src.object package

Submodules

src.object.function module

class src.object.function.Coordinate(*args, **metadata)

Bases: CoordinateMixin, ABC

Parameters:
class src.object.function.CoordinateMixin(*args, **metadata)

Bases: FunctionMixin, ABC

Parameters:
class src.object.function.Degree(*args, **metadata)

Bases: DegreeMixin, DenseArray

Parameters:
class src.object.function.DegreeMixin(*args, **metadata)

Bases: FunctionMixin, ABC

Parameters:
class src.object.function.FunctionMixin(*args, **metadata)

Bases: ObjectMixin, BaseArray, ABC

Parameters:
classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype

alias of float64

fixed_ndim: int = 2
metadata: Metadata
property nfuncs: int
property npts: int
class src.object.function.KNNDistance(*args, **metadata)

Bases: KNNDistanceMixin, DenseArray

Parameters:

  • args (Any)

  • metadata (Any)

class src.object.function.KNNDistanceMixin(*args, **metadata)

Bases: FunctionMixin, ABC

Parameters:

  • args (Any)

  • metadata (Any)

class src.object.function.LocalDistortion(*args, **metadata)

Bases: LocalDistortionMixin, DenseArray

Parameters:

  • args (Any)

  • metadata (Any)

class src.object.function.LocalDistortionMixin(*args, **metadata)

Bases: FunctionMixin, ABC

Parameters:

  • args (Any)

  • metadata (Any)

class src.object.function.NeighborCount(*args, **metadata)

Bases: NeighborCountMixin, DenseArray

Parameters:

  • args (Any)

  • metadata (Any)

class src.object.function.NeighborCountMixin(*args, **metadata)

Bases: FunctionMixin, ABC

Parameters:

  • args (Any)

  • metadata (Any)

fixed_dtype

alias of int64

src.object.geometry_matrix_mixin module

class src.object.geometry_matrix_mixin.GeometryMatrixMixin(*args, **metadata)

Bases: ObjectMixin, BaseArray, ABC

Mixin class providing common functionality for geometry-related matrices.

Adds default dimensionality of 2 of and default dtype of np.float64, along with convenience properties for querying matrix shape and point counts. Typically used as a base for distance, affinity, and Laplacian matrices.

Parameters:
classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype

alias of float64

fixed_ndim: int = 2
property from_npts: int

Number of points corresponding to the rows of the matrix.

Returns:

Number of rows in the matrix.

Return type:

int

property is_square: bool

Check if the matrix is square (number of rows equals number of columns).

Returns:

True if the matrix is square, False otherwise.

Return type:

bool

property npts: int

Number of points in the matrix, assuming it is square.

Returns:

Number of points (rows or columns) if the matrix is square.

Return type:

int

Raises:

ValueError – If the matrix is not square.

property to_npts: int

Number of points corresponding to the columns of the matrix.

Returns:

Number of columns in the matrix.

Return type:

int

src.object.metadata module

class src.object.metadata.Metadata(name=None, dist_type=None, aff_type=None, lap_type=None, radius=None, eps=None, ks=None, radii=None, ds=None, degree_type=None, aff_minus_id=None)

Bases: object

Parameters:

  • name (str | None)

  • dist_type (Literal['euclidean', 'cityblock', 'sqeuclidean'] | None)

  • aff_type (Literal['gaussian'] | None)

  • lap_type (Literal['geometric', 'random_walk', 'symmetric'] | None)

  • radius (float | None)

  • eps (float | None)

  • ks (Tuple[int, ...] | None)

  • radii (Tuple[float, ...] | None)

  • ds (Tuple[int, ...] | None)

  • degree_type (Literal['adjacency', 'affinity'] | None)

  • aff_minus_id (bool | None)

aff_minus_id: bool | None
aff_type: Literal['gaussian'] | None
degree_type: Literal['adjacency', 'affinity'] | None
dist_type: Literal['euclidean', 'cityblock', 'sqeuclidean'] | None
ds: Tuple[int, ...] | None
eps: float | None
ks: Tuple[int, ...] | None
lap_type: Literal['geometric', 'random_walk', 'symmetric'] | None
name: str | None
radii: Tuple[float, ...] | None
radius: float | None
update_with(other)

Return a new Metadata with non-None values from ‘other’ overriding self.

Parameters:

other (Metadata)

Return type:

Metadata

src.object.metadata.optional_tuple_of_floats(instance, attribute, value)
Parameters:

  • instance (object)

  • attribute (Attribute)

  • value (Tuple[float, ...] | None)

Return type:

None

src.object.metadata.optional_tuple_of_ints(instance, attribute, value)
Parameters:

  • instance (object)

  • attribute (Attribute)

  • value (Tuple[int, ...] | None)

Return type:

None

src.object.object_mixin module

class src.object.object_mixin.ObjectMixin(*args, **kwargs)

Bases: ABC

Abstract base class providing metadata handling and type/dimension validation.

