Returns the mapping from column names to column values.

Return Vec3Array corresponding to zeros of given shape.

Performs a shallow copy but with specified fields updated.

Pads an array to a given shape. Wrapper around np.pad(). Args: arr: numpy array to pad shape: target shape, use None for axes that should s

Returns the path to a given test data. Args: data_name: the name of the test data file relative to data_dir. If not set, this will

Returns a sliced column array using a key array or a boolean mask.

Gets a column by name.

Returns sanitised version of the name that can be used as a filename.

Compute dot product between 'self' and 'other'.

Compute Norm of Vec3Array, clipped to epsilon.

Compute the msa features.

Validates that the JSON doesn't contain any extra unwanted keys.

Returns a copy of this table with the specified changes applied. Args: **new_column_by_column_name: New values for the specified columns.

Filters the structure by field values and returns a new structure. Predicates are specified as keyword arguments, with names following the pa

Creates GatherInfo from a given dictionary.

Returns the number of residues in this Structure. Args: count_unresolved: Whether to include unresolved (empty) residues. Returns:

Returns a sliced table using an index (!= key) array.

Copies and pads the layout to the requested shape. Args: shape: new shape for the atom layout Returns: a copy of the atom layout

Returns a copy of the table with the specified columns remapped. Args: **mapping_by_col: Each kwarg key should be the name of one of this t

Loads the input from the AlphaFold JSON string.

A mapping from internal chain ID to a sequence of residue names. The residue names are the full residue names rather than single letter codes

Returns a mapping from chain ID to a single letter residue sequence. Args: include_missing_residues: Whether to include residues that have

Compute cross product between 'self' and 'other'.

Returns the MSA in the A3M format.

Unstacks a multi-model structure into a list of Structures. This method is the inverse of `stack`. Example usage: ``` strucs = multi

Boolean label indicating if each atom is from entities or not.

Filters the structure to only include the selected entity types. This convenience method abstracts away the specifics of mmCIF entity type na

Gets the row with specified key value.

Returns an Mmcif representing the structure.

Performs a cascade delete operation on the structure's tables. Cascade delete ensures all the tables are consistent after any table fields ar

Reads a maybe compressed (gzip, xz, zstd) file from the given path. Args: path: The path to the file to read. This can be either absolute path,

Unwrap an object from a zero-dim np.ndarray.

Applies Rot3Array to point.

Filters the table using mask and/or predicates and returns a new table. Predicates can be either: 1. A constant value, e.g. `'CA'`. In this

Returns a new structure with the specified elements removed.

Constructs Ligand from the AlphaFoldServer JSON dict.

Creates an empty Msa containing just the query sequence.

Iterates over the chains in the structure.

Yields rows from the table. This can be used to easily convert a table to a Pandas dataframe: ```py df = pd.DataFrame(table.iterrows())

Mapping from author chain ID to internal chain ID, polymers only. This mapping is well defined only for polymers (protein, DNA, RNA), but not

Iterator for (start, end) pairs from an array of start indices.

Adaptive LayerNorm.

Constructs Ligand from the AlphaFold JSON dict.

Returns the indices of the from/dest atoms in the atom_key array.

Iterates over the atoms in the structure.

Returns a boolean array of rows to keep, or None if all can be kept. Args: mask: See `Table.filter`. apply_per_element: See `Table.fi

Return unit vector with optional clipping.

Converts to a sequence of CCD codes.

Transition Block.

Returns a view of this table with mapped column names.

Applies index_arr to the atom table using NumPy-style array indexing. Args: index_arr: A 1D NumPy array that will be used to index into the

Iterator for (start, end) pairs from an array of start indices.

Adaptive zero init, from AdaLN-zero.

Adds a residue (and its chain) to the tables.

Apply Rigid3Array transform to point.

Returns a shallow copy with the global columns updated.

Returns a new Templates object with only the hits that pass all filters. This also filters on query_release_date and max_template_date. Args

Parses the single A3M and builds the Msa object.

Instantiates an AtomConfidence from a structure. Args: inference_result: Inference result from AlphaFold. Returns: Scores in Ato

Get bond_layout for all bonds between two sets of chain types. There is a mask (all_mask) that runs through this script, and each bond pair needs

Returns the transforms required to generate the given assembly.

Returns inverse of Rot3Array.

Returns a new structure with renamed chain IDs (label_asym_ids). The chains' auth_asym_ids will be updated to be identical to the chain ID si

Converts Ligand to an AlphaFold JSON dict.

Returns an mmCIF string representing the structure. Args: coords_decimal_places: The number of decimal places to keep for atom coor

Nan-mean with weighting -- empty slices return NaN.

Checks that given attributes of the object have the expected shape.

Builds a HMM for the aligned sequences given as an MSA string. Args: msa: A string with the aligned sequences, in A3M or STO format.

Computes the pTM metrics from PAE.