common_utils.py 30.8 KB
Newer Older
1
import collections.abc
2
import contextlib
3
4
import dataclasses
import enum
5
import functools
6
import itertools
7
import os
8
import pathlib
9
import random
10
import re
11
import shutil
12
import sys
13
import tempfile
14
import warnings
15
from collections import defaultdict
16
from subprocess import CalledProcessError, check_output, STDOUT
17
from typing import Callable, Sequence, Tuple, Union
18
19

import numpy as np
20
21
22

import PIL.Image
import pytest
eellison's avatar
eellison committed
23
import torch
24
import torch.testing
25
from PIL import Image
26

27
28
29
from torch.testing._comparison import BooleanPair, NonePair, not_close_error_metas, NumberPair, TensorLikePair
from torchvision import datapoints, io
from torchvision.transforms._functional_tensor import _max_value as get_max_value
30
from torchvision.transforms.v2.functional import to_dtype_image_tensor, to_image_pil, to_image_tensor
31

32

33
IN_OSS_CI = any(os.getenv(var) == "true" for var in ["CIRCLECI", "GITHUB_ACTIONS"])
Philip Meier's avatar
Philip Meier committed
34
35
IN_RE_WORKER = os.environ.get("INSIDE_RE_WORKER") is not None
IN_FBCODE = os.environ.get("IN_FBCODE_TORCHVISION") == "1"
36
CUDA_NOT_AVAILABLE_MSG = "CUDA device not available"
37
MPS_NOT_AVAILABLE_MSG = "MPS device not available"
38
OSS_CI_GPU_NO_CUDA_MSG = "We're in an OSS GPU machine, and this test doesn't need cuda."
39

40
41
42
43
44
45
46
47
48
49
50

@contextlib.contextmanager
def get_tmp_dir(src=None, **kwargs):
    tmp_dir = tempfile.mkdtemp(**kwargs)
    if src is not None:
        os.rmdir(tmp_dir)
        shutil.copytree(src, tmp_dir)
    try:
        yield tmp_dir
    finally:
        shutil.rmtree(tmp_dir)
eellison's avatar
eellison committed
51
52


53
54
55
56
57
def set_rng_seed(seed):
    torch.manual_seed(seed)
    random.seed(seed)


58
class MapNestedTensorObjectImpl:
eellison's avatar
eellison committed
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
    def __init__(self, tensor_map_fn):
        self.tensor_map_fn = tensor_map_fn

    def __call__(self, object):
        if isinstance(object, torch.Tensor):
            return self.tensor_map_fn(object)

        elif isinstance(object, dict):
            mapped_dict = {}
            for key, value in object.items():
                mapped_dict[self(key)] = self(value)
            return mapped_dict

        elif isinstance(object, (list, tuple)):
            mapped_iter = []
            for iter in object:
                mapped_iter.append(self(iter))
            return mapped_iter if not isinstance(object, tuple) else tuple(mapped_iter)

        else:
            return object


def map_nested_tensor_object(object, tensor_map_fn):
    impl = MapNestedTensorObjectImpl(tensor_map_fn)
    return impl(object)


87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
def is_iterable(obj):
    try:
        iter(obj)
        return True
    except TypeError:
        return False


@contextlib.contextmanager
def freeze_rng_state():
    rng_state = torch.get_rng_state()
    if torch.cuda.is_available():
        cuda_rng_state = torch.cuda.get_rng_state()
    yield
    if torch.cuda.is_available():
        torch.cuda.set_rng_state(cuda_rng_state)
    torch.set_rng_state(rng_state)
104
105


106
def cycle_over(objs):
107
    for idx, obj1 in enumerate(objs):
108
        for obj2 in objs[:idx] + objs[idx + 1 :]:
109
            yield obj1, obj2
110
111
112


def int_dtypes():
113
    return (torch.uint8, torch.int8, torch.int16, torch.int32, torch.int64)
114
115
116


def float_dtypes():
117
    return (torch.float32, torch.float64)
118
119
120
121
122
123
124
125


@contextlib.contextmanager
def disable_console_output():
    with contextlib.ExitStack() as stack, open(os.devnull, "w") as devnull:
        stack.enter_context(contextlib.redirect_stdout(devnull))
        stack.enter_context(contextlib.redirect_stderr(devnull))
        yield
126
127


128
def cpu_and_cuda():
129
    import pytest  # noqa
130
131

    return ("cpu", pytest.param("cuda", marks=pytest.mark.needs_cuda))
132
133


134
135
136
137
def cpu_and_cuda_and_mps():
    return cpu_and_cuda() + (pytest.param("mps", marks=pytest.mark.needs_mps),)


