xref: /dports/devel/py-dask/dask-2021.11.2/dask/array/core.py

from __future__ import annotations

import contextlib
import math
import operator
import os
import pickle
import re
import sys
import traceback
import uuid
import warnings
from bisect import bisect
from collections.abc import (
    Collection,
    Hashable,
    Iterable,
    Iterator,
    Mapping,
    MutableMapping,
)
from functools import partial, reduce, wraps
from itertools import product, zip_longest
from numbers import Integral, Number
from operator import add, mul
from threading import Lock
from typing import Any, Sequence

import numpy as np
from fsspec import get_mapper
from tlz import accumulate, concat, first, frequencies, groupby, partition
from tlz.curried import pluck

from .. import compute, config, core, threaded
from ..base import (
    DaskMethodsMixin,
    compute_as_if_collection,
    dont_optimize,
    is_dask_collection,
    persist,
    tokenize,
)
from ..blockwise import broadcast_dimensions
from ..context import globalmethod
from ..core import quote
from ..delayed import Delayed, delayed
from ..highlevelgraph import HighLevelGraph
from ..layers import reshapelist
from ..sizeof import sizeof
from ..utils import (
    IndexCallable,
    M,
    SerializableLock,
    cached_property,
    concrete,
    derived_from,
    factors,
    format_bytes,
    funcname,
    has_keyword,
    is_arraylike,
    is_dataframe_like,
    is_index_like,
    is_integer,
    is_series_like,
    ndeepmap,
    ndimlist,
    parse_bytes,
    typename,
)
from ..widgets import get_template
from . import chunk
from .chunk import getitem
from .chunk_types import is_valid_array_chunk, is_valid_chunk_type

# Keep einsum_lookup and tensordot_lookup here for backwards compatibility
from .dispatch import concatenate_lookup, einsum_lookup, tensordot_lookup  # noqa: F401
from .numpy_compat import _numpy_120, _Recurser
from .slicing import cached_cumsum, replace_ellipsis, setitem_array, slice_array

config.update_defaults({"array": {"chunk-size": "128MiB", "rechunk-threshold": 4}})

unknown_chunk_message = (
    "\n\n"
    "A possible solution: "
    "https://docs.dask.org/en/latest/array-chunks.html#unknown-chunks\n"
    "Summary: to compute chunks sizes, use\n\n"
    "   x.compute_chunk_sizes()  # for Dask Array `x`\n"
    "   ddf.to_dask_array(lengths=True)  # for Dask DataFrame `ddf`"
)


class PerformanceWarning(Warning):
    """A warning given when bad chunking may cause poor performance"""


def getter(a, b, asarray=True, lock=None):
    if isinstance(b, tuple) and any(x is None for x in b):
        b2 = tuple(x for x in b if x is not None)
        b3 = tuple(
            None if x is None else slice(None, None)
            for x in b
            if not isinstance(x, Integral)
        )
        return getter(a, b2, asarray=asarray, lock=lock)[b3]

    if lock:
        lock.acquire()
    try:
        c = a[b]
        # Below we special-case `np.matrix` to force a conversion to
        # `np.ndarray` and preserve original Dask behavior for `getter`,
        # as for all purposes `np.matrix` is array-like and thus
        # `is_arraylike` evaluates to `True` in that case.
        if asarray and (not is_arraylike(c) or isinstance(c, np.matrix)):
            c = np.asarray(c)
    finally:
        if lock:
            lock.release()
    return c


def getter_nofancy(a, b, asarray=True, lock=None):
    """A simple wrapper around ``getter``.

    Used to indicate to the optimization passes that the backend doesn't
    support fancy indexing.
    """
    return getter(a, b, asarray=asarray, lock=lock)


def getter_inline(a, b, asarray=True, lock=None):
    """A getter function that optimizations feel comfortable inlining

    Slicing operations with this function may be inlined into a graph, such as
    in the following rewrite

    **Before**

    >>> a = x[:10]  # doctest: +SKIP
    >>> b = a + 1  # doctest: +SKIP
    >>> c = a * 2  # doctest: +SKIP

    **After**

    >>> b = x[:10] + 1  # doctest: +SKIP
    >>> c = x[:10] * 2  # doctest: +SKIP

    This inlining can be relevant to operations when running off of disk.
    """
    return getter(a, b, asarray=asarray, lock=lock)


from .optimization import fuse_slice, optimize

# __array_function__ dict for mapping aliases and mismatching names
_HANDLED_FUNCTIONS = {}


def implements(*numpy_functions):
    """Register an __array_function__ implementation for dask.array.Array

    Register that a function implements the API of a NumPy function (or several
    NumPy functions in case of aliases) which is handled with
    ``__array_function__``.

    Parameters
    ----------
    \\*numpy_functions : callables
        One or more NumPy functions that are handled by ``__array_function__``
        and will be mapped by `implements` to a `dask.array` function.
    """

    def decorator(dask_func):
        for numpy_function in numpy_functions:
            _HANDLED_FUNCTIONS[numpy_function] = dask_func

        return dask_func

    return decorator


def _should_delegate(other) -> bool:
    """Check whether Dask should delegate to the other.
    This implementation follows NEP-13:
    https://numpy.org/neps/nep-0013-ufunc-overrides.html#behavior-in-combination-with-python-s-binary-operations
    """
    if hasattr(other, "__array_ufunc__") and other.__array_ufunc__ is None:
        return True
    elif (
        hasattr(other, "__array_ufunc__")
        and not is_valid_array_chunk(other)
        and type(other).__array_ufunc__ is not Array.__array_ufunc__
    ):
        return True
    return False


def check_if_handled_given_other(f):
    """Check if method is handled by Dask given type of other

    Ensures proper deferral to upcast types in dunder operations without
    assuming unknown types are automatically downcast types.
    """

    @wraps(f)
    def wrapper(self, other):
        if _should_delegate(other):
            return NotImplemented
        else:
            return f(self, other)

    return wrapper


def slices_from_chunks(chunks):
    """Translate chunks tuple to a set of slices in product order

    >>> slices_from_chunks(((2, 2), (3, 3, 3)))  # doctest: +NORMALIZE_WHITESPACE
     [(slice(0, 2, None), slice(0, 3, None)),
      (slice(0, 2, None), slice(3, 6, None)),
      (slice(0, 2, None), slice(6, 9, None)),
      (slice(2, 4, None), slice(0, 3, None)),
      (slice(2, 4, None), slice(3, 6, None)),
      (slice(2, 4, None), slice(6, 9, None))]
    """
    cumdims = [cached_cumsum(bds, initial_zero=True) for bds in chunks]
    slices = [
        [slice(s, s + dim) for s, dim in zip(starts, shapes)]
        for starts, shapes in zip(cumdims, chunks)
    ]
    return list(product(*slices))


def getem(
    arr,
    chunks,
    getitem=getter,
    shape=None,
    out_name=None,
    lock=False,
    asarray=True,
    dtype=None,
):
    """Dask getting various chunks from an array-like

    >>> getem('X', chunks=(2, 3), shape=(4, 6))  # doctest: +SKIP
    {('X', 0, 0): (getter, 'X', (slice(0, 2), slice(0, 3))),
     ('X', 1, 0): (getter, 'X', (slice(2, 4), slice(0, 3))),
     ('X', 1, 1): (getter, 'X', (slice(2, 4), slice(3, 6))),
     ('X', 0, 1): (getter, 'X', (slice(0, 2), slice(3, 6)))}

    >>> getem('X', chunks=((2, 2), (3, 3)))  # doctest: +SKIP
    {('X', 0, 0): (getter, 'X', (slice(0, 2), slice(0, 3))),
     ('X', 1, 0): (getter, 'X', (slice(2, 4), slice(0, 3))),
     ('X', 1, 1): (getter, 'X', (slice(2, 4), slice(3, 6))),
     ('X', 0, 1): (getter, 'X', (slice(0, 2), slice(3, 6)))}
    """
    out_name = out_name or arr
    chunks = normalize_chunks(chunks, shape, dtype=dtype)
    keys = product([out_name], *(range(len(bds)) for bds in chunks))
    slices = slices_from_chunks(chunks)

    if (
        has_keyword(getitem, "asarray")
        and has_keyword(getitem, "lock")
        and (not asarray or lock)
    ):
        values = [(getitem, arr, x, asarray, lock) for x in slices]
    else:
        # Common case, drop extra parameters
        values = [(getitem, arr, x) for x in slices]

    return dict(zip(keys, values))


def dotmany(A, B, leftfunc=None, rightfunc=None, **kwargs):
    """Dot product of many aligned chunks

    >>> x = np.array([[1, 2], [1, 2]])
    >>> y = np.array([[10, 20], [10, 20]])
    >>> dotmany([x, x, x], [y, y, y])
    array([[ 90, 180],
           [ 90, 180]])

    Optionally pass in functions to apply to the left and right chunks

    >>> dotmany([x, x, x], [y, y, y], rightfunc=np.transpose)
    array([[150, 150],
           [150, 150]])
    """
    if leftfunc:
        A = map(leftfunc, A)
    if rightfunc:
        B = map(rightfunc, B)
    return sum(map(partial(np.dot, **kwargs), A, B))


def _concatenate2(arrays, axes=[]):
    """Recursively concatenate nested lists of arrays along axes

    Each entry in axes corresponds to each level of the nested list.  The
    length of axes should correspond to the level of nesting of arrays.
    If axes is an empty list or tuple, return arrays, or arrays[0] if
    arrays is a list.

    >>> x = np.array([[1, 2], [3, 4]])
    >>> _concatenate2([x, x], axes=[0])
    array([[1, 2],
           [3, 4],
           [1, 2],
           [3, 4]])

    >>> _concatenate2([x, x], axes=[1])
    array([[1, 2, 1, 2],
           [3, 4, 3, 4]])

    >>> _concatenate2([[x, x], [x, x]], axes=[0, 1])
    array([[1, 2, 1, 2],
           [3, 4, 3, 4],
           [1, 2, 1, 2],
           [3, 4, 3, 4]])

    Supports Iterators
    >>> _concatenate2(iter([x, x]), axes=[1])
    array([[1, 2, 1, 2],
           [3, 4, 3, 4]])

    Special Case
    >>> _concatenate2([x, x], axes=())
    array([[1, 2],
           [3, 4]])
    """
    if axes == ():
        if isinstance(arrays, list):
            return arrays[0]
        else:
            return arrays

    if isinstance(arrays, Iterator):
        arrays = list(arrays)
    if not isinstance(arrays, (list, tuple)):
        return arrays
    if len(axes) > 1:
        arrays = [_concatenate2(a, axes=axes[1:]) for a in arrays]
    concatenate = concatenate_lookup.dispatch(
        type(max(arrays, key=lambda x: getattr(x, "__array_priority__", 0)))
    )
    if isinstance(arrays[0], dict):
        # Handle concatenation of `dict`s, used as a replacement for structured
        # arrays when that's not supported by the array library (e.g., CuPy).
        keys = list(arrays[0].keys())
        assert all(list(a.keys()) == keys for a in arrays)
        ret = dict()
        for k in keys:
            ret[k] = concatenate(list(a[k] for a in arrays), axis=axes[0])
        return ret
    else:
        return concatenate(arrays, axis=axes[0])


def apply_infer_dtype(func, args, kwargs, funcname, suggest_dtype="dtype", nout=None):
    """
    Tries to infer output dtype of ``func`` for a small set of input arguments.

    Parameters
    ----------
    func: Callable
        Function for which output dtype is to be determined

    args: List of array like
        Arguments to the function, which would usually be used. Only attributes
        ``ndim`` and ``dtype`` are used.

    kwargs: dict
        Additional ``kwargs`` to the ``func``

    funcname: String
        Name of calling function to improve potential error messages

    suggest_dtype: None/False or String
        If not ``None`` adds suggestion to potential error message to specify a dtype
        via the specified kwarg. Defaults to ``'dtype'``.

    nout: None or Int
        ``None`` if function returns single output, integer if many.
        Deafults to ``None``.

    Returns
    -------
    : dtype or List of dtype
        One or many dtypes (depending on ``nout``)
    """
    from .utils import meta_from_array

    # make sure that every arg is an evaluated array
    args = [
        np.ones_like(meta_from_array(x), shape=((1,) * x.ndim), dtype=x.dtype)
        if is_arraylike(x)
        else x
        for x in args
    ]
    try:
        with np.errstate(all="ignore"):
            o = func(*args, **kwargs)
    except Exception as e:
        exc_type, exc_value, exc_traceback = sys.exc_info()
        tb = "".join(traceback.format_tb(exc_traceback))
        suggest = (
            (
                "Please specify the dtype explicitly using the "
                "`{dtype}` kwarg.\n\n".format(dtype=suggest_dtype)
            )
            if suggest_dtype
            else ""
        )
        msg = (
            f"`dtype` inference failed in `{funcname}`.\n\n"
            f"{suggest}"
            "Original error is below:\n"
            "------------------------\n"
            f"{e!r}\n\n"
            "Traceback:\n"
            "---------\n"
            f"{tb}"
        )
    else:
        msg = None
    if msg is not None:
        raise ValueError(msg)
    return o.dtype if nout is None else tuple(e.dtype for e in o)


def normalize_arg(x):
    """Normalize user provided arguments to blockwise or map_blocks

    We do a few things:

    1.  If they are string literals that might collide with blockwise_token then we
        quote them
    2.  IF they are large (as defined by sizeof) then we put them into the
        graph on their own by using dask.delayed
    """
    if is_dask_collection(x):
        return x
    elif isinstance(x, str) and re.match(r"_\d+", x):
        return delayed(x)
    elif isinstance(x, list) and len(x) >= 10:
        return delayed(x)
    elif sizeof(x) > 1e6:
        return delayed(x)
    else:
        return x


def _pass_extra_kwargs(func, keys, *args, **kwargs):
    """Helper for :func:`dask.array.map_blocks` to pass `block_info` or `block_id`.

    For each element of `keys`, a corresponding element of args is changed
    to a keyword argument with that key, before all arguments re passed on
    to `func`.
    """
    kwargs.update(zip(keys, args))
    return func(*args[len(keys) :], **kwargs)


def map_blocks(
    func,
    *args,
    name=None,
    token=None,
    dtype=None,
    chunks=None,
    drop_axis=[],
    new_axis=None,
    meta=None,
    **kwargs,
):
    """Map a function across all blocks of a dask array.

    Note that ``map_blocks`` will attempt to automatically determine the output
    array type by calling ``func`` on 0-d versions of the inputs. Please refer to
    the ``meta`` keyword argument below if you expect that the function will not
    succeed when operating on 0-d arrays.

    Parameters
    ----------
    func : callable
        Function to apply to every block in the array.
    args : dask arrays or other objects
    dtype : np.dtype, optional
        The ``dtype`` of the output array. It is recommended to provide this.
        If not provided, will be inferred by applying the function to a small
        set of fake data.
    chunks : tuple, optional
        Chunk shape of resulting blocks if the function does not preserve
        shape. If not provided, the resulting array is assumed to have the same
        block structure as the first input array.
    drop_axis : number or iterable, optional
        Dimensions lost by the function.
    new_axis : number or iterable, optional
        New dimensions created by the function. Note that these are applied
        after ``drop_axis`` (if present).
    token : string, optional
        The key prefix to use for the output array. If not provided, will be
        determined from the function name.
    name : string, optional
        The key name to use for the output array. Note that this fully
        specifies the output key name, and must be unique. If not provided,
        will be determined by a hash of the arguments.
    meta : array-like, optional
        The ``meta`` of the output array, when specified is expected to be an
        array of the same type and dtype of that returned when calling ``.compute()``
        on the array returned by this function. When not provided, ``meta`` will be
        inferred by applying the function to a small set of fake data, usually a
        0-d array. It's important to ensure that ``func`` can successfully complete
        computation without raising exceptions when 0-d is passed to it, providing
        ``meta`` will be required otherwise. If the output type is known beforehand
        (e.g., ``np.ndarray``, ``cupy.ndarray``), an empty array of such type dtype
        can be passed, for example: ``meta=np.array((), dtype=np.int32)``.
    **kwargs :
        Other keyword arguments to pass to function. Values must be constants
        (not dask.arrays)

    See Also
    --------
    dask.array.blockwise : Generalized operation with control over block alignment.

    Examples
    --------
    >>> import dask.array as da
    >>> x = da.arange(6, chunks=3)

    >>> x.map_blocks(lambda x: x * 2).compute()
    array([ 0,  2,  4,  6,  8, 10])

    The ``da.map_blocks`` function can also accept multiple arrays.

    >>> d = da.arange(5, chunks=2)
    >>> e = da.arange(5, chunks=2)

    >>> f = da.map_blocks(lambda a, b: a + b**2, d, e)
    >>> f.compute()
    array([ 0,  2,  6, 12, 20])

    If the function changes shape of the blocks then you must provide chunks
    explicitly.

    >>> y = x.map_blocks(lambda x: x[::2], chunks=((2, 2),))

    You have a bit of freedom in specifying chunks.  If all of the output chunk
    sizes are the same, you can provide just that chunk size as a single tuple.

    >>> a = da.arange(18, chunks=(6,))
    >>> b = a.map_blocks(lambda x: x[:3], chunks=(3,))

    If the function changes the dimension of the blocks you must specify the
    created or destroyed dimensions.

    >>> b = a.map_blocks(lambda x: x[None, :, None], chunks=(1, 6, 1),
    ...                  new_axis=[0, 2])

    If ``chunks`` is specified but ``new_axis`` is not, then it is inferred to
    add the necessary number of axes on the left.

    Map_blocks aligns blocks by block positions without regard to shape. In the
    following example we have two arrays with the same number of blocks but
    with different shape and chunk sizes.

    >>> x = da.arange(1000, chunks=(100,))
    >>> y = da.arange(100, chunks=(10,))

    The relevant attribute to match is numblocks.

    >>> x.numblocks
    (10,)
    >>> y.numblocks
    (10,)

    If these match (up to broadcasting rules) then we can map arbitrary
    functions across blocks

    >>> def func(a, b):
    ...     return np.array([a.max(), b.max()])

    >>> da.map_blocks(func, x, y, chunks=(2,), dtype='i8')
    dask.array<func, shape=(20,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>

    >>> _.compute()
    array([ 99,   9, 199,  19, 299,  29, 399,  39, 499,  49, 599,  59, 699,
            69, 799,  79, 899,  89, 999,  99])

    Your block function get information about where it is in the array by
    accepting a special ``block_info`` or ``block_id`` keyword argument.

    >>> def func(block_info=None):
    ...     pass

    This will receive the following information:

    >>> block_info  # doctest: +SKIP
    {0: {'shape': (1000,),
         'num-chunks': (10,),
         'chunk-location': (4,),
         'array-location': [(400, 500)]},
     None: {'shape': (1000,),
            'num-chunks': (10,),
            'chunk-location': (4,),
            'array-location': [(400, 500)],
            'chunk-shape': (100,),
            'dtype': dtype('float64')}}

    For each argument and keyword arguments that are dask arrays (the positions
    of which are the first index), you will receive the shape of the full
    array, the number of chunks of the full array in each dimension, the chunk
    location (for example the fourth chunk over in the first dimension), and
    the array location (for example the slice corresponding to ``40:50``). The
    same information is provided for the output, with the key ``None``, plus
    the shape and dtype that should be returned.

    These features can be combined to synthesize an array from scratch, for
    example:

    >>> def func(block_info=None):
    ...     loc = block_info[None]['array-location'][0]
    ...     return np.arange(loc[0], loc[1])

    >>> da.map_blocks(func, chunks=((4, 4),), dtype=np.float_)
    dask.array<func, shape=(8,), dtype=float64, chunksize=(4,), chunktype=numpy.ndarray>

    >>> _.compute()
    array([0, 1, 2, 3, 4, 5, 6, 7])

    ``block_id`` is similar to ``block_info`` but contains only the ``chunk_location``:

    >>> def func(block_id=None):
    ...     pass

    This will receive the following information:

    >>> block_id  # doctest: +SKIP
    (4, 3)

    You may specify the key name prefix of the resulting task in the graph with
    the optional ``token`` keyword argument.

