nan
1994d38ccd
- Introduced tests for color image handling, ensuring source preservation and premultiplied alpha usage. - Implemented tests for brush stroke operations, validating segment boundaries and brush anti-aliasing. - Added tests for frame anti-aliasing, confirming correct rendering behavior with various alpha values and depth settings. - Created tests for frame projection, ensuring accurate overlap calculations and half-plane clipping. - Developed tests for image premultiplied retention, verifying behavior during loading and packing processes. - Added undo functionality tests for mask operations, ensuring single transaction behavior during edits.
137 lines
5.5 KiB
Python
137 lines
5.5 KiB
Python
import sys
|
|
from pathlib import Path
|
|
from unittest.mock import patch
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
sys.path.insert(0, str(Path(__file__).parents[1] / "src"))
|
|
|
|
from anyimage.common.image import premultiplied_rgba
|
|
from anyimage.operators.image_edit_tool import frame
|
|
|
|
|
|
def source(rgba, *, bounds=(-1, 1, -1, 1), depth=1, slope=0, active=True):
|
|
rgba = np.asarray(rgba, dtype=np.float32)
|
|
height, width = rgba.shape[:2]
|
|
world = np.eye(4)
|
|
world[2, 3] = -depth
|
|
world[2, 0] = slope
|
|
result = frame.frame_source_projection(
|
|
np.eye(4), np.eye(4), world, bounds, (1, 1), perspective=False
|
|
)
|
|
result.update(
|
|
rgba=rgba, pixels=premultiplied_rgba(np.flipud(rgba).ravel(), (width, height)),
|
|
image_size=(width, height), active=active, name="Active" if active else "Bottom",
|
|
)
|
|
return result
|
|
|
|
|
|
def render(sources, size=(1, 1)):
|
|
return np.flipud(frame.composite_frame_pixels(
|
|
sources, ((0, 1), (1, 1), (1, 0), (0, 0)), size
|
|
).reshape(size[1], size[0], 4))
|
|
|
|
|
|
def test_white_half_coverage_does_not_mix_canvas_black():
|
|
image = source([[[1, 1, 1, 1]]], bounds=(-1, 0, -1, 1))
|
|
np.testing.assert_allclose(render((image,)), [[[1, 1, 1, 0.5]]])
|
|
|
|
|
|
def test_uncovered_canvas_has_no_rgb_to_reveal_with_mask_add():
|
|
from anyimage.common.selection import SelectionMask
|
|
from anyimage.operators.image_edit_tool.mask import apply_alpha_mask
|
|
|
|
result = render((source([[[1, 1, 1, 1]]], bounds=(-0.5, 0.5, -0.5, 0.5)),), (8, 8))
|
|
uncovered = result[..., 3] == 0
|
|
assert np.any(uncovered)
|
|
np.testing.assert_array_equal(result[uncovered], 0)
|
|
restored = np.flipud(apply_alpha_mask(
|
|
result, SelectionMask(np.ones((8, 8)), (0, 0, 8, 8)), "ADD"
|
|
).reshape((8, 8, 4)))
|
|
np.testing.assert_array_equal(restored[..., :3][uncovered], 0)
|
|
np.testing.assert_array_equal(restored[..., 3], 1)
|
|
|
|
|
|
def test_partial_foreground_over_transparent_blue():
|
|
bottom = source([[[0, 0, 1, 0]]], depth=2, active=False)
|
|
top = source([[[1, 0, 0, 0.5]]], bounds=(-1, 0, -1, 1))
|
|
np.testing.assert_allclose(render((bottom, top)), [[[0.25, 0, 0.75, 0.25]]])
|
|
|
|
|
|
@pytest.mark.parametrize("alpha", [0, 1e-12, 1e-8, 1e-7])
|
|
def test_near_zero_alpha_keeps_covered_rgb(alpha):
|
|
result = render((source([[[0.2, 0.4, 0.8, alpha]]]),))
|
|
np.testing.assert_allclose(result[..., :3], [[[0.2, 0.4, 0.8]]])
|
|