138
139
def needs_cuda(test_func):
    import pytest  # noqa
140

141
    return pytest.mark.needs_cuda(test_func)
Nicolas Hug's avatar
Nicolas Hug committed
142
143


144
145
146
147
148
149
def needs_mps(test_func):
    import pytest  # noqa

    return pytest.mark.needs_mps(test_func)


Nicolas Hug's avatar
Nicolas Hug committed
150
151
152
def _create_data(height=3, width=3, channels=3, device="cpu"):
    # TODO: When all relevant tests are ported to pytest, turn this into a module-level fixture
    tensor = torch.randint(0, 256, (channels, height, width), dtype=torch.uint8, device=device)
153
154
155
156
157
158
    data = tensor.permute(1, 2, 0).contiguous().cpu().numpy()
    mode = "RGB"
    if channels == 1:
        mode = "L"
        data = data[..., 0]
    pil_img = Image.fromarray(data, mode=mode)
Nicolas Hug's avatar
Nicolas Hug committed
159
160
161
162
163
    return tensor, pil_img


def _create_data_batch(height=3, width=3, channels=3, num_samples=4, device="cpu"):
    # TODO: When all relevant tests are ported to pytest, turn this into a module-level fixture
164
    batch_tensor = torch.randint(0, 256, (num_samples, channels, height, width), dtype=torch.uint8, device=device)
Nicolas Hug's avatar
Nicolas Hug committed
165
166
167
    return batch_tensor


168
169
170
171
172
173
174
175
176
177
178
179
def get_list_of_videos(tmpdir, num_videos=5, sizes=None, fps=None):
    names = []
    for i in range(num_videos):
        if sizes is None:
            size = 5 * (i + 1)
        else:
            size = sizes[i]
        if fps is None:
            f = 5
        else:
            f = fps[i]
        data = torch.randint(0, 256, (size, 300, 400, 3), dtype=torch.uint8)
180
        name = os.path.join(tmpdir, f"{i}.mp4")
181
182
183
184
185
186
        names.append(name)
        io.write_video(name, data, fps=f)

    return names


Nicolas Hug's avatar
Nicolas Hug committed
187
def _assert_equal_tensor_to_pil(tensor, pil_image, msg=None):
188
    # FIXME: this is handled automatically by `assert_equal` below. Let's remove this in favor of it
Nicolas Hug's avatar
Nicolas Hug committed
189
190
191
192
193
    np_pil_image = np.array(pil_image)
    if np_pil_image.ndim == 2:
        np_pil_image = np_pil_image[:, :, None]
    pil_tensor = torch.as_tensor(np_pil_image.transpose((2, 0, 1)))
    if msg is None:
194
        msg = f"tensor:\n{tensor} \ndid not equal PIL tensor:\n{pil_tensor}"
195
    assert_equal(tensor.cpu(), pil_tensor, msg=msg)
Nicolas Hug's avatar
Nicolas Hug committed
196
197


198
199
200
def _assert_approx_equal_tensor_to_pil(
    tensor, pil_image, tol=1e-5, msg=None, agg_method="mean", allowed_percentage_diff=None
):
201
    # FIXME: this is handled automatically by `assert_close` below. Let's remove this in favor of it
Nicolas Hug's avatar
Nicolas Hug committed
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
    # TODO: we could just merge this into _assert_equal_tensor_to_pil
    np_pil_image = np.array(pil_image)
    if np_pil_image.ndim == 2:
        np_pil_image = np_pil_image[:, :, None]
    pil_tensor = torch.as_tensor(np_pil_image.transpose((2, 0, 1))).to(tensor)

    if allowed_percentage_diff is not None:
        # Assert that less than a given %age of pixels are different
        assert (tensor != pil_tensor).to(torch.float).mean() <= allowed_percentage_diff

    # error value can be mean absolute error, max abs error
    # Convert to float to avoid underflow when computing absolute difference
    tensor = tensor.to(torch.float)
    pil_tensor = pil_tensor.to(torch.float)
    err = getattr(torch, agg_method)(torch.abs(tensor - pil_tensor)).item()
217
    assert err < tol, f"{err} vs {tol}"
Nicolas Hug's avatar
Nicolas Hug committed
218
219
220
221
222
223
224


def _test_fn_on_batch(batch_tensors, fn, scripted_fn_atol=1e-8, **fn_kwargs):
    transformed_batch = fn(batch_tensors, **fn_kwargs)
    for i in range(len(batch_tensors)):
        img_tensor = batch_tensors[i, ...]
        transformed_img = fn(img_tensor, **fn_kwargs)
225
        torch.testing.assert_close(transformed_img, transformed_batch[i, ...], rtol=0, atol=1e-6)
Nicolas Hug's avatar
Nicolas Hug committed
226
227
228
229
230
231

