(MAINT)(TEST) increase coverage to 90%, add boilerplate CI workflow

.github/workflows/ci.yml

@@ -0,0 +1,34 @@
+name: CI
+on:
+  push:
+    branches: main
+
+  pull_request:
+    branches: main
+
+jobs:
+  run:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+      with:
+        fetch-depth: ‘2’
+        submodules: recursive
+
+    - name: Setup Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: |
+            3.10
+            3.11
+
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install --upgrade nox
+
+    - name: Execute nox
+      run: |
+        nox

tests/test_imops.py

@@ -28,116 +28,126 @@ def img_and_meta(*args, side='left'):
         return img_and_meta
 
 
-class TestImageMaths:
-
-    def test_thresh(self):
-        key = jax.random.PRNGKey(0)
-        arg = jax.random.uniform(key, (10, 10, 10))
-        grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
-        op_str = 'ARGa -thr (ARGb -bin[0.5])'
-        f = grammar.compile(op_str)
-        img_out, _ = f(arg, arg)
-        img_dataobj = arg
-        img_thr = (img_dataobj > 0.5)
-        img_ref = jnp.where(img_dataobj > img_thr, img_dataobj, 0)
-        assert (img_out == img_ref).all()
-
-    def test_morphological_atoms(self):
-        grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
-
-        data = np.zeros((10, 10))
-        data[2:8, 2:8] = 1
-        data[2, 3] = 0
-        data[4, 6] = 0
-
-        op_str = 'ARG -dil[1]'
-        f = grammar.compile(op_str)
-        dil, _ = f(data)
-
-        op_str = 'ARG -ero[1]'
-        f = grammar.compile(op_str)
-        ero, _ = f(data)
-
-        op_str = 'ARG -opening[1]'
-        f = grammar.compile(op_str)
-        opened, _ = f(data)
-
-        op_str = 'ARG -closing[1]'
-        f = grammar.compile(op_str)
-        closed, _ = f(data)
-
-        op_str = 'ARG -fillholes[1|5]'
-        f = grammar.compile(op_str)
-        filled, _ = f(data)
-
-        fig, ax = plt.subplots(2, 3, figsize=(18, 12))
-        ax[0, 0].imshow(data, cmap='gray')
-        ax[0, 1].imshow(dil, cmap='gray')
-        ax[0, 2].imshow(ero, cmap='gray')
-        ax[1, 0].imshow(opened, cmap='gray')
-        ax[1, 1].imshow(closed, cmap='gray')
-        ax[1, 2].imshow(filled, cmap='gray')
-
-        results = resource_filename(
-            'gramform',
-            '_results/'
-        )
-        fig.savefig(f'{results}/imops_morphology.png', bbox_inches='tight')
-
-    def test_logical_atoms(self):
-        grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
-
-        data0 = np.zeros((10, 10))
-        data0[:, :7] = 1
-        data0[:, :1] = 0
-        data1 = np.zeros((10, 10))
-        data1[:, 3:] = 1
-        data1[:, 9:] = 0
-
-        op_str = 'ARG -neg'
-        f = grammar.compile(op_str)
-        negation0, _ = f(data0)
-        negation1, _ = f(data1)
-
-        op_str = 'ARGa -or ARGb'
-        f = grammar.compile(op_str)
-        union, _ = f(data0, data1)
-
-        op_str = 'ARGa -and ARGb'
-        f = grammar.compile(op_str)
-        intersection, _ = f(data0, data1)
-
-        fig, ax = plt.subplots(2, 3, figsize=(18, 12))
-        ax[0, 0].imshow(data0, cmap='gray')
-        ax[0, 1].imshow(data1, cmap='gray')
-        ax[0, 2].imshow(negation0, cmap='gray')
-        ax[1, 0].imshow(negation1, cmap='gray')
-        ax[1, 1].imshow(union, cmap='gray')
-        ax[1, 2].imshow(intersection, cmap='gray')
-
-        results = resource_filename(
-            'gramform',
-            '_results/'
-        )
-        fig.savefig(f'{results}/imops_logic.png', bbox_inches='tight')
-
-    def test_compositions(self):
-        grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
-
-        key = jax.random.PRNGKey(0)
-        key0, key1, key2 = jax.random.split(key, 3)
-        data0 = jax.random.uniform(key0, (10, 10))
-        data1 = jax.random.uniform(key1, (10, 10))
-        data2 = jax.random.uniform(key2, (10, 10))
-
-        op_str = (
-            r'ARGa -sub{1} -thr[0.1]{0.5} -mul ARGb '