    >>> x.map_blocks(lambda x: x + 1, name='increment')
    dask.array<increment, shape=(1000,), dtype=int64, chunksize=(100,), chunktype=numpy.ndarray>

    For functions that may not handle 0-d arrays, it's also possible to specify
    ``meta`` with an empty array matching the type of the expected result. In
    the example below, ``func`` will result in an ``IndexError`` when computing
    ``meta``:

    >>> da.map_blocks(lambda x: x[2], da.random.random(5), meta=np.array(()))
    dask.array<lambda, shape=(5,), dtype=float64, chunksize=(5,), chunktype=numpy.ndarray>

    Similarly, it's possible to specify a non-NumPy array to ``meta``, and provide
    a ``dtype``:

    >>> import cupy  # doctest: +SKIP
    >>> rs = da.random.RandomState(RandomState=cupy.random.RandomState)  # doctest: +SKIP
    >>> dt = np.float32
    >>> da.map_blocks(lambda x: x[2], rs.random(5, dtype=dt), meta=cupy.array((), dtype=dt))  # doctest: +SKIP
    dask.array<lambda, shape=(5,), dtype=float32, chunksize=(5,), chunktype=cupy.ndarray>
    """
    if not callable(func):
        msg = (
            "First argument must be callable function, not %s\n"
            "Usage:   da.map_blocks(function, x)\n"
            "   or:   da.map_blocks(function, x, y, z)"
        )
        raise TypeError(msg % type(func).__name__)
    if token:
        warnings.warn("The token= keyword to map_blocks has been moved to name=")
        name = token

    name = f"{name or funcname(func)}-{tokenize(func, *args, **kwargs)}"
    new_axes = {}

    if isinstance(drop_axis, Number):
        drop_axis = [drop_axis]
    if isinstance(new_axis, Number):
        new_axis = [new_axis]  # TODO: handle new_axis

    arrs = [a for a in args if isinstance(a, Array)]

    argpairs = [
        (a, tuple(range(a.ndim))[::-1]) if isinstance(a, Array) else (a, None)
        for a in args
    ]
    if arrs:
        out_ind = tuple(range(max(a.ndim for a in arrs)))[::-1]
    else:
        out_ind = ()

    original_kwargs = kwargs

    if dtype is None and meta is None:
        try:
            meta = compute_meta(func, dtype, *args, **kwargs)
        except Exception:
            pass

        dtype = apply_infer_dtype(func, args, original_kwargs, "map_blocks")

    if drop_axis:
        ndim_out = len(out_ind)
        if any(i < -ndim_out or i >= ndim_out for i in drop_axis):
            raise ValueError(
                f"drop_axis out of range (drop_axis={drop_axis}, "
                f"but output is {ndim_out}d)."
            )
        drop_axis = [i % ndim_out for i in drop_axis]
        out_ind = tuple(x for i, x in enumerate(out_ind) if i not in drop_axis)
    if new_axis is None and chunks is not None and len(out_ind) < len(chunks):
        new_axis = range(len(chunks) - len(out_ind))
    if new_axis:
        # new_axis = [x + len(drop_axis) for x in new_axis]
        out_ind = list(out_ind)
        for ax in sorted(new_axis):
            n = len(out_ind) + len(drop_axis)
            out_ind.insert(ax, n)
            if chunks is not None:
                new_axes[n] = chunks[ax]
            else:
                new_axes[n] = 1
        out_ind = tuple(out_ind)
        if max(new_axis) > max(out_ind):
            raise ValueError("New_axis values do not fill in all dimensions")

    if chunks is not None:
        if len(chunks) != len(out_ind):
            raise ValueError(
                f"Provided chunks have {len(chunks)} dims; expected {len(out_ind)} dims"
            )
        adjust_chunks = dict(zip(out_ind, chunks))
    else:
        adjust_chunks = None

    out = blockwise(
        func,
        out_ind,
        *concat(argpairs),
        name=name,
        new_axes=new_axes,
        dtype=dtype,
        concatenate=True,
        align_arrays=False,
        adjust_chunks=adjust_chunks,
        meta=meta,
        **kwargs,
    )

    extra_argpairs = []
    extra_names = []
    # If func has block_id as an argument, construct an array of block IDs and
    # prepare to inject it.
    if has_keyword(func, "block_id"):
        block_id_name = "block-id-" + out.name
        block_id_dsk = {
            (block_id_name,) + block_id: block_id
            for block_id in product(*(range(len(c)) for c in out.chunks))
        }
        block_id_array = Array(
            block_id_dsk,
            block_id_name,
            chunks=tuple((1,) * len(c) for c in out.chunks),
            dtype=np.object_,
        )
        extra_argpairs.append((block_id_array, out_ind))
        extra_names.append("block_id")

    # If func has block_info as an argument, construct an array of block info
    # objects and prepare to inject it.
    if has_keyword(func, "block_info"):
        starts = {}
        num_chunks = {}
        shapes = {}

        for i, (arg, in_ind) in enumerate(argpairs):
            if in_ind is not None:
                shapes[i] = arg.shape
                if drop_axis:
                    # We concatenate along dropped axes, so we need to treat them
                    # as if there is only a single chunk.
                    starts[i] = [
                        (
                            cached_cumsum(arg.chunks[j], initial_zero=True)
                            if ind in out_ind
                            else [0, arg.shape[j]]
                        )
                        for j, ind in enumerate(in_ind)
                    ]
                    num_chunks[i] = tuple(len(s) - 1 for s in starts[i])
                else:
                    starts[i] = [
                        cached_cumsum(c, initial_zero=True) for c in arg.chunks
                    ]
                    num_chunks[i] = arg.numblocks
        out_starts = [cached_cumsum(c, initial_zero=True) for c in out.chunks]

        block_info_name = "block-info-" + out.name
        block_info_dsk = {}
        for block_id in product(*(range(len(c)) for c in out.chunks)):
            # Get position of chunk, indexed by axis labels
            location = {out_ind[i]: loc for i, loc in enumerate(block_id)}
            info = {}
            for i, shape in shapes.items():
                # Compute chunk key in the array, taking broadcasting into
                # account. We don't directly know which dimensions are
                # broadcast, but any dimension with only one chunk can be
                # treated as broadcast.
                arr_k = tuple(
                    location.get(ind, 0) if num_chunks[i][j] > 1 else 0
                    for j, ind in enumerate(argpairs[i][1])
                )
                info[i] = {
                    "shape": shape,
                    "num-chunks": num_chunks[i],
                    "array-location": [
                        (starts[i][ij][j], starts[i][ij][j + 1])
                        for ij, j in enumerate(arr_k)
                    ],
                    "chunk-location": arr_k,
                }

            info[None] = {
                "shape": out.shape,
                "num-chunks": out.numblocks,
                "array-location": [
                    (out_starts[ij][j], out_starts[ij][j + 1])
                    for ij, j in enumerate(block_id)
                ],
                "chunk-location": block_id,
                "chunk-shape": tuple(
                    out.chunks[ij][j] for ij, j in enumerate(block_id)
                ),
                "dtype": dtype,
            }
            block_info_dsk[(block_info_name,) + block_id] = info

        block_info = Array(
            block_info_dsk,
            block_info_name,
            chunks=tuple((1,) * len(c) for c in out.chunks),
            dtype=np.object_,
        )
        extra_argpairs.append((block_info, out_ind))
        extra_names.append("block_info")

    if extra_argpairs:
        # Rewrite the Blockwise layer. It would be nice to find a way to
        # avoid doing it twice, but it's currently needed to determine
        # out.chunks from the first pass. Since it constructs a Blockwise
        # rather than an expanded graph, it shouldn't be too expensive.
        out = blockwise(
            _pass_extra_kwargs,
            out_ind,
            func,
            None,
            tuple(extra_names),
            None,
            *concat(extra_argpairs),
            *concat(argpairs),
            name=out.name,
            dtype=out.dtype,
            concatenate=True,
            align_arrays=False,
            adjust_chunks=dict(zip(out_ind, out.chunks)),
            meta=meta,
            **kwargs,
        )

    return out


def broadcast_chunks(*chunkss):
    """Construct a chunks tuple that broadcasts many chunks tuples

    >>> a = ((5, 5),)
    >>> b = ((5, 5),)
    >>> broadcast_chunks(a, b)
    ((5, 5),)

    >>> a = ((10, 10, 10), (5, 5),)
    >>> b = ((5, 5),)
    >>> broadcast_chunks(a, b)
    ((10, 10, 10), (5, 5))

    >>> a = ((10, 10, 10), (5, 5),)
    >>> b = ((1,), (5, 5),)
    >>> broadcast_chunks(a, b)
    ((10, 10, 10), (5, 5))

    >>> a = ((10, 10, 10), (5, 5),)
    >>> b = ((3, 3,), (5, 5),)
    >>> broadcast_chunks(a, b)
    Traceback (most recent call last):
        ...
    ValueError: Chunks do not align: [(10, 10, 10), (3, 3)]
    """
    if not chunkss:
        return ()
    elif len(chunkss) == 1:
        return chunkss[0]
    n = max(map(len, chunkss))
    chunkss2 = [((1,),) * (n - len(c)) + c for c in chunkss]
    result = []
    for i in range(n):
        step1 = [c[i] for c in chunkss2]
        if all(c == (1,) for c in step1):
            step2 = step1
        else:
            step2 = [c for c in step1 if c != (1,)]
        if len(set(step2)) != 1:
            raise ValueError("Chunks do not align: %s" % str(step2))
        result.append(step2[0])
    return tuple(result)


def store(
    sources: Array | Collection[Array],
    targets,
    lock: bool | Lock = True,
    regions: tuple[slice, ...] | Collection[tuple[slice, ...]] | None = None,
    compute: bool = True,
    return_stored: bool = False,
    **kwargs,
):
    """Store dask arrays in array-like objects, overwrite data in target

    This stores dask arrays into object that supports numpy-style setitem
    indexing.  It stores values chunk by chunk so that it does not have to
    fill up memory.  For best performance you can align the block size of
    the storage target with the block size of your array.

    If your data fits in memory then you may prefer calling
    ``np.array(myarray)`` instead.

    Parameters
    ----------

    sources: Array or collection of Arrays
    targets: array-like or Delayed or collection of array-likes and/or Delayeds
        These should support setitem syntax ``target[10:20] = ...``
    lock: boolean or threading.Lock, optional
        Whether or not to lock the data stores while storing.
        Pass True (lock each file individually), False (don't lock) or a
        particular :class:`threading.Lock` object to be shared among all writes.
    regions: tuple of slices or collection of tuples of slices
        Each ``region`` tuple in ``regions`` should be such that
        ``target[region].shape = source.shape``
        for the corresponding source and target in sources and targets,
        respectively. If this is a tuple, the contents will be assumed to be
        slices, so do not provide a tuple of tuples.
    compute: boolean, optional
        If true compute immediately; return :class:`dask.delayed.Delayed` otherwise.
    return_stored: boolean, optional
        Optionally return the stored result (default False).
    kwargs:
        Parameters passed to compute/persist (only used if compute=True)

    Returns
    -------

    If return_stored=True
        tuple of Arrays
    If return_stored=False and compute=True
        None
    If return_stored=False and compute=False
        Delayed

    Examples
    --------

    >>> import h5py  # doctest: +SKIP
    >>> f = h5py.File('myfile.hdf5', mode='a')  # doctest: +SKIP
    >>> dset = f.create_dataset('/data', shape=x.shape,
    ...                                  chunks=x.chunks,
    ...                                  dtype='f8')  # doctest: +SKIP

    >>> store(x, dset)  # doctest: +SKIP

    Alternatively store many arrays at the same time

    >>> store([x, y, z], [dset1, dset2, dset3])  # doctest: +SKIP
    """

    if isinstance(sources, Array):
        sources = [sources]
        targets = [targets]

    if any(not isinstance(s, Array) for s in sources):
        raise ValueError("All sources must be dask array objects")

    if len(sources) != len(targets):
        raise ValueError(
            "Different number of sources [%d] and targets [%d]"
            % (len(sources), len(targets))
        )

    if isinstance(regions, tuple) or regions is None:
        regions = [regions]

    if len(sources) > 1 and len(regions) == 1:
        regions *= len(sources)

    if len(sources) != len(regions):
        raise ValueError(
            "Different number of sources [%d] and targets [%d] than regions [%d]"
            % (len(sources), len(targets), len(regions))
        )

    # Optimize all sources together
    sources_hlg = HighLevelGraph.merge(*[e.__dask_graph__() for e in sources])
    sources_layer = Array.__dask_optimize__(
        sources_hlg, list(core.flatten([e.__dask_keys__() for e in sources]))
    )
    sources_name = "store-sources-" + tokenize(sources)
    layers = {sources_name: sources_layer}
    dependencies = {sources_name: set()}

    # Optimize all targets together
    targets_keys = []
    targets_dsks = []
    for t in targets:
        if isinstance(t, Delayed):
            targets_keys.append(t.key)
            targets_dsks.append(t.__dask_graph__())
        elif is_dask_collection(t):
            raise TypeError("Targets must be either Delayed objects or array-likes")

    if targets_dsks:
        targets_hlg = HighLevelGraph.merge(*targets_dsks)
        targets_layer = Delayed.__dask_optimize__(targets_hlg, targets_keys)
        targets_name = "store-targets-" + tokenize(targets_keys)
        layers[targets_name] = targets_layer
        dependencies[targets_name] = set()

    load_stored = return_stored and not compute

    map_names = [
        "store-map-" + tokenize(s, t if isinstance(t, Delayed) else id(t), r)
        for s, t, r in zip(sources, targets, regions)
    ]
    map_keys = []

    for s, t, n, r in zip(sources, targets, map_names, regions):
        map_layer = insert_to_ooc(
            keys=s.__dask_keys__(),
            chunks=s.chunks,
            out=t.key if isinstance(t, Delayed) else t,
            name=n,
            lock=lock,
            region=r,
            return_stored=return_stored,
            load_stored=load_stored,
        )
        layers[n] = map_layer
        if isinstance(t, Delayed):
            dependencies[n] = {sources_name, targets_name}
        else:
            dependencies[n] = {sources_name}
        map_keys += map_layer.keys()

    if return_stored:
        store_dsk = HighLevelGraph(layers, dependencies)
        load_store_dsk = store_dsk
        if compute:
            store_dlyds = [Delayed(k, store_dsk) for k in map_keys]
            store_dlyds = persist(*store_dlyds, **kwargs)
            store_dsk_2 = HighLevelGraph.merge(*[e.dask for e in store_dlyds])
            load_store_dsk = retrieve_from_ooc(map_keys, store_dsk, store_dsk_2)
            map_names = ["load-" + n for n in map_names]

        return tuple(
            Array(load_store_dsk, n, s.chunks, meta=s)
            for s, n in zip(sources, map_names)
        )

    elif compute:
        store_dsk = HighLevelGraph(layers, dependencies)
        compute_as_if_collection(Array, store_dsk, map_keys, **kwargs)
        return None

    else:
        key = "store-" + tokenize(map_names)
        layers[key] = {key: map_keys}
        dependencies[key] = set(map_names)
        store_dsk = HighLevelGraph(layers, dependencies)
        return Delayed(key, store_dsk)


def blockdims_from_blockshape(shape, chunks):
    """

    >>> blockdims_from_blockshape((10, 10), (4, 3))
    ((4, 4, 2), (3, 3, 3, 1))
    >>> blockdims_from_blockshape((10, 0), (4, 0))
    ((4, 4, 2), (0,))
    """
    if chunks is None:
        raise TypeError("Must supply chunks= keyword argument")
    if shape is None:
        raise TypeError("Must supply shape= keyword argument")
    if np.isnan(sum(shape)) or np.isnan(sum(chunks)):
        raise ValueError(
            "Array chunk sizes are unknown. shape: %s, chunks: %s%s"
            % (shape, chunks, unknown_chunk_message)
        )
    if not all(map(is_integer, chunks)):
        raise ValueError("chunks can only contain integers.")
    if not all(map(is_integer, shape)):
        raise ValueError("shape can only contain integers.")
    shape = tuple(map(int, shape))
    chunks = tuple(map(int, chunks))
    return tuple(
        ((bd,) * (d // bd) + ((d % bd,) if d % bd else ()) if d else (0,))
        for d, bd in zip(shape, chunks)
    )


def finalize(results):
    if not results:
        return concatenate3(results)
    results2 = results
    while isinstance(results2, (tuple, list)):
        if len(results2) > 1:
            return concatenate3(results)
        else:
            results2 = results2[0]
    return unpack_singleton(results)


CHUNKS_NONE_ERROR_MESSAGE = """
You must specify a chunks= keyword argument.
This specifies the chunksize of your array blocks.

See the following documentation page for details:
  https://docs.dask.org/en/latest/array-creation.html#chunks
""".strip()


class Array(DaskMethodsMixin):
    """Parallel Dask Array

    A parallel nd-array comprised of many numpy arrays arranged in a grid.

    This constructor is for advanced uses only.  For normal use see the
    :func:`dask.array.from_array` function.

    Parameters
    ----------
    dask : dict
        Task dependency graph
    name : string
        Name of array in dask
    shape : tuple of ints
        Shape of the entire array
    chunks: iterable of tuples
        block sizes along each dimension
    dtype : str or dtype
        Typecode or data-type for the new Dask Array
    meta : empty ndarray
        empty ndarray created with same NumPy backend, ndim and dtype as the
        Dask Array being created (overrides dtype)

    See Also
    --------
    dask.array.from_array
    """

    __slots__ = "dask", "__name", "_cached_keys", "__chunks", "_meta", "__dict__"

    def __new__(cls, dask, name, chunks, dtype=None, meta=None, shape=None):
        self = super().__new__(cls)
        assert isinstance(dask, Mapping)
        if not isinstance(dask, HighLevelGraph):
            dask = HighLevelGraph.from_collections(name, dask, dependencies=())
        self.dask = dask
        self._name = str(name)
        meta = meta_from_array(meta, dtype=dtype)

        if (
            isinstance(chunks, str)
            or isinstance(chunks, tuple)
            and chunks
            and any(isinstance(c, str) for c in chunks)
        ):
            dt = meta.dtype
        else:
            dt = None
        self._chunks = normalize_chunks(chunks, shape, dtype=dt)
        if self.chunks is None:
            raise ValueError(CHUNKS_NONE_ERROR_MESSAGE)
        self._meta = meta_from_array(meta, ndim=self.ndim, dtype=dtype)

        for plugin in config.get("array_plugins", ()):
            result = plugin(self)
            if result is not None:
                self = result

        try:
            layer = self.dask.layers[name]
        except (AttributeError, KeyError):
            # self is no longer an Array after applying the plugins, OR
            # a plugin replaced the HighLevelGraph with a plain dict, OR
            # name is not the top layer's name (this can happen after the layer is
            # manipulated, to avoid a collision)
            pass
        else:
            if layer.collection_annotations is None:
                layer.collection_annotations = {
                    "shape": self.shape,
                    "dtype": self.dtype,
                    "chunksize": self.chunksize,
                    "chunks": self.chunks,
                    "type": typename(type(self)),
                    "chunk_type": typename(type(self._meta)),
                }
            else:
                layer.collection_annotations.update(
                    {
                        "shape": self.shape,
                        "dtype": self.dtype,
                        "chunksize": self.chunksize,
                        "chunks": self.chunks,
                        "type": typename(type(self)),
                        "chunk_type": typename(type(self._meta)),
                    }
                )

        return self

    def __reduce__(self):
        return (Array, (self.dask, self.name, self.chunks, self.dtype))

    def __dask_graph__(self):
        return self.dask

    def __dask_layers__(self):
        return (self.name,)

    def __dask_keys__(self):
        if self._cached_keys is not None:
            return self._cached_keys

        name, chunks, numblocks = self.name, self.chunks, self.numblocks

        def keys(*args):
            if not chunks:
                return [(name,)]
            ind = len(args)
            if ind + 1 == len(numblocks):
                result = [(name,) + args + (i,) for i in range(numblocks[ind])]
            else:
                result = [keys(*(args + (i,))) for i in range(numblocks[ind])]
            return result

        self._cached_keys = result = keys()
        return result

    def __dask_tokenize__(self):
        return self.name

    __dask_optimize__ = globalmethod(
        optimize, key="array_optimize", falsey=dont_optimize
    )
    __dask_scheduler__ = staticmethod(threaded.get)

    def __dask_postcompute__(self):
        return finalize, ()

    def __dask_postpersist__(self):
        return self._rebuild, ()

    def _rebuild(self, dsk, *, rename=None):
        name = self._name
        if rename:
            name = rename.get(name, name)
        return Array(dsk, name, self.chunks, self.dtype, self._meta)

    def _reset_cache(self, key=None):
        """
        Reset cached properties.

        Parameters
        ----------
        key : str, optional
            Remove specified key. The default removes all items.
        """
        if key is None:
            self.__dict__.clear()
        else:
            self.__dict__.pop(key, None)

    @cached_property
    def _key_array(self):
        return np.array(self.__dask_keys__(), dtype=object)

    @cached_property
    def numblocks(self):
        return tuple(map(len, self.chunks))

    @cached_property
    def npartitions(self):
        return reduce(mul, self.numblocks, 1)

    def compute_chunk_sizes(self):
        """
        Compute the chunk sizes for a Dask array. This is especially useful
        when the chunk sizes are unknown (e.g., when indexing one Dask array
        with another).

        Notes
        -----
        This function modifies the Dask array in-place.