    if scripted_fn_atol >= 0:
        scripted_fn = torch.jit.script(fn)
        # scriptable function test
        s_transformed_batch = scripted_fn(batch_tensors, **fn_kwargs)
        torch.testing.assert_close(transformed_batch, s_transformed_batch, rtol=1e-5, atol=scripted_fn_atol)
232
233
234


def cache(fn):
235
236
    """Similar to :func:`functools.cache` (Python >= 3.8) or :func:`functools.lru_cache` with infinite cache size,
    but this also caches exceptions.
237
238
239
    """
    sentinel = object()
    out_cache = {}
240
    exc_tb_cache = {}
241
242
243
244
245
246
247
248
249

    @functools.wraps(fn)
    def wrapper(*args, **kwargs):
        key = args + tuple(kwargs.values())

        out = out_cache.get(key, sentinel)
        if out is not sentinel:
            return out

250
251
252
        exc_tb = exc_tb_cache.get(key, sentinel)
        if exc_tb is not sentinel:
            raise exc_tb[0].with_traceback(exc_tb[1])
253
254
255
256

        try:
            out = fn(*args, **kwargs)
        except Exception as exc:
257
258
259
260
            # We need to cache the traceback here as well. Otherwise, each re-raise will add the internal pytest
            # traceback frames anew, but they will only be removed once. Thus, the traceback will be ginormous hiding
            # the actual information in the noise. See https://github.com/pytest-dev/pytest/issues/10363 for details.
            exc_tb_cache[key] = exc, exc.__traceback__
261
262
263
264
265
266
            raise exc

        out_cache[key] = out
        return out

    return wrapper
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307


def combinations_grid(**kwargs):
    """Creates a grid of input combinations.

    Each element in the returned sequence is a dictionary containing one possible combination as values.

    Example:
        >>> combinations_grid(foo=("bar", "baz"), spam=("eggs", "ham"))
        [
            {'foo': 'bar', 'spam': 'eggs'},
            {'foo': 'bar', 'spam': 'ham'},
            {'foo': 'baz', 'spam': 'eggs'},
            {'foo': 'baz', 'spam': 'ham'}
        ]
    """
    return [dict(zip(kwargs.keys(), values)) for values in itertools.product(*kwargs.values())]


class ImagePair(TensorLikePair):
    def __init__(
        self,
        actual,
        expected,
        *,
        mae=False,
        **other_parameters,
    ):
        if all(isinstance(input, PIL.Image.Image) for input in [actual, expected]):
            actual, expected = [to_image_tensor(input) for input in [actual, expected]]

        super().__init__(actual, expected, **other_parameters)
        self.mae = mae

    def compare(self) -> None:
        actual, expected = self.actual, self.expected

        self._compare_attributes(actual, expected)
        actual, expected = self._equalize_attributes(actual, expected)

        if self.mae:
308
309
            if actual.dtype is torch.uint8:
                actual, expected = actual.to(torch.int), expected.to(torch.int)
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
            mae = float(torch.abs(actual - expected).float().mean())
            if mae > self.atol:
                self._fail(
                    AssertionError,
                    f"The MAE of the images is {mae}, but only {self.atol} is allowed.",
                )
        else:
            super()._compare_values(actual, expected)


def assert_close(
    actual,
    expected,
    *,
    allow_subclasses=True,
    rtol=None,
    atol=None,
    equal_nan=False,
    check_device=True,
    check_dtype=True,
    check_layout=True,
    check_stride=False,
    msg=None,
    **kwargs,
):
    """Superset of :func:`torch.testing.assert_close` with support for PIL vs. tensor image comparison"""
    __tracebackhide__ = True

    error_metas = not_close_error_metas(
        actual,
        expected,
        pair_types=(
            NonePair,
            BooleanPair,
            NumberPair,
            ImagePair,
            TensorLikePair,
        ),
        allow_subclasses=allow_subclasses,
        rtol=rtol,
        atol=atol,
        equal_nan=equal_nan,
        check_device=check_device,
        check_dtype=check_dtype,
        check_layout=check_layout,
        check_stride=check_stride,
        **kwargs,
    )

    if error_metas:
        raise error_metas[0].to_error(msg)


assert_equal = functools.partial(assert_close, rtol=0, atol=0)


def parametrized_error_message(*args, **kwargs):
    def to_str(obj):
368
        if isinstance(obj, torch.Tensor) and obj.numel() > 30:
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
            return f"tensor(shape={list(obj.shape)}, dtype={obj.dtype}, device={obj.device})"
        elif isinstance(obj, enum.Enum):
            return f"{type(obj).__name__}.{obj.name}"
        else:
            return repr(obj)

