56 lines
2.4 KiB
Python
56 lines
2.4 KiB
Python
import torch
|
|
import pytest
|
|
from more_math.ImageMathNode import ImageMathNode
|
|
from more_math.AudioMathNode import AudioMathNode
|
|
from more_math.MaskMathNode import MaskMathNode
|
|
|
|
def test_image_mismatch_broadcast_a_longer():
|
|
# a: 2 frames, b: 1 frame
|
|
a = torch.ones((2, 64, 64, 3))
|
|
b = torch.ones((1, 64, 64, 3)) * 0.5
|
|
|
|
# a + b -> [1+0.5, 1+0.5]
|
|
result, = ImageMathNode.execute(V={"V0": a, "V1": b}, F={}, Expression="a + b", length_mismatch="tile")
|
|
assert result.shape[0] == 2
|
|
assert torch.allclose(result, torch.ones((2, 64, 64, 3)) * 1.5)
|
|
|
|
def test_image_mismatch_broadcast_b_longer():
|
|
# a: 1 frame, b: 2 frames
|
|
a = torch.ones((1, 64, 64, 3))
|
|
b = torch.ones((2, 64, 64, 3)) * 0.5
|
|
|
|
# a + b -> [1+0.5, 1+0.5] (a broadcasted to length 2)
|
|
result, = ImageMathNode.execute(V={"V0": a, "V1": b}, F={}, Expression="a + b", length_mismatch="tile")
|
|
assert result.shape[0] == 2
|
|
assert torch.allclose(result, torch.ones((2, 64, 64, 3)) * 1.5)
|
|
|
|
def test_image_mismatch_error():
|
|
a = torch.ones((2, 64, 64, 3))
|
|
b = torch.ones((1, 64, 64, 3)) * 0.5
|
|
|
|
with pytest.raises(ValueError, match="Input 'V1' has shape 1, expected 2 to match largest input."):
|
|
ImageMathNode.execute(V={"V0": a, "V1": b}, F={}, Expression="a + b", length_mismatch="error")
|
|
|
|
def test_image_mismatch_pad_b_longer():
|
|
# a: 1 frame, b: 2 frames
|
|
a = torch.ones((1, 64, 64, 3))
|
|
b = torch.ones((2, 64, 64, 3)) * 0.5
|
|
|
|
# a + b -> [1+0.5, 0+0.5] = [1.5, 0.5]
|
|
result, = ImageMathNode.execute(V={"V0": a, "V1": b}, F={}, Expression="a + b", length_mismatch="pad")
|
|
assert result.shape[0] == 2
|
|
assert torch.allclose(result[0], torch.ones((64, 64, 3)) * 1.5)
|
|
assert torch.allclose(result[1], torch.ones((64, 64, 3)) * 0.5)
|
|
|
|
def test_audio_mismatch_error():
|
|
a = torch.ones((1, 2, 100))
|
|
b = torch.ones((1, 2, 50)) * 0.5
|
|
with pytest.raises(ValueError, match=r"Input 'V1' has shape \(1, 50\), expected \(1, 100\) to match input."):
|
|
AudioMathNode.execute(V={"V0": {"waveform": a, "sample_rate": 44100}, "V1": {"waveform": b, "sample_rate": 44100}}, F={}, Expression="a + b", length_mismatch="error")
|
|
|
|
def test_mask_mismatch_error():
|
|
a = torch.ones((2, 64, 64))
|
|
b = torch.ones((1, 64, 64)) * 0.5
|
|
with pytest.raises(ValueError, match="Input 'V1' has shape 1, expected 2 to match largest input."):
|
|
MaskMathNode.execute(V={"V0": a, "V1": b}, F={}, Expression="a + b", length_mismatch="error")
|