        Examples
        --------
        >>> import dask.array as da
        >>> import numpy as np
        >>> x = da.from_array([-2, -1, 0, 1, 2], chunks=2)
        >>> x.chunks
        ((2, 2, 1),)
        >>> y = x[x <= 0]
        >>> y.chunks
        ((nan, nan, nan),)
        >>> y.compute_chunk_sizes()  # in-place computation
        dask.array<getitem, shape=(3,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>
        >>> y.chunks
        ((2, 1, 0),)

        """
        x = self
        chunk_shapes = x.map_blocks(
            _get_chunk_shape,
            dtype=int,
            chunks=tuple(len(c) * (1,) for c in x.chunks) + ((x.ndim,),),
            new_axis=x.ndim,
        )

        c = []
        for i in range(x.ndim):
            s = x.ndim * [0] + [i]
            s[i] = slice(None)
            s = tuple(s)

            c.append(tuple(chunk_shapes[s]))

        # `map_blocks` assigns numpy dtypes
        # cast chunk dimensions back to python int before returning
        x._chunks = tuple(
            tuple(int(chunk) for chunk in chunks) for chunks in compute(tuple(c))[0]
        )
        return x

    @cached_property
    def shape(self):
        return tuple(cached_cumsum(c, initial_zero=True)[-1] for c in self.chunks)

    @property
    def chunksize(self):
        return tuple(max(c) for c in self.chunks)

    @property
    def dtype(self):
        if isinstance(self._meta, tuple):
            dtype = self._meta[0].dtype
        else:
            dtype = self._meta.dtype
        return dtype

    @property
    def _chunks(self):
        """Non-public chunks property. Allows setting a chunk value."""
        return self.__chunks

    @_chunks.setter
    def _chunks(self, chunks):
        self.__chunks = chunks

        # When the chunks changes the cached properties that was
        # dependent on it needs to be deleted:
        for key in ["numblocks", "npartitions", "shape", "ndim", "size", "_key_array"]:
            self._reset_cache(key)

    @property
    def chunks(self):
        """Chunks property."""
        return self.__chunks

    @chunks.setter
    def chunks(self, chunks):
        raise TypeError(
            "Can not set chunks directly\n\n"
            "Please use the rechunk method instead:\n"
            f"  x.rechunk({chunks})\n\n"
            "If trying to avoid unknown chunks, use\n"
            "  x.compute_chunk_sizes()"
        )

    def __len__(self):
        if not self.chunks:
            raise TypeError("len() of unsized object")
        return sum(self.chunks[0])

    def __array_ufunc__(self, numpy_ufunc, method, *inputs, **kwargs):
        out = kwargs.get("out", ())
        for x in inputs + out:
            if _should_delegate(x):
                return NotImplemented

        if method == "__call__":
            if numpy_ufunc is np.matmul:
                from .routines import matmul

                # special case until apply_gufunc handles optional dimensions
                return matmul(*inputs, **kwargs)
            if numpy_ufunc.signature is not None:
                from .gufunc import apply_gufunc

                return apply_gufunc(
                    numpy_ufunc, numpy_ufunc.signature, *inputs, **kwargs
                )
            if numpy_ufunc.nout > 1:
                from . import ufunc

                try:
                    da_ufunc = getattr(ufunc, numpy_ufunc.__name__)
                except AttributeError:
                    return NotImplemented
                return da_ufunc(*inputs, **kwargs)
            else:
                return elemwise(numpy_ufunc, *inputs, **kwargs)
        elif method == "outer":
            from . import ufunc

            try:
                da_ufunc = getattr(ufunc, numpy_ufunc.__name__)
            except AttributeError:
                return NotImplemented
            return da_ufunc.outer(*inputs, **kwargs)
        else:
            return NotImplemented

    def __repr__(self):
        """

        >>> import dask.array as da
        >>> da.ones((10, 10), chunks=(5, 5), dtype='i4')
        dask.array<..., shape=(10, 10), dtype=int32, chunksize=(5, 5), chunktype=numpy.ndarray>
        """
        chunksize = str(self.chunksize)
        name = self.name.rsplit("-", 1)[0]
        return (
            "dask.array<{}, shape={}, dtype={}, chunksize={}, chunktype={}.{}>".format(
                name,
                self.shape,
                self.dtype,
                chunksize,
                type(self._meta).__module__.split(".")[0],
                type(self._meta).__name__,
            )
        )

    def _repr_html_(self):
        try:
            grid = self.to_svg(size=config.get("array.svg.size", 120))
        except NotImplementedError:
            grid = ""

        if "sparse" in typename(type(self._meta)):
            nbytes = None
            cbytes = None
        elif not math.isnan(self.nbytes):
            nbytes = format_bytes(self.nbytes)
            cbytes = format_bytes(np.prod(self.chunksize) * self.dtype.itemsize)
        else:
            nbytes = "unknown"
            cbytes = "unknown"

        return get_template("array.html.j2").render(
            array=self,
            grid=grid,
            nbytes=nbytes,
            cbytes=cbytes,
        )

    @cached_property
    def ndim(self):
        return len(self.shape)

    @cached_property
    def size(self):
        """Number of elements in array"""
        return reduce(mul, self.shape, 1)

    @property
    def nbytes(self):
        """Number of bytes in array"""
        return self.size * self.dtype.itemsize

    @property
    def itemsize(self):
        """Length of one array element in bytes"""
        return self.dtype.itemsize

    @property
    def _name(self):
        return self.__name

    @_name.setter
    def _name(self, val):
        self.__name = val
        # Clear the key cache when the name is reset
        self._cached_keys = None
        self._reset_cache("_key_array")

    @property
    def name(self):
        return self.__name

    @name.setter
    def name(self, val):
        raise TypeError(
            "Cannot set name directly\n\n"
            "Name is used to relate the array to the task graph.\n"
            "It is uncommon to need to change it, but if you do\n"
            "please set ``._name``"
        )

    def __iter__(self):
        for i in range(len(self)):
            yield self[i]

    __array_priority__ = 11  # higher than numpy.ndarray and numpy.matrix

    def __array__(self, dtype=None, **kwargs):
        x = self.compute()
        if dtype and x.dtype != dtype:
            x = x.astype(dtype)
        if not isinstance(x, np.ndarray):
            x = np.array(x)
        return x

    def __array_function__(self, func, types, args, kwargs):
        import dask.array as module

        def handle_nonmatching_names(func, args, kwargs):
            if func not in _HANDLED_FUNCTIONS:
                warnings.warn(
                    "The `{}` function is not implemented by Dask array. "
                    "You may want to use the da.map_blocks function "
                    "or something similar to silence this warning. "
                    "Your code may stop working in a future release.".format(
                        func.__module__ + "." + func.__name__
                    ),
                    FutureWarning,
                )
                # Need to convert to array object (e.g. numpy.ndarray or
                # cupy.ndarray) as needed, so we can call the NumPy function
                # again and it gets the chance to dispatch to the right
                # implementation.
                args, kwargs = compute(args, kwargs)
                return func(*args, **kwargs)

            return _HANDLED_FUNCTIONS[func](*args, **kwargs)

        # First, verify that all types are handled by Dask. Otherwise, return NotImplemented.
        if not all(type is Array or is_valid_chunk_type(type) for type in types):
            return NotImplemented

        # Now try to find a matching function name.  If that doesn't work, we may
        # be dealing with an alias or a function that's simply not in the Dask API.
        # Handle aliases via the _HANDLED_FUNCTIONS dict mapping, and warn otherwise.
        for submodule in func.__module__.split(".")[1:]:
            try:
                module = getattr(module, submodule)
            except AttributeError:
                return handle_nonmatching_names(func, args, kwargs)

        if not hasattr(module, func.__name__):
            return handle_nonmatching_names(func, args, kwargs)

        da_func = getattr(module, func.__name__)
        if da_func is func:
            return handle_nonmatching_names(func, args, kwargs)

        # If ``like`` is contained in ``da_func``'s signature, add ``like=self``
        # to the kwargs dictionary.
        if has_keyword(da_func, "like"):
            kwargs["like"] = self

        return da_func(*args, **kwargs)

    @property
    def _elemwise(self):
        return elemwise

    @wraps(store)
    def store(self, target, **kwargs):
        r = store([self], [target], **kwargs)

        if kwargs.get("return_stored", False):
            r = r[0]

        return r

    def to_svg(self, size=500):
        """Convert chunks from Dask Array into an SVG Image

        Parameters
        ----------
        chunks: tuple
        size: int
            Rough size of the image

        Examples
        --------
        >>> x.to_svg(size=500)  # doctest: +SKIP

        Returns
        -------
        text: An svg string depicting the array as a grid of chunks
        """
        from .svg import svg

        return svg(self.chunks, size=size)

    def to_hdf5(self, filename, datapath, **kwargs):
        """Store array in HDF5 file

        >>> x.to_hdf5('myfile.hdf5', '/x')  # doctest: +SKIP

        Optionally provide arguments as though to ``h5py.File.create_dataset``

        >>> x.to_hdf5('myfile.hdf5', '/x', compression='lzf', shuffle=True)  # doctest: +SKIP

        See Also
        --------
        da.store
        h5py.File.create_dataset
        """
        return to_hdf5(filename, datapath, self, **kwargs)

    def to_dask_dataframe(self, columns=None, index=None, meta=None):
        """Convert dask Array to dask Dataframe

        Parameters
        ----------
        columns: list or string
            list of column names if DataFrame, single string if Series
        index : dask.dataframe.Index, optional
            An optional *dask* Index to use for the output Series or DataFrame.

            The default output index depends on whether the array has any unknown
            chunks. If there are any unknown chunks, the output has ``None``
            for all the divisions (one per chunk). If all the chunks are known,
            a default index with known divsions is created.

            Specifying ``index`` can be useful if you're conforming a Dask Array
            to an existing dask Series or DataFrame, and you would like the
            indices to match.
        meta : object, optional
            An optional `meta` parameter can be passed for dask
            to specify the concrete dataframe type to use for partitions of
            the Dask dataframe. By default, pandas DataFrame is used.

        See Also
        --------
        dask.dataframe.from_dask_array
        """
        from ..dataframe import from_dask_array

        return from_dask_array(self, columns=columns, index=index, meta=meta)

    def __bool__(self):
        if self.size > 1:
            raise ValueError(
                f"The truth value of a {self.__class__.__name__} is ambiguous. "
                "Use a.any() or a.all()."
            )
        else:
            return bool(self.compute())

    __nonzero__ = __bool__  # python 2

    def _scalarfunc(self, cast_type):
        if self.size > 1:
            raise TypeError("Only length-1 arrays can be converted to Python scalars")
        else:
            return cast_type(self.compute())

    def __int__(self):
        return self._scalarfunc(int)

    __long__ = __int__  # python 2

    def __float__(self):
        return self._scalarfunc(float)

    def __complex__(self):
        return self._scalarfunc(complex)

    def __index__(self):
        return self._scalarfunc(operator.index)

    def __setitem__(self, key, value):
        if value is np.ma.masked:
            value = np.ma.masked_all(())

        ## Use the "where" method for cases when key is an Array
        if isinstance(key, Array):
            from .routines import where

            if isinstance(value, Array) and value.ndim > 1:
                raise ValueError("boolean index array should have 1 dimension")
            try:
                y = where(key, value, self)
            except ValueError as e:
                raise ValueError(
                    "Boolean index assignment in Dask "
                    "expects equally shaped arrays.\nExample: da1[da2] = da3 "
                    "where da1.shape == (4,), da2.shape == (4,) "
                    "and da3.shape == (4,)."
                ) from e
            self._meta = y._meta
            self.dask = y.dask
            self._name = y.name
            self._chunks = y.chunks
            return

        if np.isnan(self.shape).any():
            raise ValueError(f"Arrays chunk sizes are unknown. {unknown_chunk_message}")

        # Still here? Then apply the assignment to other type of
        # indices via the `setitem_array` function.
        value = asanyarray(value)

        out = "setitem-" + tokenize(self, key, value)
        dsk = setitem_array(out, self, key, value)

        meta = meta_from_array(self._meta)
        if np.isscalar(meta):
            meta = np.array(meta)

        graph = HighLevelGraph.from_collections(out, dsk, dependencies=[self])
        y = Array(graph, out, chunks=self.chunks, dtype=self.dtype, meta=meta)

        self._meta = y._meta
        self.dask = y.dask
        self._name = y.name
        self._chunks = y.chunks

    def __getitem__(self, index):
        # Field access, e.g. x['a'] or x[['a', 'b']]
        if isinstance(index, str) or (
            isinstance(index, list) and index and all(isinstance(i, str) for i in index)
        ):
            if isinstance(index, str):
                dt = self.dtype[index]
            else:
                dt = np.dtype(
                    {
                        "names": index,
                        "formats": [self.dtype.fields[name][0] for name in index],
                        "offsets": [self.dtype.fields[name][1] for name in index],
                        "itemsize": self.dtype.itemsize,
                    }
                )

            if dt.shape:
                new_axis = list(range(self.ndim, self.ndim + len(dt.shape)))
                chunks = self.chunks + tuple((i,) for i in dt.shape)
                return self.map_blocks(
                    getitem, index, dtype=dt.base, chunks=chunks, new_axis=new_axis
                )
            else:
                return self.map_blocks(getitem, index, dtype=dt)

        if not isinstance(index, tuple):
            index = (index,)

        from .slicing import (
            normalize_index,
            slice_with_bool_dask_array,
            slice_with_int_dask_array,
        )

        index2 = normalize_index(index, self.shape)
        dependencies = {self.name}
        for i in index2:
            if isinstance(i, Array):
                dependencies.add(i.name)

        if any(isinstance(i, Array) and i.dtype.kind in "iu" for i in index2):
            self, index2 = slice_with_int_dask_array(self, index2)
        if any(isinstance(i, Array) and i.dtype == bool for i in index2):
            self, index2 = slice_with_bool_dask_array(self, index2)

        if all(isinstance(i, slice) and i == slice(None) for i in index2):
            return self

        out = "getitem-" + tokenize(self, index2)
        dsk, chunks = slice_array(out, self.name, self.chunks, index2, self.itemsize)

        graph = HighLevelGraph.from_collections(out, dsk, dependencies=[self])

        meta = meta_from_array(self._meta, ndim=len(chunks))
        if np.isscalar(meta):
            meta = np.array(meta)

        return Array(graph, out, chunks, meta=meta)

    def _vindex(self, key):
        if not isinstance(key, tuple):
            key = (key,)
        if any(k is None for k in key):
            raise IndexError(
                "vindex does not support indexing with None (np.newaxis), "
                "got {}".format(key)
            )
        if all(isinstance(k, slice) for k in key):
            if all(
                k.indices(d) == slice(0, d).indices(d) for k, d in zip(key, self.shape)
            ):
                return self
            raise IndexError(
                "vindex requires at least one non-slice to vectorize over "
                "when the slices are not over the entire array (i.e, x[:]). "
                "Use normal slicing instead when only using slices. Got: {}".format(key)
            )
        return _vindex(self, *key)

    @property
    def vindex(self):
        """Vectorized indexing with broadcasting.

        This is equivalent to numpy's advanced indexing, using arrays that are
        broadcast against each other. This allows for pointwise indexing:

        >>> import dask.array as da
        >>> x = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
        >>> x = da.from_array(x, chunks=2)
        >>> x.vindex[[0, 1, 2], [0, 1, 2]].compute()
        array([1, 5, 9])

        Mixed basic/advanced indexing with slices/arrays is also supported. The
        order of dimensions in the result follows those proposed for
        `ndarray.vindex <https://github.com/numpy/numpy/pull/6256>`_:
        the subspace spanned by arrays is followed by all slices.

        Note: ``vindex`` provides more general functionality than standard
        indexing, but it also has fewer optimizations and can be significantly
        slower.
        """
        return IndexCallable(self._vindex)

    @property
    def blocks(self):
        """An array-like interface to the blocks of an array.

        This returns a ``Blockview`` object that provides an array-like interface
        to the blocks of a dask array.  Numpy-style indexing of a ``Blockview`` object
        returns a selection of blocks as a new dask array.

        You can index ``array.blocks`` like a numpy array of shape
        equal to the number of blocks in each dimension, (available as
        array.blocks.size).  The dimensionality of the output array matches
        the dimension of this array, even if integer indices are passed.
        Slicing with ``np.newaxis`` or multiple lists is not supported.

        Examples
        --------
        >>> import dask.array as da
        >>> x = da.arange(8, chunks=2)
        >>> x.blocks.shape # aliases x.numblocks
        (4,)
        >>> x.blocks[0].compute()
        array([0, 1])
        >>> x.blocks[:3].compute()
        array([0, 1, 2, 3, 4, 5])
        >>> x.blocks[::2].compute()
        array([0, 1, 4, 5])
        >>> x.blocks[[-1, 0]].compute()
        array([6, 7, 0, 1])
        >>> x.blocks.ravel() # doctest: +NORMALIZE_WHITESPACE
        [dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>]

        Returns
        -------
        An instance of ``dask.array.Blockview``
        """
        return BlockView(self)

    @property
    def partitions(self):
        """Slice an array by partitions. Alias of dask array .blocks attribute.

        This alias allows you to write agnostic code that works with both
        dask arrays and dask dataframes.

        This returns a ``Blockview`` object that provides an array-like interface
        to the blocks of a dask array.  Numpy-style indexing of a ``Blockview`` object
        returns a selection of blocks as a new dask array.

        You can index ``array.blocks`` like a numpy array of shape
        equal to the number of blocks in each dimension, (available as
        array.blocks.size).  The dimensionality of the output array matches
        the dimension of this array, even if integer indices are passed.
        Slicing with ``np.newaxis`` or multiple lists is not supported.

        Examples
        --------
        >>> import dask.array as da
        >>> x = da.arange(8, chunks=2)
        >>> x.partitions.shape # aliases x.numblocks
        (4,)
        >>> x.partitions[0].compute()
        array([0, 1])
        >>> x.partitions[:3].compute()
        array([0, 1, 2, 3, 4, 5])
        >>> x.partitions[::2].compute()
        array([0, 1, 4, 5])
        >>> x.partitions[[-1, 0]].compute()
        array([6, 7, 0, 1])
        >>> x.partitions.ravel() # doctest: +NORMALIZE_WHITESPACE
        [dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
         dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>]

        Returns
        -------
        An instance of ``da.array.Blockview``
        """
        return self.blocks

    @derived_from(np.ndarray)
    def dot(self, other):
        from .routines import tensordot

        return tensordot(self, other, axes=((self.ndim - 1,), (other.ndim - 2,)))

    @property
    def A(self):
        return self

    @property
    def T(self):
        return self.transpose()

    @derived_from(np.ndarray)
    def transpose(self, *axes):
        from .routines import transpose

        if not axes:
            axes = None
        elif len(axes) == 1 and isinstance(axes[0], Iterable):
            axes = axes[0]
        if (axes == tuple(range(self.ndim))) or (axes == tuple(range(-self.ndim, 0))):
            # no transpose necessary
            return self
        else:
            return transpose(self, axes=axes)

    @derived_from(np.ndarray)
    def ravel(self):
        from .routines import ravel

        return ravel(self)

    flatten = ravel

    @derived_from(np.ndarray)
    def choose(self, choices):
        from .routines import choose

        return choose(self, choices)

    @derived_from(np.ndarray)
    def reshape(self, *shape, merge_chunks=True):
        """
        .. note::

           See :meth:`dask.array.reshape` for an explanation of
           the ``merge_chunks`` keyword.
        """
        from .reshape import reshape

        if len(shape) == 1 and not isinstance(shape[0], Number):
            shape = shape[0]
        return reshape(self, shape, merge_chunks=merge_chunks)

    def topk(self, k, axis=-1, split_every=None):
        """The top k elements of an array.

        See :func:`dask.array.topk` for docstring.

        """
        from .reductions import topk

        return topk(self, k, axis=axis, split_every=split_every)

    def argtopk(self, k, axis=-1, split_every=None):
        """The indices of the top k elements of an array.

        See :func:`dask.array.argtopk` for docstring.

        """
        from .reductions import argtopk

        return argtopk(self, k, axis=axis, split_every=split_every)

    def astype(self, dtype, **kwargs):
        """Copy of the array, cast to a specified type.

        Parameters
        ----------
        dtype : str or dtype
            Typecode or data-type to which the array is cast.
        casting : {'no', 'equiv', 'safe', 'same_kind', 'unsafe'}, optional
            Controls what kind of data casting may occur. Defaults to 'unsafe'
            for backwards compatibility.