    if args or kwargs:
        postfix = "\n".join(
            [
                "",
                "Failure happened for the following parameters:",
                "",
                *[to_str(arg) for arg in args],
                *[f"{name}={to_str(kwarg)}" for name, kwarg in kwargs.items()],
            ]
        )
    else:
        postfix = ""

    def wrapper(msg):
        return msg + postfix

    return wrapper


class ArgsKwargs:
    def __init__(self, *args, **kwargs):
        self.args = args
        self.kwargs = kwargs

    def __iter__(self):
        yield self.args
        yield self.kwargs

    def load(self, device="cpu"):
        return ArgsKwargs(
            *(arg.load(device) if isinstance(arg, TensorLoader) else arg for arg in self.args),
            **{
                keyword: arg.load(device) if isinstance(arg, TensorLoader) else arg
                for keyword, arg in self.kwargs.items()
            },
        )


413
414
415
# new v2 default
DEFAULT_SIZE = (17, 11)
# old v2 defaults
416
417
418
419
420
421
422
423
424
425
DEFAULT_SQUARE_SPATIAL_SIZE = 15
DEFAULT_LANDSCAPE_SPATIAL_SIZE = (7, 33)
DEFAULT_PORTRAIT_SPATIAL_SIZE = (31, 9)
DEFAULT_SPATIAL_SIZES = (
    DEFAULT_LANDSCAPE_SPATIAL_SIZE,
    DEFAULT_PORTRAIT_SPATIAL_SIZE,
    DEFAULT_SQUARE_SPATIAL_SIZE,
)


426
def _parse_size(size, *, name="size"):
427
    if size == "random":
428
        raise ValueError("This should never happen")
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
    elif isinstance(size, int) and size > 0:
        return (size, size)
    elif (
        isinstance(size, collections.abc.Sequence)
        and len(size) == 2
        and all(isinstance(length, int) and length > 0 for length in size)
    ):
        return tuple(size)
    else:
        raise pytest.UsageError(
            f"'{name}' can either be `'random'`, a positive integer, or a sequence of two positive integers,"
            f"but got {size} instead."
        )


VALID_EXTRA_DIMS = ((), (4,), (2, 3))
DEGENERATE_BATCH_DIMS = ((0,), (5, 0), (0, 5))

DEFAULT_EXTRA_DIMS = (*VALID_EXTRA_DIMS, *DEGENERATE_BATCH_DIMS)


def from_loader(loader_fn):
    def wrapper(*args, **kwargs):
        device = kwargs.pop("device", "cpu")
        loader = loader_fn(*args, **kwargs)
        return loader.load(device)

    return wrapper


def from_loaders(loaders_fn):
    def wrapper(*args, **kwargs):
        device = kwargs.pop("device", "cpu")
        loaders = loaders_fn(*args, **kwargs)
        for loader in loaders:
            yield loader.load(device)

    return wrapper


@dataclasses.dataclass
class TensorLoader:
    fn: Callable[[Sequence[int], torch.dtype, Union[str, torch.device]], torch.Tensor]
    shape: Sequence[int]
    dtype: torch.dtype

    def load(self, device):
        return self.fn(self.shape, self.dtype, device)


@dataclasses.dataclass
class ImageLoader(TensorLoader):
481
    spatial_size: Tuple[int, int] = dataclasses.field(init=False)
482
    num_channels: int = dataclasses.field(init=False)
483
    memory_format: torch.memory_format = torch.contiguous_format
Philip Meier's avatar
Philip Meier committed
484
    canvas_size: Tuple[int, int] = dataclasses.field(init=False)
485
486

    def __post_init__(self):
487
        self.spatial_size = self.canvas_size = self.shape[-2:]
488
489
        self.num_channels = self.shape[-3]

490
491
492
    def load(self, device):
        return self.fn(self.shape, self.dtype, device, memory_format=self.memory_format)

493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508

NUM_CHANNELS_MAP = {
    "GRAY": 1,
    "GRAY_ALPHA": 2,
    "RGB": 3,
    "RGBA": 4,
}


def get_num_channels(color_space):
    num_channels = NUM_CHANNELS_MAP.get(color_space)
    if not num_channels:
        raise pytest.UsageError(f"Can't determine the number of channels for color space {color_space}")
    return num_channels


