[ecc58e72] | 1 | """ |
---|
| 2 | Test for the BaseComponent.evalDistribution |
---|
| 3 | See method documentation for more details |
---|
| 4 | """ |
---|
| 5 | import unittest |
---|
| 6 | import numpy, math |
---|
| 7 | from sans.models.SphereModel import SphereModel |
---|
| 8 | from sans.models.Cos import Cos |
---|
| 9 | |
---|
| 10 | |
---|
| 11 | class TestEvalPythonMethods(unittest.TestCase): |
---|
| 12 | """ Testing evalDistribution for pure python models """ |
---|
| 13 | |
---|
| 14 | def setUp(self): |
---|
| 15 | self.model= Cos() |
---|
| 16 | |
---|
| 17 | def test_scalar_methods(self): |
---|
| 18 | """ |
---|
| 19 | Simple test comparing the run(), runXY() and |
---|
| 20 | evalDistribution methods |
---|
| 21 | """ |
---|
| 22 | q1 = self.model.run(0.001) |
---|
| 23 | q2 = self.model.runXY(0.001) |
---|
| 24 | qlist3 = numpy.asarray([0.001, 0.002]) |
---|
| 25 | q3 = self.model.evalDistribution(qlist3) |
---|
| 26 | q4 = self.model.run(0.002) |
---|
| 27 | |
---|
| 28 | self.assertEqual(q1, q2) |
---|
| 29 | self.assertEqual(q1, q3[0]) |
---|
| 30 | self.assertEqual(q4, q3[1]) |
---|
| 31 | |
---|
| 32 | def test_XY_methods(self): |
---|
| 33 | """ |
---|
| 34 | Compare to the runXY() method for 2D models. |
---|
| 35 | |
---|
| 36 | +--------+--------+--------+ |
---|
| 37 | qy=0.009 | | | | |
---|
| 38 | +--------+--------+--------+ |
---|
| 39 | qy-0.006 | | | | |
---|
| 40 | +--------+--------+--------+ |
---|
| 41 | qy=0.003 | | | | |
---|
| 42 | +--------+--------+--------+ |
---|
| 43 | qx=0.001 0.002 0.003 |
---|
| 44 | |
---|
| 45 | """ |
---|
| 46 | # These are the expected values for all bins |
---|
| 47 | expected = numpy.zeros([3,3]) |
---|
| 48 | for i in range(3): |
---|
| 49 | for j in range(3): |
---|
| 50 | q_length = math.sqrt( (0.001*(i+1.0))*(0.001*(i+1.0)) + (0.003*(j+1.0))*(0.003*(j+1.0)) ) |
---|
| 51 | expected[i][j] = self.model.run(q_length) |
---|
| 52 | |
---|
| 53 | qx_values = [0.001, 0.002, 0.003] |
---|
| 54 | qy_values = [0.003, 0.006, 0.009] |
---|
| 55 | |
---|
| 56 | qx = numpy.asarray(qx_values) |
---|
| 57 | qy = numpy.asarray(qy_values) |
---|
[7e99f5c] | 58 | |
---|
| 59 | new_x = numpy.tile(qx, (len(qy),1)) |
---|
| 60 | new_y = numpy.tile(qy, (len(qx),1)) |
---|
| 61 | new_y = new_y.swapaxes(0,1) |
---|
| 62 | |
---|
| 63 | #iq is 1d array now (since 03-12-2010) |
---|
| 64 | qx_prime = new_x.flatten() |
---|
| 65 | qy_prime = new_y.flatten() |
---|
[ecc58e72] | 66 | |
---|
| 67 | iq = self.model.evalDistribution([qx_prime, qy_prime]) |
---|
| 68 | |
---|
| 69 | for i in range(3): |
---|
| 70 | for j in range(3): |
---|
[7e99f5c] | 71 | k = i+len(qx)*j |
---|
| 72 | self.assertAlmostEquals(iq[k], expected[i][j]) |
---|
[ecc58e72] | 73 | |
---|
| 74 | def test_rectangle_methods(self): |
---|
| 75 | """ |
---|
| 76 | Compare to the runXY() method for 2D models |
---|
| 77 | with a non-square matrix. |
---|
| 78 | TODO: Doesn't work for C models apparently |
---|
| 79 | |
---|
| 80 | +--------+--------+--------+ |
---|
| 81 | qy-0.006 | | | | |
---|
| 82 | +--------+--------+--------+ |
---|
| 83 | qy=0.003 | | | | |
---|
| 84 | +--------+--------+--------+ |
---|
| 85 | qx=0.001 0.002 0.003 |
---|
| 86 | |
---|
| 87 | """ |
---|
| 88 | # These are the expected values for all bins |
---|
| 89 | expected = numpy.zeros([3,3]) |
---|
| 90 | for i in range(3): |
---|
| 91 | for j in range(2): |
---|
| 92 | q_length = math.sqrt( (0.001*(i+1.0))*(0.001*(i+1.0)) + (0.003*(j+1.0))*(0.003*(j+1.0)) ) |
