Port test/test_transforms.py to pytest

[NicolasHug]

Currently, most tests in test/test_transforms.py rely on unittest.TestCase. Now that we support pytest, we want to remove the use of the unittest module.

For a similar issue see #3956

Instructions

There are many tests in this file, so I bundled them in multiple related groups below. If you're interested in working on this issue, please comment below with "I'm working on <group X>, <group Y>, etc..." so that others don't pick the same tests as you do. Feel free to pick as many groups as you wish, but please don't submit more than 2 groups per PR in order to keep the reviews manageable. Before picking a group, make sure it wasn't picked by another contributor first. Thanks!!

How to port a test to pytest

Porting a test from unittest to pytest is usually fairly straightforward. For a typical example, see https://github.com/pytorch/vision/pull/3907/files:

• take the test method out of the Tester(unittest.TestCase) class and just declare it as a function
• Replace @unittest.skipIf with pytest.mark.skipif(cond, reason=...)
• remove any use of self.assertXYZ.
  • Typically assertEqual(a, b) can be replaced by assert a == b when a and b are pure python objects (scalars, tuples, lists), and otherwise we can rely on assert_equal which is already used in the file.
  • self.assertRaises should be replaced with the pytest.raises(Exp, match=...): context manager, as done in https://github.com/pytorch/vision/pull/3907/files. Same for warnings with pytest.warns
  • self.assertTrue should be replaced with a plain assert
• When a function uses for loops to tests multiple parameter values, one should usepytest.mark.parametrize instead, as done e.g. in https://github.com/pytorch/vision/pull/3907/files.
• It may make sense to keep related tests within a single class. For example here, the tests in group A could be grouped into a TestToPILImage class, the tests in group N could be in TestPad, etc. Not all groups need a dedicated class though, it's on a case-by-case basis.

CC @ShrillShrestha and @zhiqwang in case you're interested in this too :)

---

Groups:

• Groups A: Port Group A test_transforms.py to pytest #3959

  • test_1_channel_ndarray_to_pil_image
  • test_1_channel_tensor_to_pil_image
  • test_2_channel_ndarray_to_pil_image
  • test_2_channel_tensor_to_pil_image
  • test_2d_ndarray_to_pil_image
  • test_2d_tensor_to_pil_image
  • test_3_channel_ndarray_to_pil_image
  • test_3_channel_tensor_to_pil_image
  • test_4_channel_ndarray_to_pil_image
  • test_4_channel_tensor_to_pil_image
  • test_ndarray_bad_types_to_pil_image
  • test_tensor_bad_types_to_pil_image

• Group B port test_accimage_xxx in test_transforms to pytest #3985

  • test_accimage_crop
  • test_accimage_pil_to_tensor
  • test_accimage_resize
  • test_accimage_to_tensor

• Group C Port affine transform tests in test_tranforms to pytest #3984

  • test_affine
  • test_random_affine

• Group D Group G,Group D #3981

  • test_autoaugment
  • test_center_crop
  • test_center_crop_2
  • test_color_jitter

• Group E Group E test_transforms.py port to pytest #4026

  • test_convert_image_dtype_float_to_float
  • test_convert_image_dtype_float_to_int
  • test_convert_image_dtype_int_to_float
  • test_convert_image_dtype_int_to_int
  • test_convert_image_dtype_int_to_int_consistency

• Group F Port to_tensor tests in test_transforms to pytest #3966

  • test_pil_to_tensor
  • test_to_tensor
  • test_to_tensor_with_other_default_dtypes

• Group G Group G,Group D #3981

  • test_five_crop
  • test_ten_crop
  • test_max_value
  • test_linear_transformation

• Group H Port group H test_transforms.py to pytest #3964

  • test_random_apply
  • test_random_choice
  • test_random_order

• Group I Port some test_random_xxx in test_transforms to pytest #3972

  • test_random_crop
  • test_random_erasing
  • test_random_rotation
  • test_randomperspective
  • test_randomperspective_fill

• Group J Port test resize in test_transforms to pytest #3952

  • test_randomresized_params
  • test_resize

• Group K port tests in test_transforms to pytest #3957

  • test_rotate
  • test_rotate_fill
  • test_gaussian_blur_asserts
  • test_lambda