509
510
511
512
513
514
515
516
517
def make_image(
    size=DEFAULT_SIZE,
    *,
    color_space="RGB",
    batch_dims=(),
    dtype=None,
    device="cpu",
    memory_format=torch.contiguous_format,
):
Philip Meier's avatar
Philip Meier committed
518
    dtype = dtype or torch.uint8
519
520
521
522
523
    max_value = get_max_value(dtype)
    data = torch.testing.make_tensor(
        (*batch_dims, get_num_channels(color_space), *size),
        low=0,
        high=max_value,
Philip Meier's avatar
Philip Meier committed
524
        dtype=dtype,
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
        device=device,
        memory_format=memory_format,
    )
    if color_space in {"GRAY_ALPHA", "RGBA"}:
        data[..., -1, :, :] = max_value

    return datapoints.Image(data)


def make_image_tensor(*args, **kwargs):
    return make_image(*args, **kwargs).as_subclass(torch.Tensor)


def make_image_pil(*args, **kwargs):
    return to_image_pil(make_image(*args, **kwargs))


542
def make_image_loader(
543
    size=DEFAULT_PORTRAIT_SPATIAL_SIZE,
544
545
546
547
548
    *,
    color_space="RGB",
    extra_dims=(),
    dtype=torch.float32,
    constant_alpha=True,
549
    memory_format=torch.contiguous_format,
550
):
551
552
    if not constant_alpha:
        raise ValueError("This should never happen")
553
    size = _parse_size(size)
554
555
    num_channels = get_num_channels(color_space)

556
    def fn(shape, dtype, device, memory_format):
557
558
559
560
561
562
563
564
        *batch_dims, _, height, width = shape
        return make_image(
            (height, width),
            color_space=color_space,
            batch_dims=batch_dims,
            dtype=dtype,
            device=device,
            memory_format=memory_format,
565
        )
566

567
    return ImageLoader(fn, shape=(*extra_dims, num_channels, *size), dtype=dtype, memory_format=memory_format)
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589


def make_image_loaders(
    *,
    sizes=DEFAULT_SPATIAL_SIZES,
    color_spaces=(
        "GRAY",
        "GRAY_ALPHA",
        "RGB",
        "RGBA",
    ),
    extra_dims=DEFAULT_EXTRA_DIMS,
    dtypes=(torch.float32, torch.float64, torch.uint8),
    constant_alpha=True,
):
    for params in combinations_grid(size=sizes, color_space=color_spaces, extra_dims=extra_dims, dtype=dtypes):
        yield make_image_loader(**params, constant_alpha=constant_alpha)


make_images = from_loaders(make_image_loaders)


590
def make_image_loader_for_interpolation(
591
    size=(233, 147), *, color_space="RGB", dtype=torch.uint8, memory_format=torch.contiguous_format
592
):
593
    size = _parse_size(size)
594
595
    num_channels = get_num_channels(color_space)

596
    def fn(shape, dtype, device, memory_format):
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
        height, width = shape[-2:]

        image_pil = (
            PIL.Image.open(pathlib.Path(__file__).parent / "assets" / "encode_jpeg" / "grace_hopper_517x606.jpg")
            .resize((width, height))
            .convert(
                {
                    "GRAY": "L",
                    "GRAY_ALPHA": "LA",
                    "RGB": "RGB",
                    "RGBA": "RGBA",
                }[color_space]
            )
        )

612
613
614
615
616
        image_tensor = to_image_tensor(image_pil)
        if memory_format == torch.contiguous_format:
            image_tensor = image_tensor.to(device=device, memory_format=memory_format, copy=True)
        else:
            image_tensor = image_tensor.to(device=device)
617
        image_tensor = to_dtype_image_tensor(image_tensor, dtype=dtype, scale=True)
618
619
620

        return datapoints.Image(image_tensor)

621
    return ImageLoader(fn, shape=(num_channels, *size), dtype=dtype, memory_format=memory_format)
622
623
624
625
626
627


def make_image_loaders_for_interpolation(
    sizes=((233, 147),),
    color_spaces=("RGB",),
    dtypes=(torch.uint8,),
628
    memory_formats=(torch.contiguous_format, torch.channels_last),
629
):
630
    for params in combinations_grid(size=sizes, color_space=color_spaces, dtype=dtypes, memory_format=memory_formats):
631
632
633
634
        yield make_image_loader_for_interpolation(**params)