            * 'no' means the data types should not be cast at all.
            * 'equiv' means only byte-order changes are allowed.
            * 'safe' means only casts which can preserve values are allowed.
            * 'same_kind' means only safe casts or casts within a kind,
                like float64 to float32, are allowed.
            * 'unsafe' means any data conversions may be done.
        copy : bool, optional
            By default, astype always returns a newly allocated array. If this
            is set to False and the `dtype` requirement is satisfied, the input
            array is returned instead of a copy.
        """
        # Scalars don't take `casting` or `copy` kwargs - as such we only pass
        # them to `map_blocks` if specified by user (different than defaults).
        extra = set(kwargs) - {"casting", "copy"}
        if extra:
            raise TypeError(
                f"astype does not take the following keyword arguments: {list(extra)}"
            )
        casting = kwargs.get("casting", "unsafe")
        dtype = np.dtype(dtype)
        if self.dtype == dtype:
            return self
        elif not np.can_cast(self.dtype, dtype, casting=casting):
            raise TypeError(
                f"Cannot cast array from {self.dtype!r} to {dtype!r} "
                f"according to the rule {casting!r}"
            )
        return self.map_blocks(chunk.astype, dtype=dtype, astype_dtype=dtype, **kwargs)

    def __abs__(self):
        return elemwise(operator.abs, self)

    @check_if_handled_given_other
    def __add__(self, other):
        return elemwise(operator.add, self, other)

    @check_if_handled_given_other
    def __radd__(self, other):
        return elemwise(operator.add, other, self)

    @check_if_handled_given_other
    def __and__(self, other):
        return elemwise(operator.and_, self, other)

    @check_if_handled_given_other
    def __rand__(self, other):
        return elemwise(operator.and_, other, self)

    @check_if_handled_given_other
    def __div__(self, other):
        return elemwise(operator.div, self, other)

    @check_if_handled_given_other
    def __rdiv__(self, other):
        return elemwise(operator.div, other, self)

    @check_if_handled_given_other
    def __eq__(self, other):
        return elemwise(operator.eq, self, other)

    @check_if_handled_given_other
    def __gt__(self, other):
        return elemwise(operator.gt, self, other)

    @check_if_handled_given_other
    def __ge__(self, other):
        return elemwise(operator.ge, self, other)

    def __invert__(self):
        return elemwise(operator.invert, self)

    @check_if_handled_given_other
    def __lshift__(self, other):
        return elemwise(operator.lshift, self, other)

    @check_if_handled_given_other
    def __rlshift__(self, other):
        return elemwise(operator.lshift, other, self)

    @check_if_handled_given_other
    def __lt__(self, other):
        return elemwise(operator.lt, self, other)

    @check_if_handled_given_other
    def __le__(self, other):
        return elemwise(operator.le, self, other)

    @check_if_handled_given_other
    def __mod__(self, other):
        return elemwise(operator.mod, self, other)

    @check_if_handled_given_other
    def __rmod__(self, other):
        return elemwise(operator.mod, other, self)

    @check_if_handled_given_other
    def __mul__(self, other):
        return elemwise(operator.mul, self, other)

    @check_if_handled_given_other
    def __rmul__(self, other):
        return elemwise(operator.mul, other, self)

    @check_if_handled_given_other
    def __ne__(self, other):
        return elemwise(operator.ne, self, other)

    def __neg__(self):
        return elemwise(operator.neg, self)

    @check_if_handled_given_other
    def __or__(self, other):
        return elemwise(operator.or_, self, other)

    def __pos__(self):
        return self

    @check_if_handled_given_other
    def __ror__(self, other):
        return elemwise(operator.or_, other, self)

    @check_if_handled_given_other
    def __pow__(self, other):
        return elemwise(operator.pow, self, other)

    @check_if_handled_given_other
    def __rpow__(self, other):
        return elemwise(operator.pow, other, self)

    @check_if_handled_given_other
    def __rshift__(self, other):
        return elemwise(operator.rshift, self, other)

    @check_if_handled_given_other
    def __rrshift__(self, other):
        return elemwise(operator.rshift, other, self)

    @check_if_handled_given_other
    def __sub__(self, other):
        return elemwise(operator.sub, self, other)

    @check_if_handled_given_other
    def __rsub__(self, other):
        return elemwise(operator.sub, other, self)

    @check_if_handled_given_other
    def __truediv__(self, other):
        return elemwise(operator.truediv, self, other)

    @check_if_handled_given_other
    def __rtruediv__(self, other):
        return elemwise(operator.truediv, other, self)

    @check_if_handled_given_other
    def __floordiv__(self, other):
        return elemwise(operator.floordiv, self, other)

    @check_if_handled_given_other
    def __rfloordiv__(self, other):
        return elemwise(operator.floordiv, other, self)

    @check_if_handled_given_other
    def __xor__(self, other):
        return elemwise(operator.xor, self, other)

    @check_if_handled_given_other
    def __rxor__(self, other):
        return elemwise(operator.xor, other, self)

    @check_if_handled_given_other
    def __matmul__(self, other):
        from .routines import matmul

        return matmul(self, other)

    @check_if_handled_given_other
    def __rmatmul__(self, other):
        from .routines import matmul

        return matmul(other, self)

    @check_if_handled_given_other
    def __divmod__(self, other):
        from .ufunc import divmod

        return divmod(self, other)

    @check_if_handled_given_other
    def __rdivmod__(self, other):
        from .ufunc import divmod

        return divmod(other, self)

    @derived_from(np.ndarray)
    def any(self, axis=None, keepdims=False, split_every=None, out=None):
        from .reductions import any

        return any(self, axis=axis, keepdims=keepdims, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def all(self, axis=None, keepdims=False, split_every=None, out=None):
        from .reductions import all

        return all(self, axis=axis, keepdims=keepdims, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def min(self, axis=None, keepdims=False, split_every=None, out=None):
        from .reductions import min

        return min(self, axis=axis, keepdims=keepdims, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def max(self, axis=None, keepdims=False, split_every=None, out=None):
        from .reductions import max

        return max(self, axis=axis, keepdims=keepdims, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def argmin(self, axis=None, split_every=None, out=None):
        from .reductions import argmin

        return argmin(self, axis=axis, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def argmax(self, axis=None, split_every=None, out=None):
        from .reductions import argmax

        return argmax(self, axis=axis, split_every=split_every, out=out)

    @derived_from(np.ndarray)
    def sum(self, axis=None, dtype=None, keepdims=False, split_every=None, out=None):
        from .reductions import sum

        return sum(
            self,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            split_every=split_every,
            out=out,
        )

    @derived_from(np.ndarray)
    def trace(self, offset=0, axis1=0, axis2=1, dtype=None):
        from .reductions import trace

        return trace(self, offset=offset, axis1=axis1, axis2=axis2, dtype=dtype)

    @derived_from(np.ndarray)
    def prod(self, axis=None, dtype=None, keepdims=False, split_every=None, out=None):
        from .reductions import prod

        return prod(
            self,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            split_every=split_every,
            out=out,
        )

    @derived_from(np.ndarray)
    def mean(self, axis=None, dtype=None, keepdims=False, split_every=None, out=None):
        from .reductions import mean

        return mean(
            self,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            split_every=split_every,
            out=out,
        )

    @derived_from(np.ndarray)
    def std(
        self, axis=None, dtype=None, keepdims=False, ddof=0, split_every=None, out=None
    ):
        from .reductions import std

        return std(
            self,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            ddof=ddof,
            split_every=split_every,
            out=out,
        )

    @derived_from(np.ndarray)
    def var(
        self, axis=None, dtype=None, keepdims=False, ddof=0, split_every=None, out=None
    ):
        from .reductions import var

        return var(
            self,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            ddof=ddof,
            split_every=split_every,
            out=out,
        )

    def moment(
        self,
        order,
        axis=None,
        dtype=None,
        keepdims=False,
        ddof=0,
        split_every=None,
        out=None,
    ):
        """Calculate the nth centralized moment.

        Parameters
        ----------
        order : int
            Order of the moment that is returned, must be >= 2.
        axis : int, optional
            Axis along which the central moment is computed. The default is to
            compute the moment of the flattened array.
        dtype : data-type, optional
            Type to use in computing the moment. For arrays of integer type the
            default is float64; for arrays of float types it is the same as the
            array type.
        keepdims : bool, optional
            If this is set to True, the axes which are reduced are left in the
            result as dimensions with size one. With this option, the result
            will broadcast correctly against the original array.
        ddof : int, optional
            "Delta Degrees of Freedom": the divisor used in the calculation is
            N - ddof, where N represents the number of elements. By default
            ddof is zero.

        Returns
        -------
        moment : ndarray

        References
        ----------
        .. [1] Pebay, Philippe (2008), "Formulas for Robust, One-Pass Parallel
           Computation of Covariances and Arbitrary-Order Statistical Moments",
           Technical Report SAND2008-6212, Sandia National Laboratories.

        """

        from .reductions import moment

        return moment(
            self,
            order,
            axis=axis,
            dtype=dtype,
            keepdims=keepdims,
            ddof=ddof,
            split_every=split_every,
            out=out,
        )

    @wraps(map_blocks)
    def map_blocks(self, func, *args, **kwargs):
        return map_blocks(func, self, *args, **kwargs)

    def map_overlap(self, func, depth, boundary=None, trim=True, **kwargs):
        """Map a function over blocks of the array with some overlap

        We share neighboring zones between blocks of the array, then map a
        function, then trim away the neighboring strips.

        Note that this function will attempt to automatically determine the output
        array type before computing it, please refer to the ``meta`` keyword argument
        in :func:`map_blocks <dask.array.core.map_blocks>` if you expect that the function will not succeed when
        operating on 0-d arrays.

        Parameters
        ----------
        func: function
            The function to apply to each extended block
        depth: int, tuple, or dict
            The number of elements that each block should share with its neighbors
            If a tuple or dict then this can be different per axis
        boundary: str, tuple, dict
            How to handle the boundaries.
            Values include 'reflect', 'periodic', 'nearest', 'none',
            or any constant value like 0 or np.nan
        trim: bool
            Whether or not to trim ``depth`` elements from each block after
            calling the map function.
            Set this to False if your mapping function already does this for you
        **kwargs:
            Other keyword arguments valid in :func:`map_blocks <dask.array.core.map_blocks>`.

        Examples
        --------
        >>> import dask.array as da
        >>> x = np.array([1, 1, 2, 3, 3, 3, 2, 1, 1])
        >>> x = da.from_array(x, chunks=5)
        >>> def derivative(x):
        ...     return x - np.roll(x, 1)

        >>> y = x.map_overlap(derivative, depth=1, boundary=0)
        >>> y.compute()
        array([ 1,  0,  1,  1,  0,  0, -1, -1,  0])

        >>> import dask.array as da
        >>> x = np.arange(16).reshape((4, 4))
        >>> d = da.from_array(x, chunks=(2, 2))
        >>> d.map_overlap(lambda x: x + x.size, depth=1).compute()
        array([[16, 17, 18, 19],
               [20, 21, 22, 23],
               [24, 25, 26, 27],
               [28, 29, 30, 31]])

        >>> func = lambda x: x + x.size
        >>> depth = {0: 1, 1: 1}
        >>> boundary = {0: 'reflect', 1: 'none'}
        >>> d.map_overlap(func, depth, boundary).compute()  # doctest: +NORMALIZE_WHITESPACE
        array([[12,  13,  14,  15],
               [16,  17,  18,  19],
               [20,  21,  22,  23],
               [24,  25,  26,  27]])

        >>> x = np.arange(16).reshape((4, 4))
        >>> d = da.from_array(x, chunks=(2, 2))
        >>> y = d.map_overlap(lambda x: x + x[2], depth=1, meta=np.array(()))
        >>> y
        dask.array<_trim, shape=(4, 4), dtype=float64, chunksize=(2, 2), chunktype=numpy.ndarray>
        >>> y.compute()
        array([[ 4,  6,  8, 10],
               [ 8, 10, 12, 14],
               [20, 22, 24, 26],
               [24, 26, 28, 30]])

        >>> import cupy  # doctest: +SKIP
        >>> x = cupy.arange(16).reshape((5, 4))  # doctest: +SKIP
        >>> d = da.from_array(x, chunks=(2, 2))  # doctest: +SKIP
        >>> y = d.map_overlap(lambda x: x + x[2], depth=1, meta=cupy.array(()))  # doctest: +SKIP
        >>> y  # doctest: +SKIP
        dask.array<_trim, shape=(4, 4), dtype=float64, chunksize=(2, 2), chunktype=cupy.ndarray>
        >>> y.compute()  # doctest: +SKIP
        array([[ 4,  6,  8, 10],
               [ 8, 10, 12, 14],
               [20, 22, 24, 26],
               [24, 26, 28, 30]])
        """
        from .overlap import map_overlap

        return map_overlap(
            func, self, depth=depth, boundary=boundary, trim=trim, **kwargs
        )

    @derived_from(np.ndarray)
    def cumsum(self, axis, dtype=None, out=None, *, method="sequential"):
        """Dask added an additional keyword-only argument ``method``.

        method : {'sequential', 'blelloch'}, optional
            Choose which method to use to perform the cumsum.  Default is 'sequential'.

            * 'sequential' performs the cumsum of each prior block before the current block.
            * 'blelloch' is a work-efficient parallel cumsum.  It exposes parallelism by
              first taking the sum of each block and combines the sums via a binary tree.
              This method may be faster or more memory efficient depending on workload,
              scheduler, and hardware.  More benchmarking is necessary.
        """
        from .reductions import cumsum

        return cumsum(self, axis, dtype, out=out, method=method)

    @derived_from(np.ndarray)
    def cumprod(self, axis, dtype=None, out=None, *, method="sequential"):
        """Dask added an additional keyword-only argument ``method``.

        method : {'sequential', 'blelloch'}, optional
            Choose which method to use to perform the cumprod.  Default is 'sequential'.

            * 'sequential' performs the cumprod of each prior block before the current block.
            * 'blelloch' is a work-efficient parallel cumprod.  It exposes parallelism by first
              taking the product of each block and combines the products via a binary tree.
              This method may be faster or more memory efficient depending on workload,
              scheduler, and hardware.  More benchmarking is necessary.
        """
        from .reductions import cumprod

        return cumprod(self, axis, dtype, out=out, method=method)

    @derived_from(np.ndarray)
    def squeeze(self, axis=None):
        from .routines import squeeze

        return squeeze(self, axis)

    def rechunk(
        self, chunks="auto", threshold=None, block_size_limit=None, balance=False
    ):
        """See da.rechunk for docstring"""
        from .rechunk import rechunk  # avoid circular import

        return rechunk(self, chunks, threshold, block_size_limit, balance)

    @property
    def real(self):
        from .ufunc import real

        return real(self)

    @property
    def imag(self):
        from .ufunc import imag

        return imag(self)

    def conj(self):
        from .ufunc import conj

        return conj(self)

    @derived_from(np.ndarray)
    def clip(self, min=None, max=None):
        from .ufunc import clip

        return clip(self, min, max)

    def view(self, dtype=None, order="C"):
        """Get a view of the array as a new data type

        Parameters
        ----------
        dtype:
            The dtype by which to view the array.
            The default, None, results in the view having the same data-type
            as the original array.
        order: string
            'C' or 'F' (Fortran) ordering

        This reinterprets the bytes of the array under a new dtype.  If that
        dtype does not have the same size as the original array then the shape
        will change.

        Beware that both numpy and dask.array can behave oddly when taking
        shape-changing views of arrays under Fortran ordering.  Under some
        versions of NumPy this function will fail when taking shape-changing
        views of Fortran ordered arrays if the first dimension has chunks of
        size one.
        """
        if dtype is None:
            dtype = self.dtype
        else:
            dtype = np.dtype(dtype)
        mult = self.dtype.itemsize / dtype.itemsize

        if order == "C":
            chunks = self.chunks[:-1] + (
                tuple(ensure_int(c * mult) for c in self.chunks[-1]),
            )
        elif order == "F":
            chunks = (
                tuple(ensure_int(c * mult) for c in self.chunks[0]),
            ) + self.chunks[1:]
        else:
            raise ValueError("Order must be one of 'C' or 'F'")

        return self.map_blocks(
            chunk.view, dtype, order=order, dtype=dtype, chunks=chunks
        )

    @derived_from(np.ndarray)
    def swapaxes(self, axis1, axis2):
        from .routines import swapaxes

        return swapaxes(self, axis1, axis2)

    @derived_from(np.ndarray)
    def round(self, decimals=0):
        from .routines import round

        return round(self, decimals=decimals)

    def copy(self):
        """
        Copy array.  This is a no-op for dask.arrays, which are immutable
        """
        if self.npartitions == 1:
            return self.map_blocks(M.copy)
        else:
            return Array(self.dask, self.name, self.chunks, meta=self)

    def __deepcopy__(self, memo):
        c = self.copy()
        memo[id(self)] = c
        return c

    def to_delayed(self, optimize_graph=True):
        """Convert into an array of :class:`dask.delayed.Delayed` objects, one per chunk.

        Parameters
        ----------
        optimize_graph : bool, optional
            If True [default], the graph is optimized before converting into
            :class:`dask.delayed.Delayed` objects.

        See Also
        --------
        dask.array.from_delayed
        """
        keys = self.__dask_keys__()
        graph = self.__dask_graph__()
        if optimize_graph:
            graph = self.__dask_optimize__(graph, keys)  # TODO, don't collape graph
            name = "delayed-" + self.name
            graph = HighLevelGraph.from_collections(name, graph, dependencies=())
        L = ndeepmap(self.ndim, lambda k: Delayed(k, graph), keys)
        return np.array(L, dtype=object)

    @derived_from(np.ndarray)
    def repeat(self, repeats, axis=None):
        from .creation import repeat

        return repeat(self, repeats, axis=axis)

    @derived_from(np.ndarray)
    def nonzero(self):
        from .routines import nonzero

        return nonzero(self)

    def to_zarr(self, *args, **kwargs):
        """Save array to the zarr storage format

        See https://zarr.readthedocs.io for details about the format.

        See function :func:`dask.array.to_zarr` for parameters.
        """
        return to_zarr(self, *args, **kwargs)

    def to_tiledb(self, uri, *args, **kwargs):
        """Save array to the TileDB storage manager

        See function :func:`dask.array.to_tiledb` for argument documentation.

        See https://docs.tiledb.io for details about the format and engine.
        """
        from .tiledb_io import to_tiledb

        return to_tiledb(self, uri, *args, **kwargs)


def ensure_int(f):
    i = int(f)
    if i != f:
        raise ValueError("Could not coerce %f to integer" % f)
    return i


def normalize_chunks(chunks, shape=None, limit=None, dtype=None, previous_chunks=None):
    """Normalize chunks to tuple of tuples

    This takes in a variety of input types and information and produces a full
    tuple-of-tuples result for chunks, suitable to be passed to Array or
    rechunk or any other operation that creates a Dask array.

    Parameters
    ----------
    chunks: tuple, int, dict, or string
        The chunks to be normalized.  See examples below for more details
    shape: Tuple[int]
        The shape of the array
    limit: int (optional)
        The maximum block size to target in bytes,
        if freedom is given to choose
    dtype: np.dtype
    previous_chunks: Tuple[Tuple[int]] optional
        Chunks from a previous array that we should use for inspiration when
        rechunking auto dimensions.  If not provided but auto-chunking exists
        then auto-dimensions will prefer square-like chunk shapes.

    Examples
    --------
    Specify uniform chunk sizes

    >>> from dask.array.core import normalize_chunks
    >>> normalize_chunks((2, 2), shape=(5, 6))
    ((2, 2, 1), (2, 2, 2))

    Also passes through fully explicit tuple-of-tuples

    >>> normalize_chunks(((2, 2, 1), (2, 2, 2)), shape=(5, 6))
    ((2, 2, 1), (2, 2, 2))

    Cleans up lists to tuples

    >>> normalize_chunks([[2, 2], [3, 3]])
    ((2, 2), (3, 3))

    Expands integer inputs 10 -> (10, 10)

    >>> normalize_chunks(10, shape=(30, 5))
    ((10, 10, 10), (5,))

    Expands dict inputs

    >>> normalize_chunks({0: 2, 1: 3}, shape=(6, 6))
    ((2, 2, 2), (3, 3))

    The values -1 and None get mapped to full size

    >>> normalize_chunks((5, -1), shape=(10, 10))
    ((5, 5), (10,))

    Use the value "auto" to automatically determine chunk sizes along certain
    dimensions.  This uses the ``limit=`` and ``dtype=`` keywords to
    determine how large to make the chunks.  The term "auto" can be used
    anywhere an integer can be used.  See array chunking documentation for more
    information.