• Group L Refactor test random horizontal flip #3971

  • test_random_horizontal_flip
  • test_random_vertical_flip
  • test_normalize
  • test_normalize_3d_tensor
  • test_normalize_different_dtype

• Group M Port grayscale tests in test_tranforms to pytest #3962

  • test_to_grayscale
  • test_random_grayscale

• Group N Port test pad in test_transforms to pytest #3954

  • test_pad
  • test_pad_raises_with_invalid_pad_sequence_len
  • test_pad_with_mode_F_images
  • test_pad_with_non_constant_padding_modes
  • test_pad_with_tuple_of_pad_values

• Group O Port Group O test to pytest ref #3945 #3955

  • test_random_invert
  • test_random_posterize
  • test_random_solarize
  • test_random_adjust_sharpness
  • test_random_autocontrast
  • test_random_equalize

• Group P port adjusts in test_transforms to pytest #3950

  • test_adjust_brightness
  • test_adjust_contrast
  • test_adjust_gamma
  • test_adjust_hue
  • test_adjust_saturation
  • test_adjust_sharpness
  • test_adjusts_L_mode

cc @pmeier

[saswatpp]

I am working on group P.
I'll try to send PR soon. :)

[ShrillShrestha]

Thanks for mentioning @NicolasHug. I will try working on Group O.

[zhiqwang]

Thanks for mentioning @NicolasHug :) I will try working on Group A.

[AnirudhDagar]

Hi @NicolasHug, I'd like to work on Group I and J.
I'll try to send a PR soon!

[pmeier]

@NicolasHug I can help review. Just ping me.

[Dbhasin1]

Hi @NicolasHug, I'll try working on Group N

[DevPranjal]

Hey @NicolasHug, I will try working on Group F.

[sahilg06]

Hi @NicolasHug, I will try working on Group K.

[anartify]

Hi @NicolasHug, I will try working on Group E.

[harishsdev]

For b,c,d,g groups i will raise PR

[NicolasHug]

Thanks @harishsdev! please make sure to submit at most 2 groups per PR to keep the review manageable :)

[Ishan-Kumar2]

Hi @NicolasHug, I will try working on Group M.

[sahilg06]

@NicolasHug in function test_rotate_fill in group K . Do I need to parameterize any variable as there are 2 for loops but 1 is dependent on the other. Please guide.

[NicolasHug]

Ah, I wish we could have covariant parametrization! This is a common struggle.
In the case of test_rotate_fill parametrizing only over modes should be good enough!

[NicolasHug]

Seems like most groups have been taken already, thanks everyone for your interest! I've created another similar issue for those who are interested: #3956

[sahilg06]

@NicolasHug but 2nd for loop in function test_rotate_fill requires nums_bands-list which depends upon modes-list. So we have to define modes-list inside the function even if we do parameterization. So is that okay as it is defeating the purpose of parameterization?

[NicolasHug]

Ah I missed that.

If I understand the logic correctly here, we should be able to remove this last for loop and just hardcode wrong_num_bands to num_bands + 1.

Basically, set(nums_bands) - {num_bands} is always 1 when num_bands is 3 and always 3 when num_bands is 1.
What's important is that wrong_num_bands is different from num_bands, I don't think we really care about the exact value.

[harishsdev]

PR Raised for

Group B

test_accimage_crop
test_accimage_pil_to_tensor
test_accimage_resize
test_accimage_to_tensor
Group C

test_affine
test_random_affine

[vivekkumar7089]

Hi, @NicolasHug I'd like to work on Group H.
I'll try to send a PR soon!!

[vivekkumar7089]

Hi, @NicolasHug I'd like to work on Group L.
I'll try to send a PR soon!!

[harishsdev]

Hi all,I am working only on Group G,Group D

please take initiate to raise PR for Group B,Group C

[AnirudhDagar]

@harishsdev no worries, I can send in a PR if you are not working on them anymore.

[Ishan-Kumar2]

Hi @AnirudhDagar are you working on both B and C? I can help with one.

[AnirudhDagar]

Sure @Ishan-Kumar2! I was about to start with B, but let me know the one you prefer. I'll take the other. :))

[Ishan-Kumar2]

Okay @AnirudhDagar you can continue with B, I'll try C.

[NicolasHug]

All done! Thank you so much everyone for your help!!