use pytest and add coverage threshold

.prettierrc.yaml

@@ -0,0 +1,5 @@
+editorconfig: true
+overrides:
+  - files: '*.{yaml,yml}'
+    options:
+      singleQuote: true

.vscode/extensions.json

@@ -10,5 +10,7 @@
         "ms-python.vscode-pylance",
         "redhat.vscode-yaml",
         "tamasfe.even-better-toml",
+        "github.vscode-github-actions",
+        "esbenp.prettier-vscode"
     ]
 }

patchmatch/__init__.py

@@ -1,3 +1,3 @@
 __app_id__ = "mauwii/PyPatchMatch"
 __app_name__ = "PyPatchMatch"
-__version__ = "1.0.0"
+__version__ = "1.0.1"

pyproject.toml

@@ -16,7 +16,7 @@ dynamic=["version"]
 license={file="LICENSE"}
 name='PyPatchMatch'
 readme={content-type="text/markdown", file="README.md"}
-requires-python=">=3.9"
+requires-python=">=3.9,<3.11"
 
 [project.urls]
 'Source Code'='https://github.com/mauwii/PyPatchMatch'
@@ -47,13 +47,9 @@ include=["patchmatch", "patchmatch.csrc"]
 exclude='''
 /(
     .git
-    | .hg
-    | .mypy_cache
-    | .tox
     | .venv
-    | _build
-    | buck-out
     | build
+    | csrc
     | dist
 )/
 '''
@@ -62,6 +58,9 @@ line-length=88
 source=['examples', 'patchmatch']
 target-version=['py39']
 
+[tool.coverage.report]
+fail_under=75
+
 [tool.isort]
 profile="black"
 

test_patch_match.py

@@ -0,0 +1,90 @@
+import tempfile
+
+import numpy as np
+import pytest
+from PIL import Image
+
+from patchmatch import patch_match
+
+
+@pytest.fixture
+def input_image():
+    return Image.open("./examples/images/forest_pruned.bmp")
+
+
+@pytest.fixture
+def mask(input_image):
+    return np.zeros((input_image.height, input_image.width), dtype=np.uint8)
+
+
+@pytest.fixture
+def ijmap(input_image):
+    return np.zeros((input_image.height, input_image.width, 3), dtype=np.float32)
+
+
+def test_inpaint_custom_patch_size(input_image):
+    result = Image.fromarray(patch_match.inpaint(image=input_image, patch_size=3))
+    assert isinstance(result, Image.Image)
+
+
+def test_inpaint_global_mask(input_image):
+    patch_match.set_verbose(True)
+    source = np.array(input_image)
+    source[:100, :100] = 255
+    global_mask = np.zeros_like(source[..., 0])
+    global_mask[:100, :100] = 1
+    result = Image.fromarray(
+        patch_match.inpaint(image=source, global_mask=global_mask, patch_size=3)
+    )
+    assert isinstance(result, Image.Image)
+
+
+def test_download_url_to_file():
+    url = "https://github.com/mauwii/PyPatchMatch/raw/main/examples/images/forest.bmp"
+    dst = tempfile.mkstemp().__dir__
+    patch_match.download_url_to_file(url=url, dst=str(dst))
+    read_image = Image.open(str(dst))
+    assert isinstance(read_image, Image.Image)
+
+
+def test_inpaint_regularity(input_image, ijmap):
+    ijmap = ijmap
+    result = patch_match.inpaint_regularity(input_image, None, ijmap, patch_size=3)
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_guide_weight(input_image, ijmap):
+    ijmap = ijmap
+    result = patch_match.inpaint_regularity(input_image, None, ijmap, guide_weight=0.5)
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_global_mask(input_image, mask, ijmap):
+    ijmap = ijmap
+    global_mask = mask
+    result = patch_match.inpaint_regularity(
+        input_image, None, ijmap, global_mask=global_mask
+    )
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_mask(input_image, mask, ijmap):
+    ijmap = ijmap
+    mask = mask
+    result = patch_match.inpaint_regularity(input_image, mask, ijmap)
+    assert isinstance(result, np.ndarray)
+
+
+def test_canonize_mask_array_2d_uint8():
+    mask = np.zeros((10, 10), dtype=np.uint8)
+    result = patch_match._canonize_mask_array(mask)
+    assert isinstance(result, np.ndarray)
+    assert result.ndim == 3
+    assert result.shape[2] == 1
+    assert result.dtype == np.uint8
+
+
+def test_canonize_mask_array_invalid_input():
+    with pytest.raises(AssertionError):
+        mask = np.zeros((10, 10, 3), dtype=np.uint8)
+        patch_match._canonize_mask_array(mask)

