Block View

class BlockView

View data as pattern-aligned blocks.

__init__(pattern_provider: CIT_DataPatterned, cache_id: object | None, x_blocks: ndarray, y_blocks: ndarray, z_blocks: ndarray) None

Construct block-view from reference to pattern-provider and block-views of X, Y, Z.

See also

In most cases, the user should use CIT_DataPatterned.view_blocks() instead. This constructor is primarily for internal use.

pattern_provider: CIT_DataPatterned

Pattern-provider used to generate this view. Primarily for internal use in convenience-functions like match_blocksize().

cache_id: object | None

unique identifier associated to the data by the data-manager, to be used for caching results of tests. None to disable caching (eg for bootstrap)

x_blocks: ndarray

shape=(n,B) with n the block-count, B the block-size

y_blocks: ndarray

shape=(n,B) with n the block-count, B the block-size

z_blocks: ndarray

shape=(n,B,k) with n the block-count, B the block-size, k=dim(Z)

copy_and_center() BlockView

Copy and subtract mean. (Used internally by cit to structure residuals correctly, no cach-id required.)

Returns:

Centered copy

Return type:

BlockView

block_size() int

Get block-size (total number of samples per block).

Returns:

block-size

Return type:

int

block_count() int

Get block-count.

Returns:

block-count

Return type:

int

sample_count_used() int

Get number of used (contained in a block) data-points.

Returns:

used sample-size

Return type:

int

z_dim() int

Get z-dimension (number of variables in the conditioning set Z).

Returns:

dim(Z)

Return type:

int

trivialize() BlockView

View data as trivial (a single block of block-size=sampe-size) blocks.

Returns:

trivial block-view

Return type:

BlockView

match_blocksize(other: BlockView) BlockView

View data as blocks with size matching another block-view.

Parameters:

other (BlockView) – block-view whose block-size should be matched

Returns:

block-view of matching size

Return type:

BlockView

apply_blockformat(X: ndarray, Y: ndarray, Z: ndarray | None = None) BlockView

Given data for X, Y, Z, match to current block-size settings. (Used internally by cit to structure residuals correctly, no cach-id required.)

Parameters:
  • X (np.ndarray) – X

  • Y (np.ndarray) – Y

  • Z (np.ndarray | None, optional) – Z, defaults to None

Returns:

block-view of given data matching current size

Return type:

BlockView

bootstrap_unaligned_blocks(rng: Generator, bootstrap_block_count: int) BlockView

Bootstrap random (unaligned) blocks.

Parameters:
  • rng (np.random.Generator (or similar, requires .integers behaving as numpy.random.Generator)) – A random number generator.

  • bootstrap_block_count (int) – Number of blocks to bootstrap

Returns:

The bootstrapped blocks (actually a copy, not a view, despite being typed as block-“view”).

Return type:

BlockView