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 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
|
from __future__ import annotations
from importlib.metadata import EntryPoint
from typing import Any
import numpy as np
import pytest
from xarray import set_options
from xarray.core.types import T_Chunks, T_DuckArray, T_NormalizedChunks
from xarray.namedarray._typing import _Chunks
from xarray.namedarray.daskmanager import DaskManager
from xarray.namedarray.parallelcompat import (
KNOWN_CHUNKMANAGERS,
ChunkManagerEntrypoint,
get_chunked_array_type,
guess_chunkmanager,
list_chunkmanagers,
load_chunkmanagers,
)
from xarray.tests import requires_dask
class DummyChunkedArray(np.ndarray):
"""
Mock-up of a chunked array class.
Adds a (non-functional) .chunks attribute by following this example in the numpy docs
https://numpy.org/doc/stable/user/basics.subclassing.html#simple-example-adding-an-extra-attribute-to-ndarray
"""
chunks: T_NormalizedChunks
def __new__(
cls,
shape,
dtype=float,
buffer=None,
offset=0,
strides=None,
order=None,
chunks=None,
):
obj = super().__new__(cls, shape, dtype, buffer, offset, strides, order)
obj.chunks = chunks
return obj
def __array_finalize__(self, obj):
if obj is None:
return
self.chunks = getattr(obj, "chunks", None)
def rechunk(self, chunks, **kwargs):
copied = self.copy()
copied.chunks = chunks
return copied
class DummyChunkManager(ChunkManagerEntrypoint):
"""Mock-up of ChunkManager class for DummyChunkedArray"""
def __init__(self):
self.array_cls = DummyChunkedArray
def is_chunked_array(self, data: Any) -> bool:
return isinstance(data, DummyChunkedArray)
def chunks(self, data: DummyChunkedArray) -> T_NormalizedChunks:
return data.chunks
def normalize_chunks(
self,
chunks: T_Chunks | T_NormalizedChunks,
shape: tuple[int, ...] | None = None,
limit: int | None = None,
dtype: np.dtype | None = None,
previous_chunks: T_NormalizedChunks | None = None,
) -> T_NormalizedChunks:
from dask.array.core import normalize_chunks
return normalize_chunks(chunks, shape, limit, dtype, previous_chunks)
def from_array(
self, data: T_DuckArray | np.typing.ArrayLike, chunks: _Chunks, **kwargs
) -> DummyChunkedArray:
from dask import array as da
return da.from_array(data, chunks, **kwargs)
def rechunk(self, data: DummyChunkedArray, chunks, **kwargs) -> DummyChunkedArray:
return data.rechunk(chunks, **kwargs)
def compute(self, *data: DummyChunkedArray, **kwargs) -> tuple[np.ndarray, ...]:
from dask.array import compute
return compute(*data, **kwargs)
def apply_gufunc(
self,
func,
signature,
*args,
axes=None,
axis=None,
keepdims=False,
output_dtypes=None,
output_sizes=None,
vectorize=None,
allow_rechunk=False,
meta=None,
**kwargs,
):
from dask.array.gufunc import apply_gufunc
return apply_gufunc(
func,
signature,
*args,
axes=axes,
axis=axis,
keepdims=keepdims,
output_dtypes=output_dtypes,
output_sizes=output_sizes,
vectorize=vectorize,
allow_rechunk=allow_rechunk,
meta=meta,
**kwargs,
)
@pytest.fixture
def register_dummy_chunkmanager(monkeypatch):
"""
Mocks the registering of an additional ChunkManagerEntrypoint.
This preserves the presence of the existing DaskManager, so a test that relies on this and DaskManager both being
returned from list_chunkmanagers() at once would still work.
The monkeypatching changes the behavior of list_chunkmanagers when called inside xarray.namedarray.parallelcompat,
but not when called from this tests file.