tests/test_patch_match.py

@@ -0,0 +1,133 @@
+import os
+
+import numpy as np
+import pytest
+from PIL import Image
+
+from patchmatch import patch_match
+
+
+@pytest.fixture
+def input_image():
+    return Image.open("./examples/images/forest_pruned.bmp")
+
+
+@pytest.fixture
+def ijmap(input_image):
+    return np.zeros((input_image.height, input_image.width, 3), dtype=np.float32)
+
+
+@pytest.fixture
+def mask(input_image):
+    return np.zeros((input_image.height, input_image.width), dtype=np.uint8)
+
+
+@pytest.fixture
+def temp_file():
+    filename = "temp.txt"
+    with open(filename, "w") as f:
+        f.write("test")
+    yield filename
+    os.remove(filename)
+
+
+def test_inpaint_custom_patch_size(input_image, mask):
+    result = patch_match.inpaint(input_image, mask, patch_size=1)
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_global_mask(input_image):
+    source = np.array(input_image)
+    source[:100, :100] = 255
+    global_mask = np.zeros_like(source[..., 0])
+    global_mask[:100, :100] = 1
+    result = Image.fromarray(
+        patch_match.inpaint(image=source, global_mask=global_mask, patch_size=3)
+    )
+    assert isinstance(result, Image.Image)
+
+
+def test_np_to_pymat():
+    npmat = np.zeros((10, 10, 3), dtype=np.uint8)
+    result = patch_match.np_to_pymat(npmat)
+    assert isinstance(result, patch_match.CMatT)
+
+
+def test_pymat_to_np():
+    pymat = patch_match.CMatT(None, patch_match.CShapeT(10, 10, 3), 0)
+    npmat = np.zeros((10, 10, 3), dtype=np.uint8)
+    pymat.data_ptr = npmat.ctypes.data
+    result = patch_match.pymat_to_np(pymat)
+    assert isinstance(result, np.ndarray)
+    assert result.shape == npmat.shape
+    assert result.dtype == npmat.dtype
+    assert np.allclose(result, npmat)
+
+
+def test_canonize_mask_array_2d_uint8():
+    mask = np.zeros((10, 10), dtype=np.uint8)
+    result = patch_match._canonize_mask_array(mask)
+    assert isinstance(result, np.ndarray)
+    assert result.ndim == 3
+    assert result.shape[2] == 1
+    assert result.dtype == np.uint8
+
+
+def test_canonize_mask_array_3d_uint8():
+    mask = np.zeros((10, 10, 1), dtype=np.uint8)
+    result = patch_match._canonize_mask_array(mask)
+    assert isinstance(result, np.ndarray)
+    assert result.ndim == 3
+    assert result.shape[2] == 1
+    assert result.dtype == np.uint8
+
+
+def test_canonize_mask_array_pil_image():
+    mask = Image.new("L", (10, 10))
+    result = patch_match._canonize_mask_array(mask)
+    assert isinstance(result, np.ndarray)
+    assert result.ndim == 3
+    assert result.shape[2] == 1
+    assert result.dtype == np.uint8
+
+
+def test_canonize_mask_array_invalid_input():
+    with pytest.raises(AssertionError):
+        mask = np.zeros((10, 10, 3), dtype=np.uint8)
+        patch_match._canonize_mask_array(mask)
+
+
+def test_inpaint_regularity(input_image, ijmap):
+    result = patch_match.inpaint_regularity(input_image, None, ijmap, patch_size=3)
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_patch_size(input_image, ijmap):
+    result = patch_match.inpaint_regularity(input_image, None, ijmap, patch_size=5)
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_guide_weight(input_image, ijmap):
+    result = patch_match.inpaint_regularity(
+        input_image, None, ijmap, patch_size=3, guide_weight=0.5
+    )
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_global_mask(input_image, ijmap, mask):
+    global_mask = mask
+    result = patch_match.inpaint_regularity(
+        input_image, None, ijmap, global_mask=global_mask, patch_size=3
+    )
+    assert isinstance(result, np.ndarray)
+
+
+def test_inpaint_regularity_custom_mask(input_image, ijmap, mask):
+    result = patch_match.inpaint_regularity(input_image, mask, ijmap, patch_size=3)
+    assert isinstance(result, np.ndarray)
+
+
+def test_download_url_to_file(temp_file):
+    url = "https://www.google.com"
+    patch_match.download_url_to_file(url, temp_file)
+    assert os.path.exists(temp_file)