    >>> normalize_chunks(("auto",), shape=(20,), limit=5, dtype='uint8')
    ((5, 5, 5, 5),)

    You can also use byte sizes (see :func:`dask.utils.parse_bytes`) in place of
    "auto" to ask for a particular size

    >>> normalize_chunks("1kiB", shape=(2000,), dtype='float32')
    ((250, 250, 250, 250, 250, 250, 250, 250),)

    Respects null dimensions

    >>> normalize_chunks((), shape=(0, 0))
    ((0,), (0,))
    """
    if dtype and not isinstance(dtype, np.dtype):
        dtype = np.dtype(dtype)
    if chunks is None:
        raise ValueError(CHUNKS_NONE_ERROR_MESSAGE)
    if isinstance(chunks, list):
        chunks = tuple(chunks)
    if isinstance(chunks, (Number, str)):
        chunks = (chunks,) * len(shape)
    if isinstance(chunks, dict):
        chunks = tuple(chunks.get(i, None) for i in range(len(shape)))
    if isinstance(chunks, np.ndarray):
        chunks = chunks.tolist()
    if not chunks and shape and all(s == 0 for s in shape):
        chunks = ((0,),) * len(shape)

    if (
        shape
        and len(shape) == 1
        and len(chunks) > 1
        and all(isinstance(c, (Number, str)) for c in chunks)
    ):
        chunks = (chunks,)

    if shape and len(chunks) != len(shape):
        raise ValueError(
            "Chunks and shape must be of the same length/dimension. "
            "Got chunks=%s, shape=%s" % (chunks, shape)
        )
    if -1 in chunks or None in chunks:
        chunks = tuple(s if c == -1 or c is None else c for c, s in zip(chunks, shape))

    # If specifying chunk size in bytes, use that value to set the limit.
    # Verify there is only one consistent value of limit or chunk-bytes used.
    for c in chunks:
        if isinstance(c, str) and c != "auto":
            parsed = parse_bytes(c)
            if limit is None:
                limit = parsed
            elif parsed != limit:
                raise ValueError(
                    "Only one consistent value of limit or chunk is allowed."
                    "Used %s != %s" % (parsed, limit)
                )
    # Substitute byte limits with 'auto' now that limit is set.
    chunks = tuple("auto" if isinstance(c, str) and c != "auto" else c for c in chunks)

    if any(c == "auto" for c in chunks):
        chunks = auto_chunks(chunks, shape, limit, dtype, previous_chunks)

    if shape is not None:
        chunks = tuple(c if c not in {None, -1} else s for c, s in zip(chunks, shape))

    if chunks and shape is not None:
        chunks = sum(
            (
                blockdims_from_blockshape((s,), (c,))
                if not isinstance(c, (tuple, list))
                else (c,)
                for s, c in zip(shape, chunks)
            ),
            (),
        )
    for c in chunks:
        if not c:
            raise ValueError(
                "Empty tuples are not allowed in chunks. Express "
                "zero length dimensions with 0(s) in chunks"
            )

    if shape is not None:
        if len(chunks) != len(shape):
            raise ValueError(
                "Input array has %d dimensions but the supplied "
                "chunks has only %d dimensions" % (len(shape), len(chunks))
            )
        if not all(
            c == s or (math.isnan(c) or math.isnan(s))
            for c, s in zip(map(sum, chunks), shape)
        ):
            raise ValueError(
                "Chunks do not add up to shape. "
                "Got chunks=%s, shape=%s" % (chunks, shape)
            )

    return tuple(tuple(int(x) if not math.isnan(x) else x for x in c) for c in chunks)


def _compute_multiplier(limit: int, dtype, largest_block: int, result):
    """
    Utility function for auto_chunk, to fin how much larger or smaller the ideal
    chunk size is relative to what we have now.
    """
    return (
        limit
        / dtype.itemsize
        / largest_block
        / np.prod(list(r if r != 0 else 1 for r in result.values()))
    )


def auto_chunks(chunks, shape, limit, dtype, previous_chunks=None):
    """Determine automatic chunks

    This takes in a chunks value that contains ``"auto"`` values in certain
    dimensions and replaces those values with concrete dimension sizes that try
    to get chunks to be of a certain size in bytes, provided by the ``limit=``
    keyword.  If multiple dimensions are marked as ``"auto"`` then they will
    all respond to meet the desired byte limit, trying to respect the aspect
    ratio of their dimensions in ``previous_chunks=``, if given.

    Parameters
    ----------
    chunks: Tuple
        A tuple of either dimensions or tuples of explicit chunk dimensions
        Some entries should be "auto"
    shape: Tuple[int]
    limit: int, str
        The maximum allowable size of a chunk in bytes
    previous_chunks: Tuple[Tuple[int]]

    See also
    --------
    normalize_chunks: for full docstring and parameters
    """
    if previous_chunks is not None:
        previous_chunks = tuple(
            c if isinstance(c, tuple) else (c,) for c in previous_chunks
        )
    chunks = list(chunks)

    autos = {i for i, c in enumerate(chunks) if c == "auto"}
    if not autos:
        return tuple(chunks)

    if limit is None:
        limit = config.get("array.chunk-size")
    if isinstance(limit, str):
        limit = parse_bytes(limit)

    if dtype is None:
        raise TypeError("dtype must be known for auto-chunking")

    if dtype.hasobject:
        raise NotImplementedError(
            "Can not use auto rechunking with object dtype. "
            "We are unable to estimate the size in bytes of object data"
        )

    for x in tuple(chunks) + tuple(shape):
        if (
            isinstance(x, Number)
            and np.isnan(x)
            or isinstance(x, tuple)
            and np.isnan(x).any()
        ):
            raise ValueError(
                "Can not perform automatic rechunking with unknown "
                "(nan) chunk sizes.%s" % unknown_chunk_message
            )

    limit = max(1, limit)

    largest_block = np.prod(
        [cs if isinstance(cs, Number) else max(cs) for cs in chunks if cs != "auto"]
    )

    if previous_chunks:
        # Base ideal ratio on the median chunk size of the previous chunks
        result = {a: np.median(previous_chunks[a]) for a in autos}

        ideal_shape = []
        for i, s in enumerate(shape):
            chunk_frequencies = frequencies(previous_chunks[i])
            mode, count = max(chunk_frequencies.items(), key=lambda kv: kv[1])
            if mode > 1 and count >= len(previous_chunks[i]) / 2:
                ideal_shape.append(mode)
            else:
                ideal_shape.append(s)

        # How much larger or smaller the ideal chunk size is relative to what we have now
        multiplier = _compute_multiplier(limit, dtype, largest_block, result)

        last_multiplier = 0
        last_autos = set()
        while (
            multiplier != last_multiplier or autos != last_autos
        ):  # while things change
            last_multiplier = multiplier  # record previous values
            last_autos = set(autos)  # record previous values

            # Expand or contract each of the dimensions appropriately
            for a in sorted(autos):
                if ideal_shape[a] == 0:
                    result[a] = 0
                    continue
                proposed = result[a] * multiplier ** (1 / len(autos))
                if proposed > shape[a]:  # we've hit the shape boundary
                    autos.remove(a)
                    largest_block *= shape[a]
                    chunks[a] = shape[a]
                    del result[a]
                else:
                    result[a] = round_to(proposed, ideal_shape[a])

            # recompute how much multiplier we have left, repeat
            multiplier = _compute_multiplier(limit, dtype, largest_block, result)

        for k, v in result.items():
            chunks[k] = v
        return tuple(chunks)

    else:
        size = (limit / dtype.itemsize / largest_block) ** (1 / len(autos))
        small = [i for i in autos if shape[i] < size]
        if small:
            for i in small:
                chunks[i] = (shape[i],)
            return auto_chunks(chunks, shape, limit, dtype)

        for i in autos:
            chunks[i] = round_to(size, shape[i])

        return tuple(chunks)


def round_to(c, s):
    """Return a chunk dimension that is close to an even multiple or factor

    We want values for c that are nicely aligned with s.

    If c is smaller than s then we want the largest factor of s that is less than the
    desired chunk size, but not less than half, which is too much.  If no such
    factor exists then we just go with the original chunk size and accept an
    uneven chunk at the end.

    If c is larger than s then we want the largest multiple of s that is still
    smaller than c.
    """
    if c <= s:
        try:
            return max(f for f in factors(s) if c / 2 <= f <= c)
        except ValueError:  # no matching factors within factor of two
            return max(1, int(c))
    else:
        return c // s * s


def _get_chunk_shape(a):
    s = np.asarray(a.shape, dtype=int)
    return s[len(s) * (None,) + (slice(None),)]


def from_array(
    x,
    chunks="auto",
    name=None,
    lock=False,
    asarray=None,
    fancy=True,
    getitem=None,
    meta=None,
    inline_array=False,
):
    """Create dask array from something that looks like an array.

    Input must have a ``.shape``, ``.ndim``, ``.dtype`` and support numpy-style slicing.

    Parameters
    ----------
    x : array_like
    chunks : int, tuple
        How to chunk the array. Must be one of the following forms:

        - A blocksize like 1000.
        - A blockshape like (1000, 1000).
        - Explicit sizes of all blocks along all dimensions like
          ((1000, 1000, 500), (400, 400)).
        - A size in bytes, like "100 MiB" which will choose a uniform
          block-like shape
        - The word "auto" which acts like the above, but uses a configuration
          value ``array.chunk-size`` for the chunk size

        -1 or None as a blocksize indicate the size of the corresponding
        dimension.
    name : str or bool, optional
        The key name to use for the array. Defaults to a hash of ``x``.

        Hashing is useful if the same value of ``x`` is used to create multiple
        arrays, as Dask can then recognise that they're the same and
        avoid duplicate computations. However, it can also be slow, and if the
        array is not contiguous it is copied for hashing. If the array uses
        stride tricks (such as :func:`numpy.broadcast_to` or
        :func:`skimage.util.view_as_windows`) to have a larger logical
        than physical size, this copy can cause excessive memory usage.

        If you don't need the deduplication provided by hashing, use
        ``name=False`` to generate a random name instead of hashing, which
        avoids the pitfalls described above. Using ``name=True`` is
        equivalent to the default.

        By default, hashing uses python's standard sha1. This behaviour can be
        changed by installing cityhash, xxhash or murmurhash. If installed,
        a large-factor speedup can be obtained in the tokenisation step.

        .. note::

           Because this ``name`` is used as the key in task graphs, you should
           ensure that it uniquely identifies the data contained within. If
           you'd like to provide a descriptive name that is still unique, combine
           the descriptive name with :func:`dask.base.tokenize` of the
           ``array_like``. See :ref:`graphs` for more.

    lock : bool or Lock, optional
        If ``x`` doesn't support concurrent reads then provide a lock here, or
        pass in True to have dask.array create one for you.
    asarray : bool, optional
        If True then call np.asarray on chunks to convert them to numpy arrays.
        If False then chunks are passed through unchanged.
        If None (default) then we use True if the ``__array_function__`` method
        is undefined.
    fancy : bool, optional
        If ``x`` doesn't support fancy indexing (e.g. indexing with lists or
        arrays) then set to False. Default is True.
    meta : Array-like, optional
        The metadata for the resulting dask array.  This is the kind of array
        that will result from slicing the input array.
        Defaults to the input array.
    inline_array : bool, default False
        How to include the array in the task graph. By default
        (``inline_array=False``) the array is included in a task by itself,
        and each chunk refers to that task by its key.

        .. code-block:: python

           >>> x = h5py.File("data.h5")["/x"]  # doctest: +SKIP
           >>> a = da.from_array(x, chunks=500)  # doctest: +SKIP
           >>> dict(a.dask)  # doctest: +SKIP
           {
              'array-original-<name>': <HDF5 dataset ...>,
              ('array-<name>', 0): (getitem, "array-original-<name>", ...),
              ('array-<name>', 1): (getitem, "array-original-<name>", ...)
           }

        With ``inline_array=True``, Dask will instead inline the array directly
        in the values of the task graph.

        .. code-block:: python

           >>> a = da.from_array(x, chunks=500, inline_array=True)  # doctest: +SKIP
           >>> dict(a.dask)  # doctest: +SKIP
           {
              ('array-<name>', 0): (getitem, <HDF5 dataset ...>, ...),
              ('array-<name>', 1): (getitem, <HDF5 dataset ...>, ...)
           }

        Note that there's no key in the task graph with just the array `x`
        anymore. Instead it's placed directly in the values.

        The right choice for ``inline_array`` depends on several factors,
        including the size of ``x``, how expensive it is to create, which
        scheduler you're using, and the pattern of downstream computations.
        As a heuristic, ``inline_array=True`` may be the right choice when
        the array ``x`` is cheap to serialize and deserialize (since it's
        included in the graph many times) and if you're experiencing ordering
        issues (see :ref:`order` for more).

        This has no effect when ``x`` is a NumPy array.

    Examples
    --------

    >>> x = h5py.File('...')['/data/path']  # doctest: +SKIP
    >>> a = da.from_array(x, chunks=(1000, 1000))  # doctest: +SKIP

    If your underlying datastore does not support concurrent reads then include
    the ``lock=True`` keyword argument or ``lock=mylock`` if you want multiple
    arrays to coordinate around the same lock.

    >>> a = da.from_array(x, chunks=(1000, 1000), lock=True)  # doctest: +SKIP

    If your underlying datastore has a ``.chunks`` attribute (as h5py and zarr
    datasets do) then a multiple of that chunk shape will be used if you
    do not provide a chunk shape.

    >>> a = da.from_array(x, chunks='auto')  # doctest: +SKIP
    >>> a = da.from_array(x, chunks='100 MiB')  # doctest: +SKIP
    >>> a = da.from_array(x)  # doctest: +SKIP

    If providing a name, ensure that it is unique

    >>> import dask.base
    >>> token = dask.base.tokenize(x)  # doctest: +SKIP
    >>> a = da.from_array('myarray-' + token)  # doctest: +SKIP

    NumPy ndarrays are eagerly sliced and then embedded in the graph.

    >>> import dask.array
    >>> a = dask.array.from_array(np.array([[1, 2], [3, 4]]), chunks=(1,1))
    >>> a.dask[a.name, 0, 0][0]
    array([1])

    Chunks with exactly-specified, different sizes can be created.

    >>> import numpy as np
    >>> import dask.array as da
    >>> x = np.random.random((100, 6))
    >>> a = da.from_array(x, chunks=((67, 33), (6,)))
    """
    if isinstance(x, Array):
        raise ValueError(
            "Array is already a dask array. Use 'asarray' or " "'rechunk' instead."
        )
    elif is_dask_collection(x):
        warnings.warn(
            "Passing an object to dask.array.from_array which is already a "
            "Dask collection. This can lead to unexpected behavior."
        )

    if isinstance(x, (list, tuple, memoryview) + np.ScalarType):
        x = np.array(x)

    if asarray is None:
        asarray = not hasattr(x, "__array_function__")

    previous_chunks = getattr(x, "chunks", None)

    chunks = normalize_chunks(
        chunks, x.shape, dtype=x.dtype, previous_chunks=previous_chunks
    )

    if name in (None, True):
        token = tokenize(x, chunks)
        original_name = "array-original-" + token
        name = name or "array-" + token
    elif name is False:
        original_name = name = "array-" + str(uuid.uuid1())
    else:
        original_name = name

    if lock is True:
        lock = SerializableLock()

    is_ndarray = type(x) is np.ndarray
    is_single_block = all(len(c) == 1 for c in chunks)
    # Always use the getter for h5py etc. Not using isinstance(x, np.ndarray)
    # because np.matrix is a subclass of np.ndarray.
    if is_ndarray and not is_single_block and not lock:
        # eagerly slice numpy arrays to prevent memory blowup
        # GH5367, GH5601
        slices = slices_from_chunks(chunks)
        keys = product([name], *(range(len(bds)) for bds in chunks))
        values = [x[slc] for slc in slices]
        dsk = dict(zip(keys, values))

    elif is_ndarray and is_single_block:
        # No slicing needed
        dsk = {(name,) + (0,) * x.ndim: x}
    else:
        if getitem is None:
            if fancy:
                getitem = getter
            else:
                getitem = getter_nofancy

        if inline_array:
            get_from = x
        else:
            get_from = original_name

        dsk = getem(
            get_from,
            chunks,
            getitem=getitem,
            shape=x.shape,
            out_name=name,
            lock=lock,
            asarray=asarray,
            dtype=x.dtype,
        )
        if not inline_array:
            dsk[original_name] = x

    # Workaround for TileDB, its indexing is 1-based,
    # and doesn't seems to support 0-length slicing
    if x.__class__.__module__.split(".")[0] == "tiledb" and hasattr(x, "_ctx_"):
        return Array(dsk, name, chunks, dtype=x.dtype)

    if meta is None:
        meta = x

    return Array(dsk, name, chunks, meta=meta, dtype=getattr(x, "dtype", None))


def from_zarr(
    url,
    component=None,
    storage_options=None,
    chunks=None,
    name=None,
    inline_array=False,
    **kwargs,
):
    """Load array from the zarr storage format

    See https://zarr.readthedocs.io for details about the format.

    Parameters
    ----------
    url: Zarr Array or str or MutableMapping
        Location of the data. A URL can include a protocol specifier like s3://
        for remote data. Can also be any MutableMapping instance, which should
        be serializable if used in multiple processes.
    component: str or None
        If the location is a zarr group rather than an array, this is the
        subcomponent that should be loaded, something like ``'foo/bar'``.
    storage_options: dict
        Any additional parameters for the storage backend (ignored for local
        paths)
    chunks: tuple of ints or tuples of ints
        Passed to :func:`dask.array.from_array`, allows setting the chunks on
        initialisation, if the chunking scheme in the on-disc dataset is not
        optimal for the calculations to follow.
    name : str, optional
         An optional keyname for the array.  Defaults to hashing the input
    kwargs:
        Passed to :class:`zarr.core.Array`.
    inline_array : bool, default False
        Whether to inline the zarr Array in the values of the task graph.
        See :meth:`dask.array.from_array` for an explanation.

    See Also
    --------
    from_array
    """
    import zarr

    storage_options = storage_options or {}
    if isinstance(url, zarr.Array):
        z = url
    elif isinstance(url, (str, os.PathLike)):
        if isinstance(url, os.PathLike):
            url = os.fspath(url)
        mapper = get_mapper(url, **storage_options)
        z = zarr.Array(mapper, read_only=True, path=component, **kwargs)
    else:
        mapper = url
        z = zarr.Array(mapper, read_only=True, path=component, **kwargs)
    chunks = chunks if chunks is not None else z.chunks
    if name is None:
        name = "from-zarr-" + tokenize(z, component, storage_options, chunks, **kwargs)
    return from_array(z, chunks, name=name, inline_array=inline_array)


def to_zarr(
    arr,
    url,
    component=None,
    storage_options=None,
    overwrite=False,
    compute=True,
    return_stored=False,
    **kwargs,
):
    """Save array to the zarr storage format

    See https://zarr.readthedocs.io for details about the format.

    Parameters
    ----------
    arr: dask.array
        Data to store
    url: Zarr Array or str or MutableMapping
        Location of the data. A URL can include a protocol specifier like s3://
        for remote data. Can also be any MutableMapping instance, which should
        be serializable if used in multiple processes.
    component: str or None
        If the location is a zarr group rather than an array, this is the
        subcomponent that should be created/over-written.
    storage_options: dict
        Any additional parameters for the storage backend (ignored for local
        paths)
    overwrite: bool
        If given array already exists, overwrite=False will cause an error,
        where overwrite=True will replace the existing data.
    compute: bool
        See :func:`~dask.array.store` for more details.
    return_stored: bool
        See :func:`~dask.array.store` for more details.
    **kwargs:
        Passed to the :func:`zarr.creation.create` function, e.g., compression options.

    Raises
    ------
    ValueError
        If ``arr`` has unknown chunk sizes, which is not supported by Zarr.

    See Also
    --------
    dask.array.store
    dask.array.Array.compute_chunk_sizes

    """
    import zarr

    if np.isnan(arr.shape).any():
        raise ValueError(
            "Saving a dask array with unknown chunk sizes is not "
            "currently supported by Zarr.%s" % unknown_chunk_message
        )

    if isinstance(url, zarr.Array):
        z = url
        if isinstance(z.store, (dict, MutableMapping)) and "distributed" in config.get(
            "scheduler", ""
        ):
            raise RuntimeError(
                "Cannot store into in memory Zarr Array using "
                "the Distributed Scheduler."
            )
        arr = arr.rechunk(z.chunks)
        return arr.store(z, lock=False, compute=compute, return_stored=return_stored)

    if not _check_regular_chunks(arr.chunks):
        raise ValueError(
            "Attempt to save array to zarr with irregular "
            "chunking, please call `arr.rechunk(...)` first."
        )

    storage_options = storage_options or {}

    if isinstance(url, str):
        mapper = get_mapper(url, **storage_options)
    else:
        # assume the object passed is already a mapper
        mapper = url

    chunks = [c[0] for c in arr.chunks]

    z = zarr.create(
        shape=arr.shape,
        chunks=chunks,
        dtype=arr.dtype,
        store=mapper,
        path=component,
        overwrite=overwrite,
        **kwargs,
    )
    return arr.store(z, lock=False, compute=compute, return_stored=return_stored)


def _check_regular_chunks(chunkset):
    """Check if the chunks are regular

    "Regular" in this context means that along every axis, the chunks all
    have the same size, except the last one, which may be smaller

    Parameters
    ----------
    chunkset: tuple of tuples of ints
        From the ``.chunks`` attribute of an ``Array``

    Returns
    -------
    True if chunkset passes, else False

    Examples
    --------
    >>> import dask.array as da
    >>> arr = da.zeros(10, chunks=(5, ))
    >>> _check_regular_chunks(arr.chunks)
    True

    >>> arr = da.zeros(10, chunks=((3, 3, 3, 1), ))
    >>> _check_regular_chunks(arr.chunks)
    True

    >>> arr = da.zeros(10, chunks=((3, 1, 3, 3), ))
    >>> _check_regular_chunks(arr.chunks)
    False
    """
    for chunks in chunkset:
        if len(chunks) == 1:
            continue
        if len(set(chunks[:-1])) > 1:
            return False
        if chunks[-1] > chunks[0]:
            return False
    return True


def from_delayed(value, shape, dtype=None, meta=None, name=None):
    """Create a dask array from a dask delayed value

    This routine is useful for constructing dask arrays in an ad-hoc fashion
    using dask delayed, particularly when combined with stack and concatenate.

