bioverse.adapters.quantum_machines.QuantumMachinesAdapter[source]

Bases: Adapter

Download small-molecule properties from the QM9 (Quantum Machines) dataset.

classmethod download()[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download revised MD17 (rMD17) molecular dynamics trajectories.

class bioverse.adapters.revised_molecular_dynamics.RevisedMolecularDynamicsAdapter[source]

Bases: Adapter

Download revised MD17 (rMD17) molecular dynamics trajectories.

classmethod download(name)[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.

Download protein-ligand complexes from the Weng et al. docking benchmark.

class bioverse.adapters.weng_docking.WengDockingAdapter[source]

Bases: Adapter

Download protein-ligand complexes from the Weng et al. docking benchmark.

classmethod download(version: str = '5.5')[source]

Download raw data and return batches, split, and assets.

Returns:

  • batches – Iterator yielding Batch shards.

  • split – Train/validation/test partition assignments.

  • assets – Auxiliary lookup tables shared across batches.