Center and randomly augment the input coordinates. Parameters ---------- atom_coords : Tensor The atom coordinates. atom_mask

Collate the data. Parameters ---------- data : list[dict[str, Tensor]] The data to collate. Returns ------- dict[str

Generate 3D coordinates using EKTDG method. Taken from `pdbeccdutils.core.component.Component`. Parameters ---------- mol: Mol

Compute the metric from the logits. Parameters ---------- logits : torch.Tensor The logits of the metric end : float

Compute the metric from the logits. Parameters ---------- logits : torch.Tensor The logits of the metric end : float

Compute the ratio between d_alpha and alpha

We always output sde according to score parametrization;

Compute the chain-chain interfaces from a gemmi structure. Parameters ---------- atom_data : List[tuple] The atom data. chain

Adds bos/eos tokens for the language model, since the structure module doesn't use these.

Compute the mean of time-dependent density p_t

Compute the vector field corresponding to p_t

Sample xt from time-dependent density p_t; rng is required

Convert an atom name to a standard format. Parameters ---------- name : str The atom name. Returns ------- tuple[int

Convert an atom name to a standard format. Parameters ---------- name : str The atom name. Returns ------- Tuple[int

Convert an atom name to a standard format. Parameters ---------- name : str The atom name. Returns ------- Tuple[int

Create local attention bias based on query window n_queries and kv window n_keys. Args: n (int): the length of quiries

Create local attention bias based on query window n_queries and kv window n_keys. Args: n (int): the length of quiries

Crop the data to a maximum number of tokens. Parameters ---------- data : Tokenized The tokenized data. m

Evaluate the structure.

Execute the external docking-related command.

returns whether val is not none

Fetch the PDB files.

Run post-processing in main thread. Parameters ---------- out_dir : Path The output directory.

https://github.com/facebookresearch/DiT/blob/main/models.py#L303

https://github.com/facebookresearch/DiT/blob/main/models.py#L303

Get the deposited, released, and last revision dates. Parameters ---------- block : gemmi.cif.Block The block to process. Re

Get the method from a gemmi structure. Parameters ---------- block : gemmi.cif.Block The block to process. Returns -----

Get the resolution from a gemmi structure. Parameters ---------- block : gemmi.cif.Block The block to process. Returns -

Load stereo_chemical_props.txt into a nice structure. Load literature values for bond lengths and bond angles and translate bond angles into

Pad the data in all dimensions to the maximum found. Parameters ---------- data : List[Tensor] List of tensors to pad. value

Process a structure. Parameters ---------- data : PDB The raw input data. resource: Resource The shared resource.

Parse an MMCIF ligand. First tries to get the SMILES string from the RCSB. Then, tries to infer atom ordering using RDKit. Parameters

Parse (covalent) connection from a gemmi Connection. Parameters ---------- connections : gemmi.ConnectionList The connection list

Parse a fasta file. The name of the fasta file is used as the name of this job. We rely on the fasta record id to determine the entity type.

Parse a structure in MMCIF format. Parameters ---------- mmcif_file : PathLike Path to the MMCIF file. components: dict[str,

Process a gemmi Polymer into a chain object. Performs alignment of the full sequence to the polymer residues. Loads coordinates and masks for

Process a sequence into a chain object. Performs alignment of the full sequence to the polymer residues. Loads coordinates and masks for the

Parse TM-score output. Args: out: str in utf-8 encoding from stdout of TM-score.

Pick a random token from a chain. Parameters ---------- tokens : np.ndarray The token data. chain_id : int The chain

Pick protein tokens from the data. Parameters ---------- tokens : np.ndarray The token data. Returns ------- np.ndar

Compute plddt loss. Parameters ---------- pred_lddt: torch.Tensor The plddt logits pred_atom_coords: torc

Get the training dataloader. Returns ------- DataLoader The training dataloader.

Prints the contents of a DictConfig as a tree structure using the Rich library. :param cfg: A DictConfig composed by Hydra. :param print_orde

Run the data processing task.

Get the atom features. Parameters ---------- data : Tokenized The tokenized data. max_atoms : int, optional The maxim

Process a target. Parameters ---------- item : PDB The raw input data. resource: Resource The shared resource. ou

Get the symmetry features. Parameters ---------- data : Tokenized The tokenized data. Returns ------- dict[str, Tens

Get the token features. Parameters ---------- data : Tokenized The tokenized data. max_tokens : int The maximum numbe

Convert rotations given as quaternions to rotation matrices. Args: quaternions: quaternions with real part first, as ten

Generate random quaternions representing rotations, i.e. versors with nonnegative real part. Args: n: Number of quaternions in a