    The dask array will consist of a single chunk.

    Examples
    --------
    >>> import dask
    >>> import dask.array as da
    >>> import numpy as np
    >>> value = dask.delayed(np.ones)(5)
    >>> array = da.from_delayed(value, (5,), dtype=float)
    >>> array
    dask.array<from-value, shape=(5,), dtype=float64, chunksize=(5,), chunktype=numpy.ndarray>
    >>> array.compute()
    array([1., 1., 1., 1., 1.])
    """
    from ..delayed import Delayed, delayed

    if not isinstance(value, Delayed) and hasattr(value, "key"):
        value = delayed(value)

    name = name or "from-value-" + tokenize(value, shape, dtype, meta)
    dsk = {(name,) + (0,) * len(shape): value.key}
    chunks = tuple((d,) for d in shape)
    # TODO: value._key may not be the name of the layer in value.dask
    # This should be fixed after we build full expression graphs
    graph = HighLevelGraph.from_collections(name, dsk, dependencies=[value])
    return Array(graph, name, chunks, dtype=dtype, meta=meta)


def from_func(func, shape, dtype=None, name=None, args=(), kwargs={}):
    """Create dask array in a single block by calling a function

    Calling the provided function with func(*args, **kwargs) should return a
    NumPy array of the indicated shape and dtype.

    Examples
    --------

    >>> a = from_func(np.arange, (3,), dtype='i8', args=(3,))
    >>> a.compute()
    array([0, 1, 2])

    This works particularly well when coupled with dask.array functions like
    concatenate and stack:

    >>> arrays = [from_func(np.array, (), dtype='i8', args=(n,)) for n in range(5)]
    >>> stack(arrays).compute()
    array([0, 1, 2, 3, 4])
    """
    name = name or "from_func-" + tokenize(func, shape, dtype, args, kwargs)
    if args or kwargs:
        func = partial(func, *args, **kwargs)
    dsk = {(name,) + (0,) * len(shape): (func,)}
    chunks = tuple((i,) for i in shape)
    return Array(dsk, name, chunks, dtype)


def common_blockdim(blockdims):
    """Find the common block dimensions from the list of block dimensions

    Currently only implements the simplest possible heuristic: the common
    block-dimension is the only one that does not span fully span a dimension.
    This is a conservative choice that allows us to avoid potentially very
    expensive rechunking.

    Assumes that each element of the input block dimensions has all the same
    sum (i.e., that they correspond to dimensions of the same size).

    Examples
    --------
    >>> common_blockdim([(3,), (2, 1)])
    (2, 1)
    >>> common_blockdim([(1, 2), (2, 1)])
    (1, 1, 1)
    >>> common_blockdim([(2, 2), (3, 1)])  # doctest: +SKIP
    Traceback (most recent call last):
        ...
    ValueError: Chunks do not align
    """
    if not any(blockdims):
        return ()
    non_trivial_dims = {d for d in blockdims if len(d) > 1}
    if len(non_trivial_dims) == 1:
        return first(non_trivial_dims)
    if len(non_trivial_dims) == 0:
        return max(blockdims, key=first)

    if np.isnan(sum(map(sum, blockdims))):
        raise ValueError(
            "Arrays' chunk sizes (%s) are unknown.\n\n"
            "A possible solution:\n"
            "  x.compute_chunk_sizes()" % blockdims
        )

    if len(set(map(sum, non_trivial_dims))) > 1:
        raise ValueError("Chunks do not add up to same value", blockdims)

    # We have multiple non-trivial chunks on this axis
    # e.g. (5, 2) and (4, 3)

    # We create a single chunk tuple with the same total length
    # that evenly divides both, e.g. (4, 1, 2)

    # To accomplish this we walk down all chunk tuples together, finding the
    # smallest element, adding it to the output, and subtracting it from all
    # other elements and remove the element itself.  We stop once we have
    # burned through all of the chunk tuples.
    # For efficiency's sake we reverse the lists so that we can pop off the end
    rchunks = [list(ntd)[::-1] for ntd in non_trivial_dims]
    total = sum(first(non_trivial_dims))
    i = 0

    out = []
    while i < total:
        m = min(c[-1] for c in rchunks)
        out.append(m)
        for c in rchunks:
            c[-1] -= m
            if c[-1] == 0:
                c.pop()
        i += m

    return tuple(out)


def unify_chunks(*args, **kwargs):
    """
    Unify chunks across a sequence of arrays

    This utility function is used within other common operations like
    :func:`dask.array.core.map_blocks` and :func:`dask.array.core.blockwise`.
    It is not commonly used by end-users directly.

    Parameters
    ----------
    *args: sequence of Array, index pairs
        Sequence like (x, 'ij', y, 'jk', z, 'i')

    Examples
    --------
    >>> import dask.array as da
    >>> x = da.ones(10, chunks=((5, 2, 3),))
    >>> y = da.ones(10, chunks=((2, 3, 5),))
    >>> chunkss, arrays = unify_chunks(x, 'i', y, 'i')
    >>> chunkss
    {'i': (2, 3, 2, 3)}

    >>> x = da.ones((100, 10), chunks=(20, 5))
    >>> y = da.ones((10, 100), chunks=(4, 50))
    >>> chunkss, arrays = unify_chunks(x, 'ij', y, 'jk', 'constant', None)
    >>> chunkss  # doctest: +SKIP
    {'k': (50, 50), 'i': (20, 20, 20, 20, 20), 'j': (4, 1, 3, 2)}

    >>> unify_chunks(0, None)
    ({}, [0])

    Returns
    -------
    chunkss : dict
        Map like {index: chunks}.
    arrays : list
        List of rechunked arrays.

    See Also
    --------
    common_blockdim
    """
    if not args:
        return {}, []

    arginds = [
        (asanyarray(a) if ind is not None else a, ind) for a, ind in partition(2, args)
    ]  # [x, ij, y, jk]
    warn = kwargs.get("warn", True)

    arrays, inds = zip(*arginds)
    if all(ind is None for ind in inds):
        return {}, list(arrays)
    if all(ind == inds[0] for ind in inds) and all(
        a.chunks == arrays[0].chunks for a in arrays
    ):
        return dict(zip(inds[0], arrays[0].chunks)), arrays

    nameinds = []
    blockdim_dict = dict()
    max_parts = 0
    for a, ind in arginds:
        if ind is not None:
            nameinds.append((a.name, ind))
            blockdim_dict[a.name] = a.chunks
            max_parts = max(max_parts, a.npartitions)
        else:
            nameinds.append((a, ind))

    chunkss = broadcast_dimensions(nameinds, blockdim_dict, consolidate=common_blockdim)
    nparts = np.prod(list(map(len, chunkss.values())))

    if warn and nparts and nparts >= max_parts * 10:
        warnings.warn(
            "Increasing number of chunks by factor of %d" % (nparts / max_parts),
            PerformanceWarning,
            stacklevel=3,
        )

    arrays = []
    for a, i in arginds:
        if i is None:
            arrays.append(a)
        else:
            chunks = tuple(
                chunkss[j]
                if a.shape[n] > 1
                else a.shape[n]
                if not np.isnan(sum(chunkss[j]))
                else None
                for n, j in enumerate(i)
            )
            if chunks != a.chunks and all(a.chunks):
                arrays.append(a.rechunk(chunks))
            else:
                arrays.append(a)
    return chunkss, arrays


def unpack_singleton(x):
    """

    >>> unpack_singleton([[[[1]]]])
    1
    >>> unpack_singleton(np.array(np.datetime64('2000-01-01')))
    array('2000-01-01', dtype='datetime64[D]')
    """
    while isinstance(x, (list, tuple)):
        try:
            x = x[0]
        except (IndexError, TypeError, KeyError):
            break
    return x


def block(arrays, allow_unknown_chunksizes=False):
    """
    Assemble an nd-array from nested lists of blocks.

    Blocks in the innermost lists are concatenated along the last
    dimension (-1), then these are concatenated along the second-last
    dimension (-2), and so on until the outermost list is reached

    Blocks can be of any dimension, but will not be broadcasted using the normal
    rules. Instead, leading axes of size 1 are inserted, to make ``block.ndim``
    the same for all blocks. This is primarily useful for working with scalars,
    and means that code like ``block([v, 1])`` is valid, where
    ``v.ndim == 1``.

    When the nested list is two levels deep, this allows block matrices to be
    constructed from their components.

    Parameters
    ----------
    arrays : nested list of array_like or scalars (but not tuples)
        If passed a single ndarray or scalar (a nested list of depth 0), this
        is returned unmodified (and not copied).

        Elements shapes must match along the appropriate axes (without
        broadcasting), but leading 1s will be prepended to the shape as
        necessary to make the dimensions match.

    allow_unknown_chunksizes: bool
        Allow unknown chunksizes, such as come from converting from dask
        dataframes.  Dask.array is unable to verify that chunks line up.  If
        data comes from differently aligned sources then this can cause
        unexpected results.

    Returns
    -------
    block_array : ndarray
        The array assembled from the given blocks.

        The dimensionality of the output is equal to the greatest of:
        * the dimensionality of all the inputs
        * the depth to which the input list is nested

    Raises
    ------
    ValueError
        * If list depths are mismatched - for instance, ``[[a, b], c]`` is
          illegal, and should be spelt ``[[a, b], [c]]``
        * If lists are empty - for instance, ``[[a, b], []]``

    See Also
    --------
    concatenate : Join a sequence of arrays together.
    stack : Stack arrays in sequence along a new dimension.
    hstack : Stack arrays in sequence horizontally (column wise).
    vstack : Stack arrays in sequence vertically (row wise).
    dstack : Stack arrays in sequence depth wise (along third dimension).
    vsplit : Split array into a list of multiple sub-arrays vertically.

    Notes
    -----

    When called with only scalars, ``block`` is equivalent to an ndarray
    call. So ``block([[1, 2], [3, 4]])`` is equivalent to
    ``array([[1, 2], [3, 4]])``.

    This function does not enforce that the blocks lie on a fixed grid.
    ``block([[a, b], [c, d]])`` is not restricted to arrays of the form::

        AAAbb
        AAAbb
        cccDD

    But is also allowed to produce, for some ``a, b, c, d``::

        AAAbb
        AAAbb
        cDDDD

    Since concatenation happens along the last axis first, `block` is _not_
    capable of producing the following directly::

        AAAbb
        cccbb
        cccDD

    Matlab's "square bracket stacking", ``[A, B, ...; p, q, ...]``, is
    equivalent to ``block([[A, B, ...], [p, q, ...]])``.
    """

    # This was copied almost verbatim from numpy.core.shape_base.block
    # See numpy license at https://github.com/numpy/numpy/blob/master/LICENSE.txt
    # or NUMPY_LICENSE.txt within this directory

    def atleast_nd(x, ndim):
        x = asanyarray(x)
        diff = max(ndim - x.ndim, 0)
        if diff == 0:
            return x
        else:
            return x[(None,) * diff + (Ellipsis,)]

    def format_index(index):
        return "arrays" + "".join(f"[{i}]" for i in index)

    rec = _Recurser(recurse_if=lambda x: type(x) is list)

    # ensure that the lists are all matched in depth
    list_ndim = None
    any_empty = False
    for index, value, entering in rec.walk(arrays):
        if type(value) is tuple:
            # not strictly necessary, but saves us from:
            #  - more than one way to do things - no point treating tuples like
            #    lists
            #  - horribly confusing behaviour that results when tuples are
            #    treated like ndarray
            raise TypeError(
                "{} is a tuple. "
                "Only lists can be used to arrange blocks, and np.block does "
                "not allow implicit conversion from tuple to ndarray.".format(
                    format_index(index)
                )
            )
        if not entering:
            curr_depth = len(index)
        elif len(value) == 0:
            curr_depth = len(index) + 1
            any_empty = True
        else:
            continue

        if list_ndim is not None and list_ndim != curr_depth:
            raise ValueError(
                "List depths are mismatched. First element was at depth {}, "
                "but there is an element at depth {} ({})".format(
                    list_ndim, curr_depth, format_index(index)
                )
            )
        list_ndim = curr_depth

    # do this here so we catch depth mismatches first
    if any_empty:
        raise ValueError("Lists cannot be empty")

    # convert all the arrays to ndarrays
    arrays = rec.map_reduce(arrays, f_map=asanyarray, f_reduce=list)

    # determine the maximum dimension of the elements
    elem_ndim = rec.map_reduce(arrays, f_map=lambda xi: xi.ndim, f_reduce=max)
    ndim = max(list_ndim, elem_ndim)

    # first axis to concatenate along
    first_axis = ndim - list_ndim

    # Make all the elements the same dimension
    arrays = rec.map_reduce(
        arrays, f_map=lambda xi: atleast_nd(xi, ndim), f_reduce=list
    )

    # concatenate innermost lists on the right, outermost on the left
    return rec.map_reduce(
        arrays,
        f_reduce=lambda xs, axis: concatenate(
            list(xs), axis=axis, allow_unknown_chunksizes=allow_unknown_chunksizes
        ),
        f_kwargs=lambda axis: dict(axis=(axis + 1)),
        axis=first_axis,
    )


def concatenate(seq, axis=0, allow_unknown_chunksizes=False):
    """
    Concatenate arrays along an existing axis

    Given a sequence of dask Arrays form a new dask Array by stacking them
    along an existing dimension (axis=0 by default)

    Parameters
    ----------
    seq: list of dask.arrays
    axis: int
        Dimension along which to align all of the arrays
    allow_unknown_chunksizes: bool
        Allow unknown chunksizes, such as come from converting from dask
        dataframes.  Dask.array is unable to verify that chunks line up.  If
        data comes from differently aligned sources then this can cause
        unexpected results.

    Examples
    --------

    Create slices

    >>> import dask.array as da
    >>> import numpy as np

    >>> data = [da.from_array(np.ones((4, 4)), chunks=(2, 2))
    ...          for i in range(3)]

    >>> x = da.concatenate(data, axis=0)
    >>> x.shape
    (12, 4)

    >>> da.concatenate(data, axis=1).shape
    (4, 12)

    Result is a new dask Array

    See Also
    --------
    stack
    """
    from . import wrap

    seq = [asarray(a, allow_unknown_chunksizes=allow_unknown_chunksizes) for a in seq]

    if not seq:
        raise ValueError("Need array(s) to concatenate")

    seq_metas = [meta_from_array(s) for s in seq]
    _concatenate = concatenate_lookup.dispatch(
        type(max(seq_metas, key=lambda x: getattr(x, "__array_priority__", 0)))
    )
    meta = _concatenate(seq_metas, axis=axis)

    # Promote types to match meta
    seq = [a.astype(meta.dtype) for a in seq]

    # Find output array shape
    ndim = len(seq[0].shape)
    shape = tuple(
        sum(a.shape[i] for a in seq) if i == axis else seq[0].shape[i]
        for i in range(ndim)
    )

    # Drop empty arrays
    seq2 = [a for a in seq if a.size]
    if not seq2:
        seq2 = seq

    if axis < 0:
        axis = ndim + axis
    if axis >= ndim:
        msg = (
            "Axis must be less than than number of dimensions"
            "\nData has %d dimensions, but got axis=%d"
        )
        raise ValueError(msg % (ndim, axis))

    n = len(seq2)
    if n == 0:
        try:
            return wrap.empty_like(meta, shape=shape, chunks=shape, dtype=meta.dtype)
        except TypeError:
            return wrap.empty(shape, chunks=shape, dtype=meta.dtype)
    elif n == 1:
        return seq2[0]

    if not allow_unknown_chunksizes and not all(
        i == axis or all(x.shape[i] == seq2[0].shape[i] for x in seq2)
        for i in range(ndim)
    ):
        if any(map(np.isnan, seq2[0].shape)):
            raise ValueError(
                "Tried to concatenate arrays with unknown"
                " shape %s.\n\nTwo solutions:\n"
                "  1. Force concatenation pass"
                " allow_unknown_chunksizes=True.\n"
                "  2. Compute shapes with "
                "[x.compute_chunk_sizes() for x in seq]" % str(seq2[0].shape)
            )
        raise ValueError("Shapes do not align: %s", [x.shape for x in seq2])

    inds = [list(range(ndim)) for i in range(n)]
    for i, ind in enumerate(inds):
        ind[axis] = -(i + 1)

    uc_args = list(concat(zip(seq2, inds)))
    _, seq2 = unify_chunks(*uc_args, warn=False)

    bds = [a.chunks for a in seq2]

    chunks = (
        seq2[0].chunks[:axis]
        + (sum((bd[axis] for bd in bds), ()),)
        + seq2[0].chunks[axis + 1 :]
    )

    cum_dims = [0] + list(accumulate(add, [len(a.chunks[axis]) for a in seq2]))

    names = [a.name for a in seq2]

    name = "concatenate-" + tokenize(names, axis)
    keys = list(product([name], *[range(len(bd)) for bd in chunks]))

    values = [
        (names[bisect(cum_dims, key[axis + 1]) - 1],)
        + key[1 : axis + 1]
        + (key[axis + 1] - cum_dims[bisect(cum_dims, key[axis + 1]) - 1],)
        + key[axis + 2 :]
        for key in keys
    ]

    dsk = dict(zip(keys, values))
    graph = HighLevelGraph.from_collections(name, dsk, dependencies=seq2)

    return Array(graph, name, chunks, meta=meta)


def load_store_chunk(x, out, index, lock, return_stored, load_stored):
    """
    A function inserted in a Dask graph for storing a chunk.

    Parameters
    ----------
    x: array-like
        An array (potentially a NumPy one)
    out: array-like
        Where to store results too.
    index: slice-like
        Where to store result from ``x`` in ``out``.
    lock: Lock-like or False
        Lock to use before writing to ``out``.
    return_stored: bool
        Whether to return ``out``.
    load_stored: bool
        Whether to return the array stored in ``out``.
        Ignored if ``return_stored`` is not ``True``.

    Examples
    --------

    >>> a = np.ones((5, 6))
    >>> b = np.empty(a.shape)
    >>> load_store_chunk(a, b, (slice(None), slice(None)), False, False, False)
    """

    result = None
    if return_stored and not load_stored:
        result = out

    if lock:
        lock.acquire()
    try:
        if x is not None:
            if is_arraylike(x):
                out[index] = x
            else:
                out[index] = np.asanyarray(x)
        if return_stored and load_stored:
            result = out[index]
    finally:
        if lock:
            lock.release()

    return result


def store_chunk(x, out, index, lock, return_stored):
    return load_store_chunk(x, out, index, lock, return_stored, False)


def load_chunk(out, index, lock):
    return load_store_chunk(None, out, index, lock, True, True)


def insert_to_ooc(
    keys: list,
    chunks: tuple[tuple[int, ...], ...],
    out,
    name: str,
    *,
    lock: Lock | bool = True,
    region: slice | None = None,
    return_stored: bool = False,
    load_stored: bool = False,
) -> dict:
    """
    Creates a Dask graph for storing chunks from ``arr`` in ``out``.

    Parameters
    ----------
    keys: list
        Dask keys of the input array
    chunks: tuple
        Dask chunks of the input array
    out: array-like
        Where to store results to
    name: str
        First element of dask keys
    lock: Lock-like or bool, optional
        Whether to lock or with what (default is ``True``,
        which means a :class:`threading.Lock` instance).
    region: slice-like, optional
        Where in ``out`` to store ``arr``'s results
        (default is ``None``, meaning all of ``out``).
    return_stored: bool, optional
        Whether to return ``out``
        (default is ``False``, meaning ``None`` is returned).
    load_stored: bool, optional
        Whether to handling loading from ``out`` at the same time.
        Ignored if ``return_stored`` is not ``True``.
        (default is ``False``, meaning defer to ``return_stored``).

    Returns
    -------
    dask graph of store operation

    Examples
    --------
    >>> import dask.array as da
    >>> d = da.ones((5, 6), chunks=(2, 3))
    >>> a = np.empty(d.shape)
    >>> insert_to_ooc(d.__dask_keys__(), d.chunks, a, "store-123")  # doctest: +SKIP
    """

    if lock is True:
        lock = Lock()

    slices = slices_from_chunks(chunks)
    if region:
        slices = [fuse_slice(region, slc) for slc in slices]

    if return_stored and load_stored:
        func = load_store_chunk
        args = (load_stored,)
    else:
        func = store_chunk
        args = ()

    dsk = {
        (name,) + t[1:]: (func, t, out, slc, lock, return_stored) + args
        for t, slc in zip(core.flatten(keys), slices)
    }
    return dsk


def retrieve_from_ooc(
    keys: Collection[Hashable], dsk_pre: Mapping, dsk_post: Mapping
) -> dict:
    """
    Creates a Dask graph for loading stored ``keys`` from ``dsk``.

