Class Blockwise

dask/blockwise.py:465–820

Tensor Operation This is a lazily constructed mapping for tensor operation graphs. This defines a dictionary using an operation and an indexing pattern. It is built for many operations like elementwise, transpose, tensordot, and so on. We choose to keep these as symbolic mappings r

class Blockwise(Layer):
    """Tensor Operation

    This is a lazily constructed mapping for tensor operation graphs.
    This defines a dictionary using an operation and an indexing pattern.
    It is built for many operations like elementwise, transpose, tensordot, and
    so on.  We choose to keep these as symbolic mappings rather than raw
    dictionaries because we are able to fuse them during optimization,
    sometimes resulting in much lower overhead.

    Parameters
    ----------
    output: str
        The name of the output collection.  Used in keynames
    output_indices: tuple
        The output indices, like ``('i', 'j', 'k')`` used to determine the
        structure of the block computations
    dsk: dict
        A small graph to apply per-output-block.  May include keys from the
        input indices.
    indices: tuple[tuple[str, tuple[str, ...] | None], ...]
        An ordered mapping from input key name, like ``'x'``
        to input indices, like ``('i', 'j')``
        Or includes literals, which have ``None`` for an index value.
        In place of input-key names, the first tuple element may also be a
        ``BlockwiseDep`` object.
    numblocks: Mapping[key, Sequence[int]]
        Number of blocks along each dimension for each input
    concatenate: bool
        Whether or not to pass contracted dimensions as a list of inputs or a
        single input to the block function
    new_axes: Mapping
        New index dimensions that may have been created and their size,
        e.g. ``{'j': 2, 'k': 3}``
    output_blocks: set[tuple[int, ...]]
        Specify a specific set of required output blocks. Since the graph
        will only contain the necessary tasks to generate these outputs,
        this kwarg can be used to "cull" the abstract layer (without needing
        to materialize the low-level graph).
    annotations: dict (optional)
        Layer annotations
    io_deps: dict[str, BlockwiseDep] (optional)
        Dictionary containing the mapping between "place-holder" collection
        keys and ``BlockwiseDep``-based objects.
        **WARNING**: This argument should only be used internally (for culling,
        fusion and cloning of existing Blockwise layers). Explicit use of this
        argument will be deprecated in the future.

    See Also
    --------
    dask.blockwise.blockwise
    dask.array.blockwise
    """

    output: str
    output_indices: tuple[str, ...]
    task: Task
    indices: tuple[tuple[str | TaskRef, tuple[str, ...] | None], ...]