replace opencv-python with scikit-image for snowball detection

changes/138.apichange.rst

@@ -0,0 +1 @@
+replace usage of ``opencv-python`` with analagous functionality from ``scikit-image``

pyproject.toml

@@ -18,7 +18,6 @@ dependencies = [
     "scipy >=1.7.2",
     "scikit-image>=0.19",
     "numpy >=1.21.2",
-    "opencv-python-headless >=4.6.0.66",
     "asdf >=2.15.0",
     "gwcs >= 0.18.1",
     "tweakwcs >=0.8.8",

src/stcal/jump/circle.py

@@ -0,0 +1,207 @@
+import random
+from typing import Optional, Union
+
+import numpy
+
+RELATIVE_TOLERANCE = 1 + 1e-14
+
+
+class Circle:
+    def __init__(self, center: tuple[float, float], radius: float):
+        self.center = numpy.array(center)
+        self.radius = radius
+
+    @classmethod
+    def from_points(cls, points: list[tuple[float, float]]) -> 'Circle':
+        """
+        Returns the smallest circle that encloses all the given points.
+        from https://www.nayuki.io/page/smallest-enclosing-circle
+
+        If 0 points are given, `None` is returned.
+        If 1 point is given, a circle of radius 0 is returned.
+        If 2 points are given, a circle with diameter between them is returned.
+        If 3 or more points are given, uses the algorithm described in this PDF:
+        https://www.cise.ufl.edu/~sitharam/COURSES/CG/kreveldnbhd.pdf
+
+        :param points: A sequence of pairs of floats or ints, e.g. [(0,5), (3.1,-2.7)].
+        :return: the smallest circle that encloses all points, to within the relative tolerance defined by the class
+        """
+
+        if len(points) == 0:
+            return None
+        elif len(points) == 1:
+            return cls(center=points[0], radius=0.0)
+        elif len(points) == 2:
+            points = numpy.array(points)
+            center = numpy.mean(points, axis=0)
+            radius = numpy.max(numpy.hypot(*(center - points).T))
+
+            return cls(center=center, radius=radius)
+        else:
+            # Convert to float and randomize order
+            shuffled_points = [(float(point[0]), float(point[1])) for point in points]
+            random.shuffle(shuffled_points)
+
+            # Progressively add points to circle or recompute circle
+            circle = None
+            for (index, point) in enumerate(shuffled_points):
+                if circle is None or point not in circle:
+                    circle = _expand_circle_from_one_point(point, shuffled_points[:index + 1])
+
+            return circle
+
+    def __getitem__(self, index: int) -> Union[tuple, float]:
+        if index == 0:
+            return tuple(self.center)
+        elif index == 1:
+            return self.radius
+        else:
+            raise IndexError(f'{self.__class__.__name__} index out of range')
+
+    def __add__(self, delta: tuple[float, float]) -> 'Circle':
+        if isinstance(delta, float):
+            delta = (delta, delta)
+        return self.__class__(center=self.center + numpy.array(delta), radius=self.radius)
+
+    def __mul__(self, factor: float) -> 'Circle':
+        return self.__class__(center=self.center, radius=self.radius + factor)
+
+    def __contains__(self, point: tuple[float, float]) -> bool:
+        return numpy.hypot(*(numpy.array(point) - self.center)) <= self.radius * RELATIVE_TOLERANCE
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Circle):
+            return False
+        return numpy.all(self.center == other.center) and self.radius == other.radius
+
+    def almost_equals(self, other: 'Circle', delta: Optional[float] = None) -> bool:
+        if delta is None:
+            delta = RELATIVE_TOLERANCE
+        return numpy.allclose(self.center, other.center, atol=delta) and \
+            numpy.allclose(self.radius, other.radius, atol=delta)
+
+    def __repr__(self) -> str:
+        return f'{self.__class__.__name__}({self.center}, {self.radius})'
+
+
+def _expand_circle_from_one_point(
+        known_boundary_point: tuple[float, float],
+        points: list[tuple[float, float]],
+) -> Circle:
+    """
+    iteratively expand a circle from one known boundary point to enclose the given set of points
+    from https://www.nayuki.io/page/smallest-enclosing-circle
+    """
+