@dataclasses.dataclass
635
class BoundingBoxesLoader(TensorLoader):
636
637
    format: datapoints.BoundingBoxFormat
    spatial_size: Tuple[int, int]
Philip Meier's avatar
Philip Meier committed
638
639
640
641
    canvas_size: Tuple[int, int] = dataclasses.field(init=False)

    def __post_init__(self):
        self.canvas_size = self.spatial_size
642
643


644
def make_bounding_box(
Philip Meier's avatar
Philip Meier committed
645
    canvas_size=DEFAULT_SIZE,
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
    *,
    format=datapoints.BoundingBoxFormat.XYXY,
    batch_dims=(),
    dtype=None,
    device="cpu",
):
    def sample_position(values, max_value):
        # We cannot use torch.randint directly here, because it only allows integer scalars as values for low and high.
        # However, if we have batch_dims, we need tensors as limits.
        return torch.stack([torch.randint(max_value - v, ()) for v in values.flatten().tolist()]).reshape(values.shape)

    if isinstance(format, str):
        format = datapoints.BoundingBoxFormat[format]

    dtype = dtype or torch.float32

    if any(dim == 0 for dim in batch_dims):
663
        return datapoints.BoundingBoxes(
Philip Meier's avatar
Philip Meier committed
664
            torch.empty(*batch_dims, 4, dtype=dtype, device=device), format=format, canvas_size=canvas_size
665
666
        )

Philip Meier's avatar
Philip Meier committed
667
668
669
    h, w = [torch.randint(1, c, batch_dims) for c in canvas_size]
    y = sample_position(h, canvas_size[0])
    x = sample_position(w, canvas_size[1])
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684

    if format is datapoints.BoundingBoxFormat.XYWH:
        parts = (x, y, w, h)
    elif format is datapoints.BoundingBoxFormat.XYXY:
        x1, y1 = x, y
        x2 = x1 + w
        y2 = y1 + h
        parts = (x1, y1, x2, y2)
    elif format is datapoints.BoundingBoxFormat.CXCYWH:
        cx = x + w / 2
        cy = y + h / 2
        parts = (cx, cy, w, h)
    else:
        raise ValueError(f"Format {format} is not supported")

685
    return datapoints.BoundingBoxes(
Philip Meier's avatar
Philip Meier committed
686
        torch.stack(parts, dim=-1).to(dtype=dtype, device=device), format=format, canvas_size=canvas_size
687
688
689
    )


690
def make_bounding_box_loader(*, extra_dims=(), format, spatial_size=DEFAULT_PORTRAIT_SPATIAL_SIZE, dtype=torch.float32):
691
692
693
    if isinstance(format, str):
        format = datapoints.BoundingBoxFormat[format]

694
    spatial_size = _parse_size(spatial_size, name="spatial_size")
695
696

    def fn(shape, dtype, device):
697
        *batch_dims, num_coordinates = shape
698
699
700
        if num_coordinates != 4:
            raise pytest.UsageError()

701
        return make_bounding_box(
702
            format=format, canvas_size=spatial_size, batch_dims=batch_dims, dtype=dtype, device=device
703
704
        )

705
    return BoundingBoxesLoader(fn, shape=(*extra_dims[-1:], 4), dtype=dtype, format=format, spatial_size=spatial_size)
706
707
708
709


def make_bounding_box_loaders(
    *,
710
    extra_dims=tuple(d for d in DEFAULT_EXTRA_DIMS if len(d) < 2),
711
    formats=tuple(datapoints.BoundingBoxFormat),
712
    spatial_size=DEFAULT_PORTRAIT_SPATIAL_SIZE,
713
714
715
    dtypes=(torch.float32, torch.float64, torch.int64),
):
    for params in combinations_grid(extra_dims=extra_dims, format=formats, dtype=dtypes):
716
        yield make_bounding_box_loader(**params, spatial_size=spatial_size)
717
718
719
720
721
722
723
724
725


make_bounding_boxes = from_loaders(make_bounding_box_loaders)


class MaskLoader(TensorLoader):
    pass


726
727
728
729
730
731
732
733
734
735
736
737
738
739
def make_detection_mask(size=DEFAULT_SIZE, *, num_objects=5, batch_dims=(), dtype=None, device="cpu"):
    """Make a "detection" mask, i.e. (*, N, H, W), where each object is encoded as one of N boolean masks"""
    return datapoints.Mask(
        torch.testing.make_tensor(
            (*batch_dims, num_objects, *size),
            low=0,
            high=2,
            dtype=dtype or torch.bool,
            device=device,
        )
    )


def make_detection_mask_loader(size=DEFAULT_PORTRAIT_SPATIAL_SIZE, *, num_objects=5, extra_dims=(), dtype=torch.uint8):
740
    # This produces "detection" masks, i.e. `(*, N, H, W)`, where `N` denotes the number of objects
741
    size = _parse_size(size)
742
743

    def fn(shape, dtype, device):
744
745
746
747
        *batch_dims, num_objects, height, width = shape
        return make_detection_mask(
            (height, width), num_objects=num_objects, batch_dims=batch_dims, dtype=dtype, device=device
        )
748
749
750
751
752
753

    return MaskLoader(fn, shape=(*extra_dims, num_objects, *size), dtype=dtype)


def make_detection_mask_loaders(
    sizes=DEFAULT_SPATIAL_SIZES,
754
    num_objects=(1, 0, 5),
755
756
757
758
759
760
761
762
763
764
    extra_dims=DEFAULT_EXTRA_DIMS,
    dtypes=(torch.uint8,),
):
    for params in combinations_grid(size=sizes, num_objects=num_objects, extra_dims=extra_dims, dtype=dtypes):
        yield make_detection_mask_loader(**params)


make_detection_masks = from_loaders(make_detection_mask_loaders)