    Parameters
    ----------
    keys: Collection
        A sequence containing Dask graph keys to load
    dsk_pre: Mapping
        A Dask graph corresponding to a Dask Array before computation
    dsk_post: Mapping
        A Dask graph corresponding to a Dask Array after computation

    Examples
    --------
    >>> import dask.array as da
    >>> d = da.ones((5, 6), chunks=(2, 3))
    >>> a = np.empty(d.shape)
    >>> g = insert_to_ooc(d.__dask_keys__(), d.chunks, a, "store-123")
    >>> retrieve_from_ooc(g.keys(), g, {k: k for k in g.keys()})  # doctest: +SKIP
    """
    load_dsk = {
        ("load-" + k[0],) + k[1:]: (load_chunk, dsk_post[k]) + dsk_pre[k][3:-1]
        for k in keys
    }

    return load_dsk


def asarray(
    a, allow_unknown_chunksizes=False, dtype=None, order=None, *, like=None, **kwargs
):
    """Convert the input to a dask array.

    Parameters
    ----------
    a : array-like
        Input data, in any form that can be converted to a dask array. This
        includes lists, lists of tuples, tuples, tuples of tuples, tuples of
        lists and ndarrays.
    allow_unknown_chunksizes: bool
        Allow unknown chunksizes, such as come from converting from dask
        dataframes.  Dask.array is unable to verify that chunks line up.  If
        data comes from differently aligned sources then this can cause
        unexpected results.
    dtype : data-type, optional
        By default, the data-type is inferred from the input data.
    order : {‘C’, ‘F’, ‘A’, ‘K’}, optional
        Memory layout. ‘A’ and ‘K’ depend on the order of input array a.
        ‘C’ row-major (C-style), ‘F’ column-major (Fortran-style) memory
        representation. ‘A’ (any) means ‘F’ if a is Fortran contiguous, ‘C’
        otherwise ‘K’ (keep) preserve input order. Defaults to ‘C’.
    like: array-like
        Reference object to allow the creation of Dask arrays with chunks
        that are not NumPy arrays. If an array-like passed in as ``like``
        supports the ``__array_function__`` protocol, the chunk type of the
        resulting array will be definde by it. In this case, it ensures the
        creation of a Dask array compatible with that passed in via this
        argument. If ``like`` is a Dask array, the chunk type of the
        resulting array will be defined by the chunk type of ``like``.
        Requires NumPy 1.20.0 or higher.

    Returns
    -------
    out : dask array
        Dask array interpretation of a.

    Examples
    --------
    >>> import dask.array as da
    >>> import numpy as np
    >>> x = np.arange(3)
    >>> da.asarray(x)
    dask.array<array, shape=(3,), dtype=int64, chunksize=(3,), chunktype=numpy.ndarray>

    >>> y = [[1, 2, 3], [4, 5, 6]]
    >>> da.asarray(y)
    dask.array<array, shape=(2, 3), dtype=int64, chunksize=(2, 3), chunktype=numpy.ndarray>
    """
    if like is None:
        if isinstance(a, Array):
            return a
        elif hasattr(a, "to_dask_array"):
            return a.to_dask_array()
        elif type(a).__module__.split(".")[0] == "xarray" and hasattr(a, "data"):
            return asarray(a.data)
        elif isinstance(a, (list, tuple)) and any(isinstance(i, Array) for i in a):
            return stack(a, allow_unknown_chunksizes=allow_unknown_chunksizes)
        elif not isinstance(getattr(a, "shape", None), Iterable):
            a = np.asarray(a, dtype=dtype, order=order)
    else:
        if not _numpy_120:
            raise RuntimeError("The use of ``like`` required NumPy >= 1.20")

        like_meta = meta_from_array(like)
        if isinstance(a, Array):
            return a.map_blocks(np.asarray, like=like_meta, dtype=dtype, order=order)
        else:
            a = np.asarray(a, like=like_meta, dtype=dtype, order=order)
    return from_array(a, getitem=getter_inline, **kwargs)


def asanyarray(a, dtype=None, order=None, *, like=None):
    """Convert the input to a dask array.

    Subclasses of ``np.ndarray`` will be passed through as chunks unchanged.

    Parameters
    ----------
    a : array-like
        Input data, in any form that can be converted to a dask array. This
        includes lists, lists of tuples, tuples, tuples of tuples, tuples of
        lists and ndarrays.
    dtype : data-type, optional
        By default, the data-type is inferred from the input data.
    order : {‘C’, ‘F’, ‘A’, ‘K’}, optional
        Memory layout. ‘A’ and ‘K’ depend on the order of input array a.
        ‘C’ row-major (C-style), ‘F’ column-major (Fortran-style) memory
        representation. ‘A’ (any) means ‘F’ if a is Fortran contiguous, ‘C’
        otherwise ‘K’ (keep) preserve input order. Defaults to ‘C’.
    like: array-like
        Reference object to allow the creation of Dask arrays with chunks
        that are not NumPy arrays. If an array-like passed in as ``like``
        supports the ``__array_function__`` protocol, the chunk type of the
        resulting array will be definde by it. In this case, it ensures the
        creation of a Dask array compatible with that passed in via this
        argument. If ``like`` is a Dask array, the chunk type of the
        resulting array will be defined by the chunk type of ``like``.
        Requires NumPy 1.20.0 or higher.

    Returns
    -------
    out : dask array
        Dask array interpretation of a.

    Examples
    --------
    >>> import dask.array as da
    >>> import numpy as np
    >>> x = np.arange(3)
    >>> da.asanyarray(x)
    dask.array<array, shape=(3,), dtype=int64, chunksize=(3,), chunktype=numpy.ndarray>

    >>> y = [[1, 2, 3], [4, 5, 6]]
    >>> da.asanyarray(y)
    dask.array<array, shape=(2, 3), dtype=int64, chunksize=(2, 3), chunktype=numpy.ndarray>
    """
    if like is None:
        if isinstance(a, Array):
            return a
        elif hasattr(a, "to_dask_array"):
            return a.to_dask_array()
        elif type(a).__module__.split(".")[0] == "xarray" and hasattr(a, "data"):
            return asanyarray(a.data)
        elif isinstance(a, (list, tuple)) and any(isinstance(i, Array) for i in a):
            return stack(a)
        elif not isinstance(getattr(a, "shape", None), Iterable):
            a = np.asanyarray(a, dtype=dtype, order=order)
    else:
        if not _numpy_120:
            raise RuntimeError("The use of ``like`` required NumPy >= 1.20")

        like_meta = meta_from_array(like)
        if isinstance(a, Array):
            return a.map_blocks(np.asanyarray, like=like_meta, dtype=dtype, order=order)
        else:
            a = np.asanyarray(a, like=like_meta, dtype=dtype, order=order)
    return from_array(a, chunks=a.shape, getitem=getter_inline, asarray=False)


def is_scalar_for_elemwise(arg):
    """

    >>> is_scalar_for_elemwise(42)
    True
    >>> is_scalar_for_elemwise('foo')
    True
    >>> is_scalar_for_elemwise(True)
    True
    >>> is_scalar_for_elemwise(np.array(42))
    True
    >>> is_scalar_for_elemwise([1, 2, 3])
    True
    >>> is_scalar_for_elemwise(np.array([1, 2, 3]))
    False
    >>> is_scalar_for_elemwise(from_array(np.array(0), chunks=()))
    False
    >>> is_scalar_for_elemwise(np.dtype('i4'))
    True
    """
    # the second half of shape_condition is essentially just to ensure that
    # dask series / frame are treated as scalars in elemwise.
    maybe_shape = getattr(arg, "shape", None)
    shape_condition = not isinstance(maybe_shape, Iterable) or any(
        is_dask_collection(x) for x in maybe_shape
    )

    return (
        np.isscalar(arg)
        or shape_condition
        or isinstance(arg, np.dtype)
        or (isinstance(arg, np.ndarray) and arg.ndim == 0)
    )


def broadcast_shapes(*shapes):
    """
    Determines output shape from broadcasting arrays.

    Parameters
    ----------
    shapes : tuples
        The shapes of the arguments.

    Returns
    -------
    output_shape : tuple

    Raises
    ------
    ValueError
        If the input shapes cannot be successfully broadcast together.
    """
    if len(shapes) == 1:
        return shapes[0]
    out = []
    for sizes in zip_longest(*map(reversed, shapes), fillvalue=-1):
        if np.isnan(sizes).any():
            dim = np.nan
        else:
            dim = 0 if 0 in sizes else np.max(sizes)
        if any(i not in [-1, 0, 1, dim] and not np.isnan(i) for i in sizes):
            raise ValueError(
                "operands could not be broadcast together with "
                "shapes {}".format(" ".join(map(str, shapes)))
            )
        out.append(dim)
    return tuple(reversed(out))


def elemwise(op, *args, out=None, **kwargs):
    """Apply elementwise function across arguments

    Respects broadcasting rules

    Parameters
    ----------
    out : dask array or None
        If out is a dask.array then this overwrites the contents of that array with
        the result

    Examples
    --------
    >>> elemwise(add, x, y)  # doctest: +SKIP
    >>> elemwise(sin, x)  # doctest: +SKIP
    >>> elemwise(sin, x, out=dask_array)  # doctest: +SKIP

    See Also
    --------
    blockwise
    """
    if not {"name", "dtype"}.issuperset(kwargs):
        msg = "%s does not take the following keyword arguments %s"
        raise TypeError(
            msg % (op.__name__, str(sorted(set(kwargs) - {"name", "dtype"})))
        )

    args = [np.asarray(a) if isinstance(a, (list, tuple)) else a for a in args]

    shapes = []
    for arg in args:
        shape = getattr(arg, "shape", ())
        if any(is_dask_collection(x) for x in shape):
            # Want to exclude Delayed shapes and dd.Scalar
            shape = ()
        shapes.append(shape)

    shapes = [s if isinstance(s, Iterable) else () for s in shapes]
    out_ndim = len(
        broadcast_shapes(*shapes)
    )  # Raises ValueError if dimensions mismatch
    expr_inds = tuple(range(out_ndim))[::-1]

    need_enforce_dtype = False
    if "dtype" in kwargs:
        need_enforce_dtype = True
        dt = kwargs["dtype"]
    else:
        # We follow NumPy's rules for dtype promotion, which special cases
        # scalars and 0d ndarrays (which it considers equivalent) by using
        # their values to compute the result dtype:
        # https://github.com/numpy/numpy/issues/6240
        # We don't inspect the values of 0d dask arrays, because these could
        # hold potentially very expensive calculations. Instead, we treat
        # them just like other arrays, and if necessary cast the result of op
        # to match.
        vals = [
            np.empty((1,) * max(1, a.ndim), dtype=a.dtype)
            if not is_scalar_for_elemwise(a)
            else a
            for a in args
        ]
        try:
            dt = apply_infer_dtype(op, vals, {}, "elemwise", suggest_dtype=False)
        except Exception:
            return NotImplemented
        need_enforce_dtype = any(
            not is_scalar_for_elemwise(a) and a.ndim == 0 for a in args
        )

    name = kwargs.get("name", None) or f"{funcname(op)}-{tokenize(op, dt, *args)}"

    blockwise_kwargs = dict(dtype=dt, name=name, token=funcname(op).strip("_"))
    if need_enforce_dtype:
        blockwise_kwargs["enforce_dtype"] = dt
        blockwise_kwargs["enforce_dtype_function"] = op
        op = _enforce_dtype
    result = blockwise(
        op,
        expr_inds,
        *concat(
            (a, tuple(range(a.ndim)[::-1]) if not is_scalar_for_elemwise(a) else None)
            for a in args
        ),
        **blockwise_kwargs,
    )

    return handle_out(out, result)


def handle_out(out, result):
    """Handle out parameters

    If out is a dask.array then this overwrites the contents of that array with
    the result
    """
    if isinstance(out, tuple):
        if len(out) == 1:
            out = out[0]
        elif len(out) > 1:
            raise NotImplementedError("The out parameter is not fully supported")
        else:
            out = None
    if isinstance(out, Array):
        if out.shape != result.shape:
            raise ValueError(
                "Mismatched shapes between result and out parameter. "
                "out=%s, result=%s" % (str(out.shape), str(result.shape))
            )
        out._chunks = result.chunks
        out.dask = result.dask
        out._meta = result._meta
        out._name = result.name
    elif out is not None:
        msg = (
            "The out parameter is not fully supported."
            " Received type %s, expected Dask Array" % type(out).__name__
        )
        raise NotImplementedError(msg)
    else:
        return result


def _enforce_dtype(*args, **kwargs):
    """Calls a function and converts its result to the given dtype.

    The parameters have deliberately been given unwieldy names to avoid
    clashes with keyword arguments consumed by blockwise

    A dtype of `object` is treated as a special case and not enforced,
    because it is used as a dummy value in some places when the result will
    not be a block in an Array.

    Parameters
    ----------
    enforce_dtype : dtype
        Result dtype
    enforce_dtype_function : callable
        The wrapped function, which will be passed the remaining arguments
    """
    dtype = kwargs.pop("enforce_dtype")
    function = kwargs.pop("enforce_dtype_function")

    result = function(*args, **kwargs)
    if hasattr(result, "dtype") and dtype != result.dtype and dtype != object:
        if not np.can_cast(result, dtype, casting="same_kind"):
            raise ValueError(
                "Inferred dtype from function %r was %r "
                "but got %r, which can't be cast using "
                "casting='same_kind'"
                % (funcname(function), str(dtype), str(result.dtype))
            )
        if np.isscalar(result):
            # scalar astype method doesn't take the keyword arguments, so
            # have to convert via 0-dimensional array and back.
            result = result.astype(dtype)
        else:
            try:
                result = result.astype(dtype, copy=False)
            except TypeError:
                # Missing copy kwarg
                result = result.astype(dtype)
    return result


def broadcast_to(x, shape, chunks=None, meta=None):
    """Broadcast an array to a new shape.

    Parameters
    ----------
    x : array_like
        The array to broadcast.
    shape : tuple
        The shape of the desired array.
    chunks : tuple, optional
        If provided, then the result will use these chunks instead of the same
        chunks as the source array. Setting chunks explicitly as part of
        broadcast_to is more efficient than rechunking afterwards. Chunks are
        only allowed to differ from the original shape along dimensions that
        are new on the result or have size 1 the input array.
    meta : empty ndarray
        empty ndarray created with same NumPy backend, ndim and dtype as the
        Dask Array being created (overrides dtype)

    Returns
    -------
    broadcast : dask array

    See Also
    --------
    :func:`numpy.broadcast_to`
    """
    x = asarray(x)
    shape = tuple(shape)

    if meta is None:
        meta = meta_from_array(x)

    if x.shape == shape and (chunks is None or chunks == x.chunks):
        return x

    ndim_new = len(shape) - x.ndim
    if ndim_new < 0 or any(
        new != old for new, old in zip(shape[ndim_new:], x.shape) if old != 1
    ):
        raise ValueError(f"cannot broadcast shape {x.shape} to shape {shape}")

    if chunks is None:
        chunks = tuple((s,) for s in shape[:ndim_new]) + tuple(
            bd if old > 1 else (new,)
            for bd, old, new in zip(x.chunks, x.shape, shape[ndim_new:])
        )
    else:
        chunks = normalize_chunks(
            chunks, shape, dtype=x.dtype, previous_chunks=x.chunks
        )
        for old_bd, new_bd in zip(x.chunks, chunks[ndim_new:]):
            if old_bd != new_bd and old_bd != (1,):
                raise ValueError(
                    "cannot broadcast chunks %s to chunks %s: "
                    "new chunks must either be along a new "
                    "dimension or a dimension of size 1" % (x.chunks, chunks)
                )

    name = "broadcast_to-" + tokenize(x, shape, chunks)
    dsk = {}

    enumerated_chunks = product(*(enumerate(bds) for bds in chunks))
    for new_index, chunk_shape in (zip(*ec) for ec in enumerated_chunks):
        old_index = tuple(
            0 if bd == (1,) else i for bd, i in zip(x.chunks, new_index[ndim_new:])
        )
        old_key = (x.name,) + old_index
        new_key = (name,) + new_index
        dsk[new_key] = (np.broadcast_to, old_key, quote(chunk_shape))

    graph = HighLevelGraph.from_collections(name, dsk, dependencies=[x])
    return Array(graph, name, chunks, dtype=x.dtype, meta=meta)


@derived_from(np)
def broadcast_arrays(*args, subok=False):
    subok = bool(subok)

    to_array = asanyarray if subok else asarray
    args = tuple(to_array(e) for e in args)

    # Unify uneven chunking
    inds = [list(reversed(range(x.ndim))) for x in args]
    uc_args = concat(zip(args, inds))
    _, args = unify_chunks(*uc_args, warn=False)

    shape = broadcast_shapes(*(e.shape for e in args))
    chunks = broadcast_chunks(*(e.chunks for e in args))

    result = [broadcast_to(e, shape=shape, chunks=chunks) for e in args]

    return result


def offset_func(func, offset, *args):
    """Offsets inputs by offset

    >>> double = lambda x: x * 2
    >>> f = offset_func(double, (10,))
    >>> f(1)
    22
    >>> f(300)
    620
    """

    def _offset(*args):
        args2 = list(map(add, args, offset))
        return func(*args2)

    with contextlib.suppress(Exception):
        _offset.__name__ = "offset_" + func.__name__

    return _offset


def chunks_from_arrays(arrays):
    """Chunks tuple from nested list of arrays

    >>> x = np.array([1, 2])
    >>> chunks_from_arrays([x, x])
    ((2, 2),)

    >>> x = np.array([[1, 2]])
    >>> chunks_from_arrays([[x], [x]])
    ((1, 1), (2,))

    >>> x = np.array([[1, 2]])
    >>> chunks_from_arrays([[x, x]])
    ((1,), (2, 2))

    >>> chunks_from_arrays([1, 1])
    ((1, 1),)
    """
    if not arrays:
        return ()
    result = []
    dim = 0

    def shape(x):
        try:
            return x.shape
        except AttributeError:
            return (1,)

    while isinstance(arrays, (list, tuple)):
        result.append(tuple(shape(deepfirst(a))[dim] for a in arrays))
        arrays = arrays[0]
        dim += 1
    return tuple(result)


def deepfirst(seq):
    """First element in a nested list

    >>> deepfirst([[[1, 2], [3, 4]], [5, 6], [7, 8]])
    1
    """
    if not isinstance(seq, (list, tuple)):
        return seq
    else:
        return deepfirst(seq[0])


def shapelist(a):
    """Get the shape of nested list"""
    if type(a) is list:
        return tuple([len(a)] + list(shapelist(a[0])))
    else:
        return ()


def transposelist(arrays, axes, extradims=0):
    """Permute axes of nested list

    >>> transposelist([[1,1,1],[1,1,1]], [2,1])
    [[[1, 1], [1, 1], [1, 1]]]

    >>> transposelist([[1,1,1],[1,1,1]], [2,1], extradims=1)
    [[[[1], [1]], [[1], [1]], [[1], [1]]]]
    """
    if len(axes) != ndimlist(arrays):
        raise ValueError("Length of axes should equal depth of nested arrays")
    if extradims < 0:
        raise ValueError("`newdims` should be positive")
    if len(axes) > len(set(axes)):
        raise ValueError("`axes` should be unique")

    ndim = max(axes) + 1
    shape = shapelist(arrays)
    newshape = [
        shape[axes.index(i)] if i in axes else 1 for i in range(ndim + extradims)
    ]

    result = list(core.flatten(arrays))
    return reshapelist(newshape, result)


def stack(seq, axis=0, allow_unknown_chunksizes=False):
    """
    Stack arrays along a new axis

    Given a sequence of dask arrays, form a new dask array by stacking them
    along a new dimension (axis=0 by default)

    Parameters
    ----------
    seq: list of dask.arrays
    axis: int
        Dimension along which to align all of the arrays
    allow_unknown_chunksizes: bool
        Allow unknown chunksizes, such as come from converting from dask
        dataframes.  Dask.array is unable to verify that chunks line up.  If
        data comes from differently aligned sources then this can cause
        unexpected results.