+    circle = Circle(known_boundary_point, 0.0)
+    for point_index, point in enumerate(points):
+        if point not in circle:
+            if circle.radius == 0.0:
+                circle = Circle.from_points([known_boundary_point, point])
+            else:
+                circle = _expand_circle_from_two_points(known_boundary_point, point, points[: point_index + 1])
+    return circle
+
+
+def _expand_circle_from_two_points(
+        known_boundary_point_a: tuple[float, float],
+        known_boundary_point_b: tuple[float, float],
+        points: list[tuple[float, float]],
+) -> Circle:
+    """
+    iteratively expand a circle from two known boundary points to enclose the given set of points
+    from https://www.nayuki.io/page/smallest-enclosing-circle
+    """
+
+    known_boundary_points = numpy.array([known_boundary_point_a, known_boundary_point_b])
+
+    circle = Circle.from_points(known_boundary_points)
+    left = None
+    right = None
+
+    # For each point not in the two-point circle
+    for point in points:
+        if circle is None or point in circle:
+            continue
+
+        # Form a circumcircle and classify it on left or right side
+        circumcircle_cross_product = _triangle_cross_product((*known_boundary_points, point))
+        circumcircle = circumcircle_from_points(known_boundary_point_a, known_boundary_point_b, point)
+
+        if circumcircle is None:
+            continue
+
+        circumcenter_cross_product = _triangle_cross_product((*known_boundary_points, circumcircle.center))
+
+        if circumcircle_cross_product > 0.0 and \
+                (
+                        left is None or
+                        circumcenter_cross_product > _triangle_cross_product((*known_boundary_points, left.center))
+                ):
+            left = circumcircle
+        elif circumcircle_cross_product < 0.0 and \
+                (
+                        right is None or
+                        circumcenter_cross_product < _triangle_cross_product((*known_boundary_points, right.center))
+                ):
+            right = circumcircle
+
+    # Select which circle to return
+    if left is None and right is None:
+        return circle
+    elif left is None:
+        return right if right is not None else Circle(center=(0.0, 0.0), radius=0.0)
+    elif right is None:
+        return left
+    else:
+        return left if (left.radius <= right.radius) else right
+
+
+def circumcircle_from_points(
+        a: tuple[float, float],
+        b: tuple[float, float],
+        c: tuple[float, float],
+) -> Circle | None:
+    """
+    build a circumcircle from three points, using the algorithm from https://en.wikipedia.org/wiki/Circumscribed_circle
+    from https://www.nayuki.io/page/smallest-enclosing-circle
+    """
+
+    points = numpy.array([a, b, c])
+    incenter = ((numpy.min(points, axis=0) + numpy.max(points, axis=0)) / 2)
+
+    relative_points = points - incenter
+    a, b, c = relative_points
+
+    intermediate = 2 * (a[0] * (b[1] - c[1]) + b[0] * (c[1] - a[1]) + c[0] * (a[1] - b[1]))
+    if intermediate == 0:
+        return None
+
+    relative_circumcenter = numpy.array([
+        (a[0] ** 2 + a[1] ** 2) * (b[1] - c[1]) +
+        (b[0] ** 2 + b[1] ** 2) * (c[1] - a[1]) +
+        (c[0] ** 2 + c[1] ** 2) * (a[1] - b[1]),
+        (a[0] ** 2 + a[1] ** 2) * (c[0] - b[0]) +
+        (b[0] ** 2 + b[1] ** 2) * (a[0] - c[0]) +
+        (c[0] ** 2 + c[1] ** 2) * (b[0] - a[0]),
+    ]) / intermediate
+
+    return Circle(
+        center=relative_circumcenter + incenter,
+        radius=numpy.max(numpy.hypot(*(relative_circumcenter - relative_points).T)),
+    )
+
+
+def _triangle_cross_product(triangle: tuple[tuple[float, float], tuple[float, float], tuple[float, float]]) -> float:
+    """
+    calculates twice the signed area of the provided triangle
+    from https://www.nayuki.io/page/smallest-enclosing-circle
+
+    :param triangle: three points defining a triangle
+    :return: twice the signed area of triangle
+    """
+
+    return (triangle[1][0] - triangle[0][0]) \
+        * (triangle[2][1] - triangle[0][1]) \
+        - (triangle[1][1] - triangle[0][1]) \
+        * (triangle[2][0] - triangle[0][0])