765
766
767
768
769
770
771
772
773
774
775
def make_segmentation_mask(size=DEFAULT_SIZE, *, num_categories=10, batch_dims=(), dtype=None, device="cpu"):
    """Make a "segmentation" mask, i.e. (*, H, W), where the category is encoded as pixel value"""
    return datapoints.Mask(
        torch.testing.make_tensor(
            (*batch_dims, *size),
            low=0,
            high=num_categories,
            dtype=dtype or torch.uint8,
            device=device,
        )
    )
776
777


778
779
780
781
def make_segmentation_mask_loader(
    size=DEFAULT_PORTRAIT_SPATIAL_SIZE, *, num_categories=10, extra_dims=(), dtype=torch.uint8
):
    # This produces "segmentation" masks, i.e. `(*, H, W)`, where the category is encoded in the values
782
    size = _parse_size(size)
783

784
785
786
787
788
    def fn(shape, dtype, device):
        *batch_dims, height, width = shape
        return make_segmentation_mask(
            (height, width), num_categories=num_categories, batch_dims=batch_dims, dtype=dtype, device=device
        )
789

790
    return MaskLoader(fn, shape=(*extra_dims, *size), dtype=dtype)
791
792
793
794
795


def make_segmentation_mask_loaders(
    *,
    sizes=DEFAULT_SPATIAL_SIZES,
796
    num_categories=(1, 2, 10),
797
798
799
800
801
802
803
804
805
806
807
808
809
    extra_dims=DEFAULT_EXTRA_DIMS,
    dtypes=(torch.uint8,),
):
    for params in combinations_grid(size=sizes, num_categories=num_categories, extra_dims=extra_dims, dtype=dtypes):
        yield make_segmentation_mask_loader(**params)


make_segmentation_masks = from_loaders(make_segmentation_mask_loaders)


def make_mask_loaders(
    *,
    sizes=DEFAULT_SPATIAL_SIZES,
810
811
    num_objects=(1, 0, 5),
    num_categories=(1, 2, 10),
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
    extra_dims=DEFAULT_EXTRA_DIMS,
    dtypes=(torch.uint8,),
):
    yield from make_detection_mask_loaders(sizes=sizes, num_objects=num_objects, extra_dims=extra_dims, dtypes=dtypes)
    yield from make_segmentation_mask_loaders(
        sizes=sizes, num_categories=num_categories, extra_dims=extra_dims, dtypes=dtypes
    )


make_masks = from_loaders(make_mask_loaders)


class VideoLoader(ImageLoader):
    pass


828
829
830
831
def make_video(size=DEFAULT_SIZE, *, num_frames=3, batch_dims=(), **kwargs):
    return datapoints.Video(make_image(size, batch_dims=(*batch_dims, num_frames), **kwargs))


832
833
834
835
def make_video_tensor(*args, **kwargs):
    return make_video(*args, **kwargs).as_subclass(torch.Tensor)


836
def make_video_loader(
837
    size=DEFAULT_PORTRAIT_SPATIAL_SIZE,
838
839
    *,
    color_space="RGB",
840
    num_frames=3,
841
842
843
    extra_dims=(),
    dtype=torch.uint8,
):
844
    size = _parse_size(size)
845

846
    def fn(shape, dtype, device, memory_format):
847
848
849
850
851
852
853
854
855
        *batch_dims, num_frames, _, height, width = shape
        return make_video(
            (height, width),
            num_frames=num_frames,
            batch_dims=batch_dims,
            color_space=color_space,
            dtype=dtype,
            device=device,
            memory_format=memory_format,
856
        )
857
858
859
860
861
862
863
864
865
866
867

    return VideoLoader(fn, shape=(*extra_dims, num_frames, get_num_channels(color_space), *size), dtype=dtype)


def make_video_loaders(
    *,
    sizes=DEFAULT_SPATIAL_SIZES,
    color_spaces=(
        "GRAY",
        "RGB",
    ),
868
    num_frames=(1, 0, 3),
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
    extra_dims=DEFAULT_EXTRA_DIMS,
    dtypes=(torch.uint8, torch.float32, torch.float64),
):
    for params in combinations_grid(
        size=sizes, color_space=color_spaces, num_frames=num_frames, extra_dims=extra_dims, dtype=dtypes
    ):
        yield make_video_loader(**params)