    Examples
    --------

    Create slices

    >>> import dask.array as da
    >>> import numpy as np

    >>> data = [da.from_array(np.ones((4, 4)), chunks=(2, 2))
    ...         for i in range(3)]

    >>> x = da.stack(data, axis=0)
    >>> x.shape
    (3, 4, 4)

    >>> da.stack(data, axis=1).shape
    (4, 3, 4)

    >>> da.stack(data, axis=-1).shape
    (4, 4, 3)

    Result is a new dask Array

    See Also
    --------
    concatenate
    """
    from . import wrap

    seq = [asarray(a, allow_unknown_chunksizes=allow_unknown_chunksizes) for a in seq]

    if not seq:
        raise ValueError("Need array(s) to stack")
    if not allow_unknown_chunksizes and not all(x.shape == seq[0].shape for x in seq):
        idx = first(i for i in enumerate(seq) if i[1].shape != seq[0].shape)
        raise ValueError(
            "Stacked arrays must have the same shape. The first array had shape "
            f"{seq[0].shape}, while array {idx[0] + 1} has shape {idx[1].shape}."
        )

    meta = np.stack([meta_from_array(a) for a in seq], axis=axis)
    seq = [x.astype(meta.dtype) for x in seq]

    ndim = meta.ndim - 1
    if axis < 0:
        axis = ndim + axis + 1
    shape = tuple(
        len(seq)
        if i == axis
        else (seq[0].shape[i] if i < axis else seq[0].shape[i - 1])
        for i in range(meta.ndim)
    )

    seq2 = [a for a in seq if a.size]
    if not seq2:
        seq2 = seq

    n = len(seq2)
    if n == 0:
        try:
            return wrap.empty_like(meta, shape=shape, chunks=shape, dtype=meta.dtype)
        except TypeError:
            return wrap.empty(shape, chunks=shape, dtype=meta.dtype)

    ind = list(range(ndim))
    uc_args = list(concat((x, ind) for x in seq2))
    _, seq2 = unify_chunks(*uc_args)

    assert len({a.chunks for a in seq2}) == 1  # same chunks
    chunks = seq2[0].chunks[:axis] + ((1,) * n,) + seq2[0].chunks[axis:]

    names = [a.name for a in seq2]
    name = "stack-" + tokenize(names, axis)
    keys = list(product([name], *[range(len(bd)) for bd in chunks]))

    inputs = [
        (names[key[axis + 1]],) + key[1 : axis + 1] + key[axis + 2 :] for key in keys
    ]
    values = [
        (
            getitem,
            inp,
            (slice(None, None, None),) * axis
            + (None,)
            + (slice(None, None, None),) * (ndim - axis),
        )
        for inp in inputs
    ]

    layer = dict(zip(keys, values))
    graph = HighLevelGraph.from_collections(name, layer, dependencies=seq2)

    return Array(graph, name, chunks, meta=meta)


def concatenate3(arrays):
    """Recursive np.concatenate

    Input should be a nested list of numpy arrays arranged in the order they
    should appear in the array itself.  Each array should have the same number
    of dimensions as the desired output and the nesting of the lists.

    >>> x = np.array([[1, 2]])
    >>> concatenate3([[x, x, x], [x, x, x]])
    array([[1, 2, 1, 2, 1, 2],
           [1, 2, 1, 2, 1, 2]])

    >>> concatenate3([[x, x], [x, x], [x, x]])
    array([[1, 2, 1, 2],
           [1, 2, 1, 2],
           [1, 2, 1, 2]])
    """
    # We need this as __array_function__ may not exist on older NumPy versions.
    # And to reduce verbosity.
    NDARRAY_ARRAY_FUNCTION = getattr(np.ndarray, "__array_function__", None)

    arrays = concrete(arrays)
    if not arrays:
        return np.empty(0)

    advanced = max(
        core.flatten(arrays, container=(list, tuple)),
        key=lambda x: getattr(x, "__array_priority__", 0),
    )

    if not all(
        NDARRAY_ARRAY_FUNCTION
        is getattr(type(arr), "__array_function__", NDARRAY_ARRAY_FUNCTION)
        for arr in core.flatten(arrays, container=(list, tuple))
    ):
        try:
            x = unpack_singleton(arrays)
            return _concatenate2(arrays, axes=tuple(range(x.ndim)))
        except TypeError:
            pass

    if concatenate_lookup.dispatch(type(advanced)) is not np.concatenate:
        x = unpack_singleton(arrays)
        return _concatenate2(arrays, axes=list(range(x.ndim)))

    ndim = ndimlist(arrays)
    if not ndim:
        return arrays
    chunks = chunks_from_arrays(arrays)
    shape = tuple(map(sum, chunks))

    def dtype(x):
        try:
            return x.dtype
        except AttributeError:
            return type(x)

    result = np.empty(shape=shape, dtype=dtype(deepfirst(arrays)))

    for (idx, arr) in zip(
        slices_from_chunks(chunks), core.flatten(arrays, container=(list, tuple))
    ):
        if hasattr(arr, "ndim"):
            while arr.ndim < ndim:
                arr = arr[None, ...]
        result[idx] = arr

    return result


def concatenate_axes(arrays, axes):
    """Recursively call np.concatenate along axes"""
    if len(axes) != ndimlist(arrays):
        raise ValueError("Length of axes should equal depth of nested arrays")

    extradims = max(0, deepfirst(arrays).ndim - (max(axes) + 1))
    return concatenate3(transposelist(arrays, axes, extradims=extradims))


def to_hdf5(filename, *args, chunks=True, **kwargs):
    """Store arrays in HDF5 file

    This saves several dask arrays into several datapaths in an HDF5 file.
    It creates the necessary datasets and handles clean file opening/closing.

    Parameters
    ----------
    chunks: tuple or ``True``
        Chunk shape, or ``True`` to pass the chunks from the dask array.
        Defaults to ``True``.

    Examples
    --------

    >>> da.to_hdf5('myfile.hdf5', '/x', x)  # doctest: +SKIP

    or

    >>> da.to_hdf5('myfile.hdf5', {'/x': x, '/y': y})  # doctest: +SKIP

    Optionally provide arguments as though to ``h5py.File.create_dataset``

    >>> da.to_hdf5('myfile.hdf5', '/x', x, compression='lzf', shuffle=True)  # doctest: +SKIP

    >>> da.to_hdf5('myfile.hdf5', '/x', x, chunks=(10,20,30))  # doctest: +SKIP

    This can also be used as a method on a single Array

    >>> x.to_hdf5('myfile.hdf5', '/x')  # doctest: +SKIP

    See Also
    --------
    da.store
    h5py.File.create_dataset
    """
    if len(args) == 1 and isinstance(args[0], dict):
        data = args[0]
    elif len(args) == 2 and isinstance(args[0], str) and isinstance(args[1], Array):
        data = {args[0]: args[1]}
    else:
        raise ValueError("Please provide {'/data/path': array} dictionary")

    import h5py

    with h5py.File(filename, mode="a") as f:
        dsets = [
            f.require_dataset(
                dp,
                shape=x.shape,
                dtype=x.dtype,
                chunks=tuple(c[0] for c in x.chunks) if chunks is True else chunks,
                **kwargs,
            )
            for dp, x in data.items()
        ]
        store(list(data.values()), dsets)


def interleave_none(a, b):
    """

    >>> interleave_none([0, None, 2, None], [1, 3])
    (0, 1, 2, 3)
    """
    result = []
    i = j = 0
    n = len(a) + len(b)
    while i + j < n:
        if a[i] is not None:
            result.append(a[i])
            i += 1
        else:
            result.append(b[j])
            i += 1
            j += 1
    return tuple(result)


def keyname(name, i, okey):
    """

    >>> keyname('x', 3, [None, None, 0, 2])
    ('x', 3, 0, 2)
    """
    return (name, i) + tuple(k for k in okey if k is not None)


def _vindex(x, *indexes):
    """Point wise indexing with broadcasting.

    >>> x = np.arange(56).reshape((7, 8))
    >>> x
    array([[ 0,  1,  2,  3,  4,  5,  6,  7],
           [ 8,  9, 10, 11, 12, 13, 14, 15],
           [16, 17, 18, 19, 20, 21, 22, 23],
           [24, 25, 26, 27, 28, 29, 30, 31],
           [32, 33, 34, 35, 36, 37, 38, 39],
           [40, 41, 42, 43, 44, 45, 46, 47],
           [48, 49, 50, 51, 52, 53, 54, 55]])

    >>> d = from_array(x, chunks=(3, 4))
    >>> result = _vindex(d, [0, 1, 6, 0], [0, 1, 0, 7])
    >>> result.compute()
    array([ 0,  9, 48,  7])
    """
    indexes = replace_ellipsis(x.ndim, indexes)

    nonfancy_indexes = []
    reduced_indexes = []
    for i, ind in enumerate(indexes):
        if isinstance(ind, Number):
            nonfancy_indexes.append(ind)
        elif isinstance(ind, slice):
            nonfancy_indexes.append(ind)
            reduced_indexes.append(slice(None))
        else:
            nonfancy_indexes.append(slice(None))
            reduced_indexes.append(ind)

    nonfancy_indexes = tuple(nonfancy_indexes)
    reduced_indexes = tuple(reduced_indexes)

    x = x[nonfancy_indexes]

    array_indexes = {}
    for i, (ind, size) in enumerate(zip(reduced_indexes, x.shape)):
        if not isinstance(ind, slice):
            ind = np.array(ind, copy=True)
            if ind.dtype.kind == "b":
                raise IndexError("vindex does not support indexing with boolean arrays")
            if ((ind >= size) | (ind < -size)).any():
                raise IndexError(
                    "vindex key has entries out of bounds for "
                    "indexing along axis %s of size %s: %r" % (i, size, ind)
                )
            ind %= size
            array_indexes[i] = ind

    if array_indexes:
        x = _vindex_array(x, array_indexes)

    return x


def _vindex_array(x, dict_indexes):
    """Point wise indexing with only NumPy Arrays."""

    try:
        broadcast_indexes = np.broadcast_arrays(*dict_indexes.values())
    except ValueError as e:
        # note: error message exactly matches numpy
        shapes_str = " ".join(str(a.shape) for a in dict_indexes.values())
        raise IndexError(
            "shape mismatch: indexing arrays could not be "
            "broadcast together with shapes " + shapes_str
        ) from e
    broadcast_shape = broadcast_indexes[0].shape

    lookup = dict(zip(dict_indexes, broadcast_indexes))
    flat_indexes = [
        lookup[i].ravel().tolist() if i in lookup else None for i in range(x.ndim)
    ]
    flat_indexes.extend([None] * (x.ndim - len(flat_indexes)))

    flat_indexes = [
        list(index) if index is not None else index for index in flat_indexes
    ]
    bounds = [list(accumulate(add, (0,) + c)) for c in x.chunks]
    bounds2 = [b for i, b in zip(flat_indexes, bounds) if i is not None]
    axis = _get_axis(flat_indexes)
    token = tokenize(x, flat_indexes)
    out_name = "vindex-merge-" + token

    points = list()
    for i, idx in enumerate(zip(*[i for i in flat_indexes if i is not None])):
        block_idx = [bisect(b, ind) - 1 for b, ind in zip(bounds2, idx)]
        inblock_idx = [
            ind - bounds2[k][j] for k, (ind, j) in enumerate(zip(idx, block_idx))
        ]
        points.append((i, tuple(block_idx), tuple(inblock_idx)))

    chunks = [c for i, c in zip(flat_indexes, x.chunks) if i is None]
    chunks.insert(0, (len(points),) if points else (0,))
    chunks = tuple(chunks)

    if points:
        per_block = groupby(1, points)
        per_block = {k: v for k, v in per_block.items() if v}

        other_blocks = list(
            product(
                *[
                    list(range(len(c))) if i is None else [None]
                    for i, c in zip(flat_indexes, x.chunks)
                ]
            )
        )

        full_slices = [slice(None, None) if i is None else None for i in flat_indexes]

        name = "vindex-slice-" + token
        vindex_merge_name = "vindex-merge-" + token
        dsk = {}
        for okey in other_blocks:
            for i, key in enumerate(per_block):
                dsk[keyname(name, i, okey)] = (
                    _vindex_transpose,
                    (
                        _vindex_slice,
                        (x.name,) + interleave_none(okey, key),
                        interleave_none(
                            full_slices, list(zip(*pluck(2, per_block[key])))
                        ),
                    ),
                    axis,
                )
            dsk[keyname(vindex_merge_name, 0, okey)] = (
                _vindex_merge,
                [list(pluck(0, per_block[key])) for key in per_block],
                [keyname(name, i, okey) for i in range(len(per_block))],
            )

        result_1d = Array(
            HighLevelGraph.from_collections(out_name, dsk, dependencies=[x]),
            out_name,
            chunks,
            x.dtype,
            meta=x._meta,
        )
        return result_1d.reshape(broadcast_shape + result_1d.shape[1:])

    # output has a zero dimension, just create a new zero-shape array with the
    # same dtype
    from .wrap import empty

    result_1d = empty(
        tuple(map(sum, chunks)), chunks=chunks, dtype=x.dtype, name=out_name
    )
    return result_1d.reshape(broadcast_shape + result_1d.shape[1:])


def _get_axis(indexes):
    """Get axis along which point-wise slicing results lie

    This is mostly a hack because I can't figure out NumPy's rule on this and
    can't be bothered to go reading.

    >>> _get_axis([[1, 2], None, [1, 2], None])
    0
    >>> _get_axis([None, [1, 2], [1, 2], None])
    1
    >>> _get_axis([None, None, [1, 2], [1, 2]])
    2
    """
    ndim = len(indexes)
    indexes = [slice(None, None) if i is None else [0] for i in indexes]
    x = np.empty((2,) * ndim)
    x2 = x[tuple(indexes)]
    return x2.shape.index(1)


def _vindex_slice(block, points):
    """Pull out point-wise slices from block"""
    points = [p if isinstance(p, slice) else list(p) for p in points]
    return block[tuple(points)]


def _vindex_transpose(block, axis):
    """Rotate block so that points are on the first dimension"""
    axes = [axis] + list(range(axis)) + list(range(axis + 1, block.ndim))
    return block.transpose(axes)


def _vindex_merge(locations, values):
    """

    >>> locations = [0], [2, 1]
    >>> values = [np.array([[1, 2, 3]]),
    ...           np.array([[10, 20, 30], [40, 50, 60]])]

    >>> _vindex_merge(locations, values)
    array([[ 1,  2,  3],
           [40, 50, 60],
           [10, 20, 30]])
    """
    locations = list(map(list, locations))
    values = list(values)

    n = sum(map(len, locations))

    shape = list(values[0].shape)
    shape[0] = n
    shape = tuple(shape)

    dtype = values[0].dtype

    x = np.empty_like(values[0], dtype=dtype, shape=shape)

    ind = [slice(None, None) for i in range(x.ndim)]
    for loc, val in zip(locations, values):
        ind[0] = loc
        x[tuple(ind)] = val

    return x


def to_npy_stack(dirname, x, axis=0):
    """Write dask array to a stack of .npy files

    This partitions the dask.array along one axis and stores each block along
    that axis as a single .npy file in the specified directory

    Examples
    --------
    >>> x = da.ones((5, 10, 10), chunks=(2, 4, 4))  # doctest: +SKIP
    >>> da.to_npy_stack('data/', x, axis=0)  # doctest: +SKIP

    The ``.npy`` files store numpy arrays for ``x[0:2], x[2:4], and x[4:5]``
    respectively, as is specified by the chunk size along the zeroth axis::

        $ tree data/
        data/
        |-- 0.npy
        |-- 1.npy
        |-- 2.npy
        |-- info

    The ``info`` file stores the dtype, chunks, and axis information of the array.
    You can load these stacks with the :func:`dask.array.from_npy_stack` function.

    >>> y = da.from_npy_stack('data/')  # doctest: +SKIP

    See Also
    --------
    from_npy_stack
    """

    chunks = tuple((c if i == axis else (sum(c),)) for i, c in enumerate(x.chunks))
    xx = x.rechunk(chunks)

    if not os.path.exists(dirname):
        os.mkdir(dirname)

    meta = {"chunks": chunks, "dtype": x.dtype, "axis": axis}

    with open(os.path.join(dirname, "info"), "wb") as f:
        pickle.dump(meta, f)

    name = "to-npy-stack-" + str(uuid.uuid1())
    dsk = {
        (name, i): (np.save, os.path.join(dirname, "%d.npy" % i), key)
        for i, key in enumerate(core.flatten(xx.__dask_keys__()))
    }

    graph = HighLevelGraph.from_collections(name, dsk, dependencies=[xx])
    compute_as_if_collection(Array, graph, list(dsk))


def from_npy_stack(dirname, mmap_mode="r"):
    """Load dask array from stack of npy files

    See :func:`dask.array.to_npy_stack` for docstring.

    Parameters
    ----------
    dirname: string
        Directory of .npy files
    mmap_mode: (None or 'r')
        Read data in memory map mode
    """
    with open(os.path.join(dirname, "info"), "rb") as f:
        info = pickle.load(f)

    dtype = info["dtype"]
    chunks = info["chunks"]
    axis = info["axis"]

    name = "from-npy-stack-%s" % dirname
    keys = list(product([name], *[range(len(c)) for c in chunks]))
    values = [
        (np.load, os.path.join(dirname, "%d.npy" % i), mmap_mode)
        for i in range(len(chunks[axis]))
    ]
    dsk = dict(zip(keys, values))

    return Array(dsk, name, chunks, dtype)


def new_da_object(dsk, name, chunks, meta=None, dtype=None):
    """Generic constructor for dask.array or dask.dataframe objects.

    Decides the appropriate output class based on the type of `meta` provided.
    """
    if is_dataframe_like(meta) or is_series_like(meta) or is_index_like(meta):
        from ..dataframe.core import new_dd_object

        assert all(len(c) == 1 for c in chunks[1:])
        divisions = [None] * (len(chunks[0]) + 1)
        return new_dd_object(dsk, name, meta, divisions)
    else:
        return Array(dsk, name=name, chunks=chunks, meta=meta, dtype=dtype)


class BlockView:
    """An array-like interface to the blocks of an array.

    ``BlockView`` provides an array-like interface
    to the blocks of a dask array.  Numpy-style indexing of a
     ``BlockView`` returns a selection of blocks as a new dask array.

    You can index ``BlockView`` like a numpy array of shape
    equal to the number of blocks in each dimension, (available as
    array.blocks.size).  The dimensionality of the output array matches
    the dimension of this array, even if integer indices are passed.
    Slicing with ``np.newaxis`` or multiple lists is not supported.

    Examples
    --------
    >>> import dask.array as da
    >>> from dask.array.core import BlockView
    >>> x = da.arange(8, chunks=2)
    >>> bv = BlockView(x)
    >>> bv.shape # aliases x.numblocks
    (4,)
    >>> bv.size
    4
    >>> bv[0].compute()
    array([0, 1])
    >>> bv[:3].compute()
    array([0, 1, 2, 3, 4, 5])
    >>> bv[::2].compute()
    array([0, 1, 4, 5])
    >>> bv[[-1, 0]].compute()
    array([6, 7, 0, 1])
    >>> bv.ravel()  # doctest: +NORMALIZE_WHITESPACE
    [dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
     dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
     dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>,
     dask.array<blocks, shape=(2,), dtype=int64, chunksize=(2,), chunktype=numpy.ndarray>]

    Returns
    -------
    An instance of ``da.array.Blockview``
    """

    def __init__(self, array: Array) -> BlockView:
        self._array = array

    def __getitem__(
        self, index: int | Sequence[int] | slice | Sequence[slice]
    ) -> Array:
        from .slicing import normalize_index

        if not isinstance(index, tuple):
            index = (index,)
        if sum(isinstance(ind, (np.ndarray, list)) for ind in index) > 1:
            raise ValueError("Can only slice with a single list")
        if any(ind is None for ind in index):
            raise ValueError("Slicing with np.newaxis or None is not supported")
        index = normalize_index(index, self._array.numblocks)
        index = tuple(slice(k, k + 1) if isinstance(k, Number) else k for k in index)

        name = "blocks-" + tokenize(self._array, index)

        new_keys = self._array._key_array[index]

        chunks = tuple(
            tuple(np.array(c)[i].tolist()) for c, i in zip(self._array.chunks, index)
        )

        keys = product(*(range(len(c)) for c in chunks))

        layer = {(name,) + key: tuple(new_keys[key].tolist()) for key in keys}

        graph = HighLevelGraph.from_collections(name, layer, dependencies=[self._array])
        return Array(graph, name, chunks, meta=self._array)

    def __eq__(self, other: Any) -> bool:
        if isinstance(other, BlockView):
            return self._array is other._array
        else:
            return NotImplemented

    @property
    def size(self) -> int:
        """
        The total number of blocks in the array.
        """
        return np.prod(self.shape)

    @property
    def shape(self) -> tuple[int, ...]:
        """
        The number of blocks per axis. Alias of ``dask.array.numblocks``.
        """
        return self._array.numblocks

    def ravel(self) -> list[Array]:
        """
        Return a flattened list of all the blocks in the array in C order.
        """
        return [self[idx] for idx in np.ndindex(self.shape)]


from .blockwise import blockwise
from .utils import compute_meta, meta_from_array