from typing import List import pytest import ezdxf import math from math import isclose from ezdxf.math import Vec3, global_bspline_interpolation, close_vectors, cad_fit_point_interpolation from ezdxf.math.parametrize import ( uniform_t_vector, distance_t_vector, centripetal_t_vector, arc_t_vector, arc_distances, estimate_tangents, ) from ezdxf.math.bspline import ( knots_from_parametrization, required_knot_values, averaged_knots_unconstrained, natural_knots_constrained, averaged_knots_constrained, natural_knots_unconstrained, double_knots, create_t_vector, ) POINTS1 = Vec3.list([(1, 1), (2, 4), (4, 1), (7, 6)]) POINTS2 = Vec3.list([(1, 1), (2, 4), (4, 1), (7, 6), (5, 8), (3, 3), (1, 7)]) @pytest.fixture(params=[POINTS1, POINTS2]) def fit_points(request): return request.param def test_uniform_t_array(fit_points): t_vector = list(uniform_t_vector(len(fit_points))) assert len(t_vector) == len(fit_points) assert t_vector[0] == 0.0 assert t_vector[-1] == 1.0 for t1, t2 in zip(t_vector, t_vector[1:]): assert t1 <= t2 def test_chord_length_t_array(fit_points): t_vector = list(distance_t_vector(fit_points)) assert len(t_vector) == len(fit_points) assert t_vector[0] == 0.0 assert t_vector[-1] == pytest.approx(1.0) for t1, t2 in zip(t_vector, t_vector[1:]): assert t1 <= t2 def test_centripetal_length_t_array(fit_points): t_vector = list(centripetal_t_vector(fit_points)) assert len(t_vector) == len(fit_points) assert t_vector[0] == 0.0 assert t_vector[-1] == pytest.approx(1.0) for t1, t2 in zip(t_vector, t_vector[1:]): assert t1 <= t2 def test_arc_distances(): p = Vec3.list([(0, 0), (2, 2), (4, 0), (6, -2), (8, 0)]) # p[1]..p[3] are a straight line, radius calculation fails and # a straight line from p[1] to p[2] is used as replacement # for the second arc radius = 2.0 arc_length = math.pi * 0.5 * radius diagonal = math.sqrt(2.0) * radius distances = list(arc_distances(p)) assert len(distances) == 4 assert isclose(distances[0], arc_length) assert isclose(distances[1], diagonal) # replacement for arc assert isclose(distances[2], arc_length) assert isclose(distances[3], arc_length) def test_arc_length_t_array(fit_points): t_vector = list(arc_t_vector(fit_points)) assert len(t_vector) == len(fit_points) assert t_vector[0] == 0.0 assert t_vector[-1] == pytest.approx(1.0) for t1, t2 in zip(t_vector, t_vector[1:]): assert t1 <= t2 def test_invalid_order_count_combination(): count = 4 order = 5 with pytest.raises(ezdxf.DXFValueError): required_knot_values(count, order) with pytest.raises(ezdxf.DXFValueError): list(knots_from_parametrization(n=count - 1, p=order - 1, t=[])) def check_knots(count: int, order: int, knots: List[float]): assert len(knots) == required_knot_values(count, order) assert len(set(knots[:order])) == 1, "first order elements have to be equal" assert len(set(knots[-order:])) == 1, "last order elements have to be equal" for k1, k2 in zip(knots, knots[1:]): assert k1 <= k2 @pytest.mark.parametrize("p", (2, 3, 4)) @pytest.mark.parametrize("method", ("average", "natural")) def test_knot_generation(p, method): fit_points = Vec3.list( [ (0, 0), (0, 10), (10, 10), (20, 10), (20, 0), (30, 0), (30, 10), (40, 10), (40, 0), ] ) count = len(fit_points) n = count - 1 order = p + 1 t_vector = distance_t_vector(fit_points) knots = list(knots_from_parametrization(n, p, t_vector, method)) check_knots(n + 1, p + 1, knots) @pytest.fixture def fit_points_2(): return Vec3.list( [ (0, 0), (0, 10), (10, 10), (20, 10), (20, 0), (30, 0), (30, 10), (40, 10), (40, 0), ] ) @pytest.mark.parametrize("p", (2, 3, 4, 5)) def test_unconstrained_averaged_knots(p, fit_points_2): t_vector = list(distance_t_vector(fit_points_2)) n = len(fit_points_2) - 1 knots = averaged_knots_unconstrained(n, p, t_vector) check_knots(n + 1, p + 1, knots) @pytest.mark.parametrize("p", (2, 3, 4, 5)) def test_constrained_averaged_knots(p, fit_points_2): t_vector = list(distance_t_vector(fit_points_2)) n = len(fit_points_2) - 1 # add 2 knots for tangents knots = averaged_knots_constrained(n + 2, p, t_vector) check_knots(n + 3, p + 1, knots) @pytest.mark.parametrize("p", (2, 3, 4, 5)) def test_unconstrained_natural_knots(p, fit_points_2): t_vector = list(distance_t_vector(fit_points_2)) n = len(fit_points_2) - 1 # add 2 knots for tangents knots = natural_knots_unconstrained(n, p, t_vector) check_knots(n + 1, p + 1, knots) @pytest.mark.parametrize("p", (2, 3, 4, 5)) def test_constrained_natural_knots(p, fit_points_2): t_vector = list(distance_t_vector(fit_points_2)) n = len(fit_points_2) - 1 # add 2 knots for tangents knots = natural_knots_constrained(n + 2, p, t_vector) check_knots(n + 3, p + 1, knots) @pytest.mark.parametrize("p", (2, 3, 4, 5)) def test_double_knots(p, fit_points_2): t_vector = list(distance_t_vector(fit_points_2)) n = len(fit_points_2) - 1 # create knots for each control point and 1st derivative knots = double_knots(n, p, t_vector) check_knots((n + 1) * 2, p + 1, knots) def test_bspline_interpolation(fit_points): spline = global_bspline_interpolation(fit_points, degree=3, method="chord") assert len(spline.control_points) == len(fit_points) t_array = list(create_t_vector(fit_points, "chord")) assert t_array[0] == 0.0 assert t_array[-1] == pytest.approx(1.0) assert len(t_array) == len(fit_points) t_points = [spline.point(t) for t in t_array] assert close_vectors(t_points, fit_points) @pytest.mark.parametrize("method", ["distance", "centripetal", "arc"]) def test_create_t_vectors_for_identical_points(method): assert len(list(create_t_vector(Vec3.list([(0, 0), (0, 0)]), method))) == 0 def test_bspline_interpolation_first_derivatives(fit_points): tangents = estimate_tangents(fit_points) spline = global_bspline_interpolation( fit_points, degree=3, tangents=tangents ) assert len(spline.control_points) == 2 * len(fit_points) expected = [ (0.0, 0.0), (0.010310831479728222, 0.32375901937484741), (0.030277278274297714, 0.6140977144241333), (0.059526983648538589, 0.87245100736618042), (0.097687594592571259, 1.1002539396286011), (0.14438676834106445, 1.2989413738250732), (0.19925214350223541, 1.469948410987854), (0.26191136240959167, 1.6147100925445557), (0.33199205994606018, 1.7346612215042114), (0.40912193059921265, 1.8312369585037231), (0.49292856454849243, 1.9058722257614136), (0.58303964138031006, 1.9600019454956055), (0.67908281087875366, 1.9950611591339111), (0.78068572282791138, 2.0124847888946533), (0.88747602701187134, 2.0137078762054443), (0.9990813136100769, 2.0001654624938965), (1.1151292324066162, 1.973292350769043), (1.2352476119995117, 1.9345237016677856), (1.3590638637542725, 1.8852944374084473), (1.4862056970596313, 1.8270395994186401), (1.6163008213043213, 1.761194109916687), (1.7489768266677856, 1.6891928911209106), (1.8838614225387573, 1.6124709844589233), (2.0205821990966797, 1.532463550567627), (2.1587669849395752, 1.4506052732467651), (2.2980430126190186, 1.3683313131332397), (2.4380383491516113, 1.2870765924453735), (2.5783803462982178, 1.2082761526107788), (2.7186968326568604, 1.1333650350570679), (2.8586153984069824, 1.0637780427932739), (2.9977638721466064, 1.0009502172470093), (3.1357693672180176, 0.94631654024124146), (3.2722601890563965, 0.90131211280822754), (3.4068636894226074, 0.86737185716629028), (3.5392072200775146, 0.84593069553375244), (3.6689188480377197, 0.83842372894287109), (3.795626163482666, 0.84628582000732422), (3.9189567565917969, 0.87095201015472412), (4.0385379791259766, 0.91385728120803833), (4.1539978981018066, 0.97643661499023438), (4.2649641036987305, 1.0601249933242798), (4.3710641860961914, 1.1663573980331421), (4.4719257354736328, 1.2965688705444336), (4.567176342010498, 1.4521942138671875), (4.6564435958862305, 1.6346687078475952), (4.7393555641174316, 1.8454270362854004), (4.8155393600463867, 2.0859043598175049), (4.8846230506896973, 2.3575356006622314), (4.9462337493896484, 2.6617558002471924), (5.0, 3.0), ] def test_check_values(): test_points = [(0.0, 0.0), (1.0, 2.0), (3.0, 1.0), (5.0, 3.0)] spline = global_bspline_interpolation( test_points, degree=3, method="distance" ) result = list(spline.approximate(49)) assert len(result) == 50 for p1, p2 in zip(result, expected): assert isclose(p1[0], p2[0], abs_tol=1e-6) assert isclose(p1[1], p2[1], abs_tol=1e-6) def test_cad_fit_point_interpolation_for_2_points(): points = Vec3.list([(0, 0), (0, 10)]) control_points, knots = cad_fit_point_interpolation(points) assert control_points[0].isclose(points[0]) assert control_points[-1].isclose(points[-1]) assert len(control_points) == 4 assert len(knots) == required_knot_values(4, 4) == 8 def test_cad_fit_point_interpolation_for_5_points(): points = Vec3.list([(0, 0), (0, 10), (10, 10), (20, 10), (20, 0)]) control_points, knots = cad_fit_point_interpolation(points) assert len(control_points) == 7 assert len(knots) == required_knot_values(7, 4) == 11 # Checked visually by BricsCAD 2022 and TrueView 2022: # See function check_visually_fit_points_to_cad_cv() in script # exploration/spline_end_tangent_estimation.py assert control_points[0].isclose(points[0]) assert control_points[1].isclose(Vec3(-0.8333333333333334, 4.285714285714286, -0.0)) assert control_points[2].isclose(Vec3(-2.5, 12.857142857142858, 0.0)) assert control_points[3].isclose(Vec3(10.0, 8.571428571428573, 0.0)) assert control_points[4].isclose(Vec3(22.5, 12.857142857142858, 0.0)) assert control_points[5].isclose(Vec3(20.833333333333332, 4.2857142857142865, -0.0)) assert control_points[6].isclose(points[-1])