make_videos = from_loaders(make_video_loaders)


class TestMark:
    def __init__(
        self,
        # Tuple of test class name and test function name that identifies the test the mark is applied to. If there is
        # no test class, i.e. a standalone test function, use `None`.
        test_id,
        # `pytest.mark.*` to apply, e.g. `pytest.mark.skip` or `pytest.mark.xfail`
        mark,
        *,
        # Callable, that will be passed an `ArgsKwargs` and should return a boolean to indicate if the mark will be
        # applied. If omitted, defaults to always apply.
        condition=None,
    ):
        self.test_id = test_id
        self.mark = mark
        self.condition = condition or (lambda args_kwargs: True)


def mark_framework_limitation(test_id, reason, condition=None):
    # The purpose of this function is to have a single entry point for skip marks that are only there, because the test
    # framework cannot handle the kernel in general or a specific parameter combination.
    # As development progresses, we can change the `mark.skip` to `mark.xfail` from time to time to see if the skip is
    # still justified.
    # We don't want to use `mark.xfail` all the time, because that actually runs the test until an error happens. Thus,
    # we are wasting CI resources for no reason for most of the time
    return TestMark(test_id, pytest.mark.skip(reason=reason), condition=condition)


class InfoBase:
    def __init__(
        self,
        *,
        # Identifier if the info that shows up the parametrization.
        id,
        # Test markers that will be (conditionally) applied to an `ArgsKwargs` parametrization.
        # See the `TestMark` class for details
        test_marks=None,
        # Additional parameters, e.g. `rtol=1e-3`, passed to `assert_close`. Keys are a 3-tuple of `test_id` (see
        # `TestMark`), the dtype, and the device.
        closeness_kwargs=None,
    ):
        self.id = id

        self.test_marks = test_marks or []
        test_marks_map = defaultdict(list)
        for test_mark in self.test_marks:
            test_marks_map[test_mark.test_id].append(test_mark)
        self._test_marks_map = dict(test_marks_map)

        self.closeness_kwargs = closeness_kwargs or dict()

    def get_marks(self, test_id, args_kwargs):
        return [
            test_mark.mark for test_mark in self._test_marks_map.get(test_id, []) if test_mark.condition(args_kwargs)
        ]

    def get_closeness_kwargs(self, test_id, *, dtype, device):
        if not (isinstance(test_id, tuple) and len(test_id) == 2):
            msg = "`test_id` should be a `Tuple[Optional[str], str]` denoting the test class and function name"
            if callable(test_id):
                msg += ". Did you forget to add the `test_id` fixture to parameters of the test?"
            else:
                msg += f", but got {test_id} instead."
            raise pytest.UsageError(msg)
        if isinstance(device, torch.device):
            device = device.type
        return self.closeness_kwargs.get((test_id, dtype, device), dict())
948
949
950
951


def assert_run_python_script(source_code):
    """Utility to check assertions in an independent Python subprocess.
952

953
    The script provided in the source code should return 0 and not print
954
955
956
957
    anything on stderr or stdout. Modified from scikit-learn test utils.

    Args:
        source_code (str): The Python source code to execute.
958
    """
959
960
961
962
    with get_tmp_dir() as root:
        path = pathlib.Path(root) / "main.py"
        with open(path, "w") as file:
            file.write(source_code)
963
964

        try:
965
            out = check_output([sys.executable, str(path)], stderr=STDOUT)
966
967
968
969
        except CalledProcessError as e:
            raise RuntimeError(f"script errored with output:\n{e.output.decode()}")
        if out != b"":
            raise AssertionError(out.decode())
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989


@contextlib.contextmanager
def assert_no_warnings():
    # The name `catch_warnings` is a misnomer as the context manager does **not** catch any warnings, but rather scopes
    # the warning filters. All changes that are made to the filters while in this context, will be reset upon exit.
    with warnings.catch_warnings():
        warnings.simplefilter("error")
        yield


@contextlib.contextmanager
def ignore_jit_no_profile_information_warning():
    # Calling a scripted object often triggers a warning like
    # `UserWarning: operator() profile_node %$INT1 : int[] = prim::profile_ivalue($INT2) does not have profile information`
    # with varying `INT1` and `INT2`. Since these are uninteresting for us and only clutter the test summary, we ignore
    # them.
    with warnings.catch_warnings():
        warnings.filterwarnings("ignore", message=re.escape("operator() profile_node %"), category=UserWarning)
        yield