Parameters:

  • args (Any)

  • kwargs (Any)

abstractmethod classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype: Type[generic]
fixed_ndim: int
metadata: Metadata
src.object.object_mixin.chunk(arr, bsize=None)
Parameters:

  • arr (T)

  • bsize (int | None)

Return type:

Iterator[T]

Module contents

class src.object.Eigen(iterable, **kwargs)

Bases: ObjectMixin, Tuple[DenseArray, DenseArray]

This class represents a pair of eigenvalues and eigenvectors. It is a Tuple of two DenseArray with additional checks and methods. Ensures correct eigenvalue and eigenvector lengths and dimensions.

Parameters:

  • iterable (Iterable[Any])

  • kwargs (Any)

Return type:

Eigen

classmethod concat_with_metadata(arrs, axis=0)

Concatenates eigenvalues and eigenvectors along an axis and returns an instance of this class with metadata of the first object in arrs.

Parameters:

  • arrs (Sequence[Self]) – Sequence of objects of this class to concatenate.

  • axis (int) – Axis along which to concatenate arrays. (default: 0).

Returns:

A new instance of this class with concatenated data and metadata taken from the first element in arrs.

Return type:

Self

property eigenvalues: DenseArray
property eigenvectors: DenseArray
fixed_tuple_length: int = 2
fixed_type: Tuple[Type, Type] = (<class 'src.array.dense.dense.DenseArray'>, <class 'src.array.dense.dense.DenseArray'>)
fixed_value_ndim: int = 1
fixed_vector_ndim: int = 2
sort_by_eigval(largest=False)

Sorts the eigenvalues and eigenvectors by eigenvalue.

Parameters:

largest (bool) – If True, sorts with largest first. If False, sorts with smallest first.

Returns:

A sorted instance of this class.

Return type:

Self

class src.object.FunctionMixin(*args, **metadata)

Bases: ObjectMixin, BaseArray, ABC

Parameters:

  • args (Any)

  • metadata (Any)

classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype

alias of float64

fixed_ndim: int = 2
metadata: Metadata
property nfuncs: int
property npts: int
class src.object.GeometryMatrixMixin(*args, **metadata)

Bases: ObjectMixin, BaseArray, ABC

Mixin class providing common functionality for geometry-related matrices.

Adds default dimensionality of 2 of and default dtype of np.float64, along with convenience properties for querying matrix shape and point counts. Typically used as a base for distance, affinity, and Laplacian matrices.

Parameters:

  • args (Any)

  • metadata (Any)

classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype

alias of float64

fixed_ndim: int = 2
property from_npts: int

Number of points corresponding to the rows of the matrix.

Returns:

Number of rows in the matrix.

Return type:

int

property is_square: bool

Check if the matrix is square (number of rows equals number of columns).

Returns:

True if the matrix is square, False otherwise.

Return type:

bool

metadata: Metadata
property npts: int

Number of points in the matrix, assuming it is square.

Returns:

Number of points (rows or columns) if the matrix is square.

Return type:

int

Raises:

ValueError – If the matrix is not square.

property to_npts: int

Number of points corresponding to the columns of the matrix.

Returns:

Number of columns in the matrix.

Return type:

int

class src.object.Metadata(name=None, dist_type=None, aff_type=None, lap_type=None, radius=None, eps=None, ks=None, radii=None, ds=None, degree_type=None, aff_minus_id=None)

Bases: object

Parameters:

  • name (str | None)

  • dist_type (Literal['euclidean', 'cityblock', 'sqeuclidean'] | None)

  • aff_type (Literal['gaussian'] | None)

  • lap_type (Literal['geometric', 'random_walk', 'symmetric'] | None)

  • radius (float | None)

  • eps (float | None)

  • ks (Tuple[int, ...] | None)

  • radii (Tuple[float, ...] | None)

  • ds (Tuple[int, ...] | None)

  • degree_type (Literal['adjacency', 'affinity'] | None)

  • aff_minus_id (bool | None)

aff_minus_id: bool | None
aff_type: Literal['gaussian'] | None
degree_type: Literal['adjacency', 'affinity'] | None
dist_type: Literal['euclidean', 'cityblock', 'sqeuclidean'] | None
ds: Tuple[int, ...] | None
eps: float | None
ks: Tuple[int, ...] | None
lap_type: Literal['geometric', 'random_walk', 'symmetric'] | None
name: str | None
radii: Tuple[float, ...] | None
radius: float | None
update_with(other)

Return a new Metadata with non-None values from ‘other’ overriding self.

Parameters:

other (Metadata)

Return type:

Metadata

class src.object.ObjectMixin(*args, **kwargs)

Bases: ABC

Abstract base class providing metadata handling and type/dimension validation.

Parameters:

  • args (Any)

  • kwargs (Any)

abstractmethod classmethod concat_with_metadata(arrs, axis=0)

Concatenates a Sequence of this class, retaining the metadata of the first in the sequence. Returns an instance of this class with data concatentated.

Parameters:

  • arrs (Sequence[Self]) – Sequence of instances to concatenate.

  • axis (int) – Axis to concatentate on. Axis must exist for each instance in arrs.

Returns:

An instance of this class with data concatentated.

Return type:

Self

fixed_dtype: Type[generic]
fixed_ndim: int
metadata: Metadata
src.object.chunk(arr, bsize=None)
Parameters:

  • arr (T)

  • bsize (int | None)

Return type:

Iterator[T]