-            r'-add{1} -uthr[-1] (ARGc -div{2,...})'
-        )
-        f = grammar.compile(op_str)
-        img_out, _ = f(data0, data1, data2)
-        img_ref = data0 - 1
-        img_ref = jnp.where(img_ref > 0.5, img_ref, 0.1)
-        img_ref = img_ref * data1 + 1
-        img_ref = jnp.where(img_ref < 2 / data2, img_ref, -1)
-        assert (img_out == img_ref).all()
+def test_thresh():
+    key = jax.random.PRNGKey(0)
+    arg = jax.random.uniform(key, (10, 10, 10))
+    grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
+    op_str = 'ARGa -thr (ARGb -bin[0.5])'
+    f = grammar.compile(op_str)
+    img_out, _ = f(arg, arg)
+    img_dataobj = arg
+    img_thr = (img_dataobj > 0.5)
+    img_ref = jnp.where(img_dataobj > img_thr, img_dataobj, 0)
+    assert (img_out == img_ref).all()
+
+
+def test_morphological_atoms():
+    grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
+
+    data = np.zeros((10, 10))
+    data[2:8, 2:8] = 1
+    data[2, 3] = 0
+    data[4, 6] = 0
+
+    op_str = 'ARG -dil[1]'
+    f = grammar.compile(op_str)
+    dil, _ = f(data)
+
+    op_str = 'ARG -ero[1]'
+    f = grammar.compile(op_str)
+    ero, _ = f(data)
+
+    op_str = 'ARG -opening[1]'
+    f = grammar.compile(op_str)
+    opened, _ = f(data)
+
+    op_str = 'ARG -closing[1]'
+    f = grammar.compile(op_str)
+    closed, _ = f(data)
+
+    op_str = 'ARG -fillholes[1|5]'
+    f = grammar.compile(op_str)
+    filled, _ = f(data)
+
+    fig, ax = plt.subplots(2, 3, figsize=(18, 12))
+    ax[0, 0].imshow(data, cmap='gray')
+    ax[0, 1].imshow(dil, cmap='gray')
+    ax[0, 2].imshow(ero, cmap='gray')
+    ax[1, 0].imshow(opened, cmap='gray')
+    ax[1, 1].imshow(closed, cmap='gray')
+    ax[1, 2].imshow(filled, cmap='gray')
+
+    results = resource_filename(
+        'gramform',
+        '_results/'
+    )
+    fig.savefig(f'{results}/imops_morphology.png', bbox_inches='tight')
+
+
+def test_logical_atoms():
+    grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
+
+    data0 = np.zeros((10, 10))
+    data0[:, :7] = 1
+    data0[:, :1] = 0
+    data1 = np.zeros((10, 10))
+    data1[:, 3:] = 1
+    data1[:, 9:] = 0
+
+    op_str = 'ARG -neg'
+    f = grammar.compile(op_str)
+    negation0, _ = f(data0)
+    negation1, _ = f(data1)
+
+    op_str = 'ARGa -or ARGb'
+    f = grammar.compile(op_str)
+    union, _ = f(data0, data1)
+
+    op_str = 'ARGa -and ARGb'
+    f = grammar.compile(op_str)
+    intersection, _ = f(data0, data1)
+
+    fig, ax = plt.subplots(2, 3, figsize=(18, 12))
+    ax[0, 0].imshow(data0, cmap='gray')
+    ax[0, 1].imshow(data1, cmap='gray')
+    ax[0, 2].imshow(negation0, cmap='gray')
+    ax[1, 0].imshow(negation1, cmap='gray')
+    ax[1, 1].imshow(union, cmap='gray')
+    ax[1, 2].imshow(intersection, cmap='gray')
+
+    results = resource_filename(
+        'gramform',
+        '_results/'
+    )
+    fig.savefig(f'{results}/imops_logic.png', bbox_inches='tight')
+
+
+def test_compositions():
+    grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
+
+    key = jax.random.PRNGKey(0)
+    key0, key1, key2 = jax.random.split(key, 3)
+    data0 = jax.random.uniform(key0, (10, 10))
+    data1 = jax.random.uniform(key1, (10, 10))
+    data2 = jax.random.uniform(key2, (10, 10))
+
+    op_str = (
+        r'ARGa -sub{1} -thr[0.1]{0.5} -mul ARGb '
+        r'-add{1} -uthr[-1] (ARGc -div{2,...})'
+    )
+    f = grammar.compile(op_str)
+    img_out, _ = f(data0, data1, data2)
+    img_ref = data0 - 1
+    img_ref = jnp.where(img_ref > 0.5, img_ref, 0.1)
+    img_ref = img_ref * data1 + 1
+    img_ref = jnp.where(img_ref < 2 / data2, img_ref, -1)
+    assert (img_out == img_ref).all()
+
+
+def test_remainder():
+    grammar = ImageMathsGrammar(default_interpreter=TensorLeafInterpreter())
+    op_str = 'ARGa -mod ARGb -mod 3'
+    f = grammar.compile(op_str)
+    img_out, _ = f(np.full((10, 10), 17), np.full((10, 10), 13))
+    assert img_out.shape == (10, 10)
+    assert (img_out == 1).all()