Merge pull request #49 from neuro-ml/C++

C++

.github/workflows/release.yml

@@ -42,7 +42,7 @@ jobs:
 
     steps:
       - uses: actions/checkout@v3
-      - uses: actions/setup-python@v3
+      - uses: actions/setup-python@v4
       - name: Install cibuildwheel
         run: python -m pip install cibuildwheel==2.10.0
       - name: Install gcc for mac

.github/workflows/test_build.yml

@@ -16,7 +16,7 @@ jobs:
 
     steps:
       - uses: actions/checkout@v3
-      - uses: actions/setup-python@v3
+      - uses: actions/setup-python@v4
       - name: Install cibuildwheel
         run: python -m pip install cibuildwheel==2.10.0
       - name: Install gcc for mac

_build_utils.py

@@ -13,7 +13,7 @@ class LazyImport(dict):
     # https://github.com/cython/cython/blob/6ad6ca0e9e7d030354b7fe7d7b56c3f6e6a4bc23/Cython/Compiler/ModuleNode.py#L773
     def __init__(self, module_name):
         self.module_name = module_name
-        return super().__init__(self, description=self.__doc__)
+        super().__init__(self, description=self.__doc__)
 
     # Must be hashable due to
     # https://github.com/cython/cython/blob/6ad6ca0e9e7d030354b7fe7d7b56c3f6e6a4bc23/Cython/Compiler/Main.py#L307

imops/__version__.py

@@ -1 +1 @@
-__version__ = '0.8.5'
+__version__ = '0.8.6'

imops/interp2d.py

@@ -53,12 +53,15 @@ def __init__(
         if triangles is not None:
             if not isinstance(triangles, np.ndarray):
                 raise TypeError(f'Wrong type of `triangles` argument, expected np.ndarray. Got {type(triangles)}')
-            if triangles.ndim != 2 or triangles.shape[1] != 3 or triangles.shape[0] * 3 != triangles.size:
+            if triangles.ndim != 2 or triangles.shape[1] != 3:
                 raise ValueError('Passed `triangles` argument has an incorrect shape')
 
         if not isinstance(points, np.ndarray):
             raise TypeError(f'Wrong type of `points` argument, expected np.ndarray. Got {type(points)}')
 
+        if points.ndim != 2 or points.shape[1] != 2:
+            raise ValueError('Passed `points` argument has an incorrect shape')
+
         if values is not None:
             if not isinstance(values, np.ndarray):
                 raise TypeError(f'Wrong type of `values` argument, expected np.ndarray. Got {type(values)}')
@@ -81,7 +84,7 @@ def __call__(self, points: np.ndarray, values: np.ndarray = None, fill_value: fl
         points: np.ndarray
             2-D array of data point coordinates to interpolate at
         values: np.ndarray
-            1-D array of fp32/fp64 values to use at initial points. If passed, existing values will be rewritten
+            1-D array of fp32/fp64 values to use at initial points
         fill_value: float
             value to fill past edges
 
@@ -90,22 +93,26 @@ def __call__(self, points: np.ndarray, values: np.ndarray = None, fill_value: fl
         new_values: np.ndarray
             interpolated values at given points
         """
-        self.values = values or self.values
+        if values is None:
+            values = self.values
 
-        if self.values is None:
+        if values is None:
             raise ValueError('`values` argument was never passed neither in __init__ or __call__ methods')
 
-        if not isinstance(self.values, np.ndarray):
-            raise TypeError(f'Wrong type of `values` argument, expected np.ndarray. Got {type(self.values)}')
+        if not isinstance(values, np.ndarray):
+            raise TypeError(f'Wrong type of `values` argument, expected np.ndarray. Got {type(values)}')
+
+        if values.ndim > 1:
+            raise ValueError(f'Wrong shape of `values` argument, expected ndim=1. Got shape {values.shape}')
+
+        if not isinstance(points, np.ndarray):
+            raise TypeError(f'Wrong type of `points` argument, expected np.ndarray. Got {type(points)}')
 
-        if self.values.ndim > 1:
-            raise ValueError(f'Wrong shape of `values` argument, expected ndim=1. Got shape {self.values.shape}')
+        if points.ndim != 2 or points.shape[1] != 2:
+            raise ValueError('Passed `points` argument has an incorrect shape')
 
         _, neighbors = self.kdtree.query(
             points, 1, **{'workers': self.num_threads} if python_version()[:3] != '3.6' else {}
         )
 
-        if not isinstance(points, np.ndarray):
-            raise TypeError(f'Wrong type of `points` argument, expected np.ndarray. Got {type(points)}')
-
-        return super().__call__(points, self.values, neighbors, fill_value)
+        return super().__call__(points, values, neighbors, fill_value)

tests/test_interp2d.py

@@ -41,7 +41,6 @@ def test_test_data(num_threads):
         scipy_values = griddata(x_points, x_values, int_points, method='linear', fill_value=0.0)
 
         delta_ds = np.abs(delaunay_values - scipy_values)
-        # delta_di = np.abs(delaunay_values - imops_values)
         delta_si = np.abs(scipy_values - imops_values)
 
         assert delta_ds.max() <= 1e-10 and delta_si.max() <= 5, f'Failed with big case, arr_{i}'
@@ -74,6 +73,15 @@ def test_no_values(example):
         Linear2DInterpolator(x_points)(int_points)
 
 
+def test_no_changes_in_values(example):
+    x_points, int_points = example
+    first_values = np.ones((x_points.shape[0],), dtype=float)
+    second_values = 2.0 * np.ones((x_points.shape[0],), dtype=float)
+    interpolator = Linear2DInterpolator(x_points, first_values)
+    interpolator(int_points, second_values)
+    assert np.all(interpolator.values == first_values), 'Failed with changes in self.values after __call__'
+
+
 def test_bad_values_dtype(example):
     x_points, int_points = example
 
@@ -99,3 +107,26 @@ def test_bad_values_lenght(example):
         Linear2DInterpolator(x_points, values=values)(int_points)
     with pytest.raises(ValueError):
         Linear2DInterpolator(x_points)(int_points, values=values)
+
+
+def test_bad_triangles_dtype(example):
+    x_points, _ = example
+
+    with pytest.raises(TypeError):
+        Linear2DInterpolator(x_points, triangles=[1, 2, 3])
+    with pytest.raises(ValueError):
+        Linear2DInterpolator(x_points, triangles=np.ones((3, 2)))
+
+
+def test_bad_points_dtype(example):
+    x_points, _ = example
+
+    values = np.ones((x_points.shape[0],))
+    with pytest.raises(TypeError):
+        Linear2DInterpolator(points=[1, 2, 3])
+    with pytest.raises(TypeError):
+        Linear2DInterpolator(points=x_points, values=values)(points=[1, 2, 3])
+    with pytest.raises(ValueError):
+        Linear2DInterpolator(points=np.ones((3, 3)))
+    with pytest.raises(ValueError):
+        Linear2DInterpolator(points=x_points, values=values)(points=np.ones((3, 3)))