---|
| 93 | expected[i][j] = self.model.run(q_length) |
---|
| 94 | |
---|
| 95 | qx_values = [0.001, 0.002, 0.003] |
---|
| 96 | qy_values = [0.003, 0.006] |
---|
| 97 | |
---|
| 98 | qx = numpy.asarray(qx_values) |
---|
| 99 | qy = numpy.asarray(qy_values) |
---|
[7e99f5c] | 100 | |
---|
| 101 | new_x = numpy.tile(qx, (len(qy),1)) |
---|
| 102 | new_y = numpy.tile(qy, (len(qx),1)) |
---|
| 103 | new_y = new_y.swapaxes(0,1) |
---|
| 104 | |
---|
| 105 | #iq is 1d array now (since 03-12-2010) |
---|
| 106 | qx_prime = new_x.flatten() |
---|
| 107 | qy_prime = new_y.flatten() |
---|
[ecc58e72] | 108 | |
---|
| 109 | iq = self.model.evalDistribution([qx_prime, qy_prime]) |
---|
| 110 | |
---|
| 111 | for i in range(3): |
---|
| 112 | for j in range(2): |
---|
[7e99f5c] | 113 | k = i+len(qx)*j |
---|
| 114 | self.assertAlmostEquals(iq[k], expected[i][j]) |
---|
[ecc58e72] | 115 | |
---|
| 116 | |
---|
| 117 | |
---|
| 118 | class TestEvalMethods(unittest.TestCase): |
---|
| 119 | """ Testing evalDistribution for C models """ |
---|
| 120 | |
---|
| 121 | def setUp(self): |
---|
| 122 | self.model= SphereModel() |
---|
| 123 | |
---|
| 124 | def test_scalar_methods(self): |
---|
| 125 | """ |
---|
| 126 | Simple test comparing the run(), runXY() and |
---|
| 127 | evalDistribution methods |
---|
| 128 | """ |
---|
| 129 | q1 = self.model.run(0.001) |
---|
| 130 | q2 = self.model.runXY(0.001) |
---|
| 131 | q4 = self.model.run(0.002) |
---|
| 132 | qlist3 = numpy.asarray([0.001, 0.002]) |
---|
| 133 | q3 = self.model.evalDistribution(qlist3) |
---|
| 134 | |
---|
| 135 | self.assertEqual(q1, q2) |
---|
| 136 | self.assertEqual(q1, q3[0]) |
---|
| 137 | self.assertEqual(q4, q3[1]) |
---|
| 138 | |
---|
| 139 | def test_XY_methods(self): |
---|
| 140 | """ |
---|
| 141 | Compare to the runXY() method for 2D models. |
---|
| 142 | |
---|
| 143 | +--------+--------+--------+ |
---|
| 144 | qy=0.009 | | | | |
---|
| 145 | +--------+--------+--------+ |
---|
| 146 | qy-0.006 | | | | |
---|
| 147 | +--------+--------+--------+ |
---|
| 148 | qy=0.003 | | | | |
---|
| 149 | +--------+--------+--------+ |
---|
| 150 | qx=0.001 0.002 0.003 |
---|
| 151 | |
---|
| 152 | """ |
---|
| 153 | # These are the expected values for all bins |
---|
| 154 | expected = numpy.zeros([3,3]) |
---|
| 155 | for i in range(3): |
---|
| 156 | for j in range(3): |
---|
| 157 | q_length = math.sqrt( (0.001*(i+1.0))*(0.001*(i+1.0)) + (0.003*(j+1.0))*(0.003*(j+1.0)) ) |
---|
| 158 | expected[i][j] = self.model.run(q_length) |
---|
| 159 | |
---|
| 160 | qx_values = [0.001, 0.002, 0.003] |
---|
| 161 | qy_values = [0.003, 0.006, 0.009] |
---|
| 162 | |
---|
| 163 | qx = numpy.asarray(qx_values) |
---|
| 164 | qy = numpy.asarray(qy_values) |
---|
[7e99f5c] | 165 | |
---|
| 166 | new_x = numpy.tile(qx, (len(qy),1)) |
---|
| 167 | new_y = numpy.tile(qy, (len(qx),1)) |
---|
| 168 | new_y = new_y.swapaxes(0,1) |
---|
| 169 | |
---|
| 170 | #iq is 1d array now (since 03-12-2010) |
---|
| 171 | qx_prime = new_x.flatten() |
---|
| 172 | qy_prime = new_y.flatten() |
---|
[ecc58e72] | 173 | |
---|
| 174 | iq = self.model.evalDistribution([qx_prime, qy_prime]) |
---|
| 175 | |
---|
| 176 | for i in range(3): |
---|
| 177 | for j in range(3): |
---|
[7e99f5c] | 178 | # convert index into 1d array |
---|
| 179 | k = i+len(qx)*j |
---|
| 180 | self.assertAlmostEquals(iq[k], expected[i][j]) |
---|
[ecc58e72] | 181 | |
---|
| 182 | |
---|
| 183 | if __name__ == '__main__': |
---|
| 184 | unittest.main() |
---|