np.testing.assert_allclose(result[..., 3], 0 if alpha <= 1e-8 else alpha, atol=1e-15)
|
|
|
|
|
|
def test_narrow_source_missed_by_center_is_sampled():
|
|
image = source([[[1, 1, 1, 1]]], bounds=(-0.8, -0.7, -1, 1))
|
|
np.testing.assert_allclose(render((image,)), [[[1, 1, 1, 0.25]]])
|
|
|
|
|
|
@pytest.mark.parametrize("alpha", [0, 1])
|
|
def test_crossing_planes_resolve_per_sample_depth(alpha):
|
|
first = source([[[1, 0, 0, alpha]]], slope=1)
|
|
second = source([[[0, 0, 1, alpha]]], slope=-1, active=False)
|
|
np.testing.assert_allclose(render((first, second)), [[[0.5, 0, 0.5, alpha]]])
|
|
|
|
|
|
def test_aligned_interior_keeps_checkerboard():
|
|
rgba = np.ones((8, 8, 4), dtype=np.float32)
|
|
rgba[..., :3] = (np.indices((8, 8)).sum(axis=0) % 2)[..., None]
|
|
np.testing.assert_allclose(render((source(rgba),), (8, 8)), rgba)
|
|
|
|
|
|
def test_oblique_boundary_matches_reference_and_chunk_sizes():
|
|
top = source([[[1, 0, 0, 0.5]]], bounds=(-0.7, 0.7, -0.7, 0.7))
|
|
angle = 0.3
|
|
world = np.eye(4)
|
|
world[:2, :2] = ((np.cos(angle), -np.sin(angle)), (np.sin(angle), np.cos(angle)))
|
|
world[2, 3] = -1
|
|
top.update(frame.frame_source_projection(
|
|
np.eye(4), np.eye(4), world, (-0.7, 0.7, -0.7, 0.7), (1, 1), perspective=False
|
|
))
|
|
bottom = source([[[0, 0, 1, 0]]], depth=2, active=False)
|
|
sources = (bottom, top)
|
|
size = 17
|
|
reference = np.zeros((size, size, 4), dtype=np.float32)
|
|
for dy in (0.125, 0.375, 0.625, 0.875):
|
|
for dx in (0.125, 0.375, 0.625, 0.875):
|
|
x, y = np.meshgrid((np.arange(size) + dx) / size, 1 - (np.arange(size) + dy) / size)
|
|
sample, _ = frame._composite_frame_samples(sources, np.stack((x, y), axis=-1))
|
|
reference += sample / 16
|
|
for rows in (1, 3, 256):
|
|
with patch.object(frame, "PROJECTIVE_CHUNK_ROWS", rows):
|
|
np.testing.assert_allclose(render(sources, (size, size)), reference, atol=1e-6)
|
|
|
|
|
|
@pytest.mark.parametrize("background", [0, 1])
|
|
def test_packed_frame_mask_add_keeps_boundary_color(background):
|
|
from io import BytesIO
|
|
import bpy
|
|
from PIL import Image
|
|
from anyimage.common.image import create_image_edit_result, image_rgba
|
|
from anyimage.common.selection import SelectionMask
|
|
from anyimage.operators.image_edit_tool.mask import apply_alpha_mask
|
|
|
|
bottom = source([[[background, background, background, 0]]], depth=2, active=False)
|
|
top = source([[[1, 0, 0, 1]]], bounds=(-1, 0, -1, 1))
|
|
rgba = render((bottom, top))
|
|
original = bpy.data.images.new("Boundary source", width=1, height=1, alpha=True)
|
|
result = create_image_edit_result(original, rgba.ravel(), (1, 1))
|
|
try:
|
|
packed = np.asarray(Image.open(BytesIO(result.packed_file.data))) / 255
|
|
expected = (np.array([1, 0, 0]) + background) / 2
|
|
np.testing.assert_allclose(packed[0, 0, :3], expected, atol=1 / 255)
|
|
restored = apply_alpha_mask(
|
|
image_rgba(result), SelectionMask(np.ones((1, 1)), (0, 0, 1, 1)), "ADD"
|
|
).reshape(1, 1, 4)
|
|
np.testing.assert_allclose(restored[0, 0, :3], expected, atol=1 / 255)
|
|
assert restored[0, 0, 3] == 1
|
|
assert result.alpha_mode == "PREMUL"
|
|
finally:
|
|
bpy.data.images.remove(original)
|
|
bpy.data.images.remove(result)
|