"""
# Should include DaskManager iff dask is available to be imported
preregistered_chunkmanagers = list_chunkmanagers()
monkeypatch.setattr(
"xarray.namedarray.parallelcompat.list_chunkmanagers",
lambda: {"dummy": DummyChunkManager()} | preregistered_chunkmanagers,
)
yield
class TestGetChunkManager:
def test_get_chunkmanger(self, register_dummy_chunkmanager) -> None:
chunkmanager = guess_chunkmanager("dummy")
assert isinstance(chunkmanager, DummyChunkManager)
def test_get_chunkmanger_via_set_options(self, register_dummy_chunkmanager) -> None:
with set_options(chunk_manager="dummy"):
chunkmanager = guess_chunkmanager(None)
assert isinstance(chunkmanager, DummyChunkManager)
def test_fail_on_known_but_missing_chunkmanager(
self, register_dummy_chunkmanager, monkeypatch
) -> None:
monkeypatch.setitem(KNOWN_CHUNKMANAGERS, "test", "test-package")
with pytest.raises(
ImportError, match="chunk manager 'test' is not available.+test-package"
):
guess_chunkmanager("test")
def test_fail_on_nonexistent_chunkmanager(
self, register_dummy_chunkmanager
) -> None:
with pytest.raises(ValueError, match="unrecognized chunk manager 'foo'"):
guess_chunkmanager("foo")
@requires_dask
def test_get_dask_if_installed(self) -> None:
chunkmanager = guess_chunkmanager(None)
assert isinstance(chunkmanager, DaskManager)
def test_no_chunk_manager_available(self, monkeypatch) -> None:
monkeypatch.setattr("xarray.namedarray.parallelcompat.list_chunkmanagers", dict)
with pytest.raises(ImportError, match="no chunk managers available"):
guess_chunkmanager("foo")
def test_no_chunk_manager_available_but_known_manager_requested(
self, monkeypatch
) -> None:
monkeypatch.setattr("xarray.namedarray.parallelcompat.list_chunkmanagers", dict)
with pytest.raises(ImportError, match="chunk manager 'dask' is not available"):
guess_chunkmanager("dask")
@requires_dask
def test_choose_dask_over_other_chunkmanagers(
self, register_dummy_chunkmanager
) -> None:
chunk_manager = guess_chunkmanager(None)
assert isinstance(chunk_manager, DaskManager)
class TestGetChunkedArrayType:
def test_detect_chunked_arrays(self, register_dummy_chunkmanager) -> None:
dummy_arr = DummyChunkedArray([1, 2, 3])
chunk_manager = get_chunked_array_type(dummy_arr)
assert isinstance(chunk_manager, DummyChunkManager)
def test_ignore_inmemory_arrays(self, register_dummy_chunkmanager) -> None:
dummy_arr = DummyChunkedArray([1, 2, 3])
chunk_manager = get_chunked_array_type(*[dummy_arr, 1.0, np.array([5, 6])])
assert isinstance(chunk_manager, DummyChunkManager)
with pytest.raises(TypeError, match="Expected a chunked array"):
get_chunked_array_type(5.0)
def test_raise_if_no_arrays_chunked(self, register_dummy_chunkmanager) -> None:
with pytest.raises(TypeError, match="Expected a chunked array "):
get_chunked_array_type(*[1.0, np.array([5, 6])])
def test_raise_if_no_matching_chunkmanagers(self) -> None:
dummy_arr = DummyChunkedArray([1, 2, 3])
with pytest.raises(
TypeError, match="Could not find a Chunk Manager which recognises"
):
get_chunked_array_type(dummy_arr)
@requires_dask
def test_detect_dask_if_installed(self) -> None:
import dask.array as da
dask_arr = da.from_array([1, 2, 3], chunks=(1,))
chunk_manager = get_chunked_array_type(dask_arr)
assert isinstance(chunk_manager, DaskManager)
@requires_dask
def test_raise_on_mixed_array_types(self, register_dummy_chunkmanager) -> None:
import dask.array as da
dummy_arr = DummyChunkedArray([1, 2, 3])
dask_arr = da.from_array([1, 2, 3], chunks=(1,))
with pytest.raises(TypeError, match="received multiple types"):
get_chunked_array_type(*[dask_arr, dummy_arr])
def test_bogus_entrypoint() -> None:
# Create a bogus entry-point as if the user broke their setup.cfg
# or is actively developing their new chunk manager
entry_point = EntryPoint(
"bogus", "xarray.bogus.doesnotwork", "xarray.chunkmanagers"
)
with pytest.warns(UserWarning, match="Failed to load chunk manager"):
assert len(load_chunkmanagers([entry_point])) == 0
|