Color Refinement#

stereomolgraph.algorithms.circular.numpy_int_multiset_hash(arr: np.ndarray[tuple[int, ...], np.dtype[np.int64]], out: None | np.ndarray[tuple[Literal[1], ...], np.dtype[np.int64]] = None) → np.ndarray#

Hash function for a multiset (order-independent with duplicates) of integers. Works by sorting the elements and then applying the tuple hashing function.

Return type:

ndarray

stereomolgraph.algorithms.circular.label_hash(mg: MolGraph, atom_labels: Collection[str] = ('atom_type',)) → ndarray[tuple[int], dtype[int64]]#

Generates a hash for each atom based on choosen attributes.

Parameters:

  • mg (MolGraph) – MolGraph object containing the atoms.

  • atom_labels (Collection[str]) – Iterable of attribute names to use for hashing. (default: ('atom_type',))

Return type:

ndarray[tuple[int], dtype[int64]]

stereomolgraph.algorithms.circular.circular_generator(mg: MolGraph, atom_labels: None | ndarray[tuple[int], dtype[int64]] = None) → Iterator[ndarray[tuple[int], dtype[int64]]]#

Color refinement algorithm for MolGraph.

This algorithm refines the atom coloring based on their connectivity. Identical to the Weisfeiler-Lehman (1-WL) algorithm.

Parameters:

  • mg (MolGraph) – MolGraph object containing the atoms and their connectivity.

  • max_iter – Maximum number of iterations for refinement. Default is None, which means it will run until convergence.

Return type:

Iterator[ndarray[tuple[int], dtype[int64]]]

stereomolgraph.algorithms.circular.color_refine_hash_mg(graph: MolGraph) → int#

Color-refined hash for plain MolGraph objects.

Return type:

int

stereomolgraph.algorithms.circular.color_refine_hash_smg(graph: StereoMolGraph) → int#

Color-refined hash for StereoMolGraph objects.

Drops the extra sentinel slot the stereo generator appends.

Return type:

int

stereomolgraph.algorithms.circular.color_refine_hash_crg(graph: CondensedReactionGraph) → int#

Color-refined hash for CondensedReactionGraph objects.

Return type:

int

stereomolgraph.algorithms.circular.color_refine_hash_scrg(graph: StereoCondensedReactionGraph) → int#

Color-refined hash for StereoCondensedReactionGraph objects.

Return type:

int

stereomolgraph.algorithms.circular.circular_fingerprint(graph: MolGraph, radius: int = 3, n_bits: int = 2048, count: bool = True, accumulate: bool = False, include_hydrogens: bool = False) → ndarray#

Build a circular fingerprint for a molecular graph.

If accumulate is True, identifiers from every radius up to the requested radius are included. Otherwise, only the last radius is used. If include_hydrogens is False, atom environments centered on hydrogens are excluded.

Parameters:

  • graph (MolGraph) – Molecular graph to fingerprint.

  • radius (int) – Maximum refinement radius to include. (default: 3)

  • n_bits (int) – Length of the folded fingerprint. If 0 or None, return the unique integer identifiers instead of a folded bit/count vector. (default: 2048)

  • count (bool) – If True, all unique values are used once. (default: True)

  • accumulate (bool) – Whether to include identifiers from all radii up to radius instead of only the final radius. (default: False)

  • include_hydrogens (bool) – Whether to include hydrogen-centered environments. (default: False)

Return type:

ndarray

stereomolgraph.algorithms.circular.circular_stereo_fingerprint(graph: StereoMolGraph, radius: int = 3, n_bits: int = 2048, count: bool = True, accumulate: bool = False, include_hydrogens: bool = False) → ndarray#

Build a circular stereo fingerprint for a molecular graph.

If accumulate is True, identifiers from every radius up to the requested radius are included. Otherwise, only the last radius is used. If include_hydrogens is False, atom environments centered on hydrogens are excluded.

Parameters:

  • graph (StereoMolGraph) – Molecular graph to fingerprint.

  • radius (int) – Maximum refinement radius to include. (default: 3)

  • n_bits (int) – Length of the folded fingerprint. If 0 or None, return the unique integer identifiers instead of a folded bit/count vector. (default: 2048)

  • count (bool) – If True, all unique values are used once. (default: True)

  • accumulate (bool) – Whether to include identifiers from all radii up to radius instead of only the final radius. (default: False)

  • include_hydrogens (bool) – Whether to include hydrogen-centered environments. (default: False)

Return type:

ndarray