1 | #!/usr/bin/env python |
---|
2 | |
---|
3 | import unittest |
---|
4 | |
---|
5 | from Matrix import * |
---|
6 | from Numeric import dot, allclose, array, transpose, arange, ones, Float |
---|
7 | from cg_solve import * |
---|
8 | from scipy import sparse, mat |
---|
9 | |
---|
10 | |
---|
11 | class TestCase(unittest.TestCase): |
---|
12 | |
---|
13 | ## def test_solve(self): |
---|
14 | ## """Small Matrix""" |
---|
15 | |
---|
16 | ## A = [[2.0, -1.0, 0.0, 0.0 ], |
---|
17 | ## [-1.0, 2.0, -1.0, 0.0], |
---|
18 | ## [0.0, -1.0, 2.0, -1.0], |
---|
19 | ## [0.0, 0.0, -1.0, 2.0]] |
---|
20 | |
---|
21 | ## A = mat(A) |
---|
22 | |
---|
23 | ## xe = asarray([0.0, 1.0, 2.0, 3.0]) |
---|
24 | ## b = A*xe |
---|
25 | ## x = asarray([0.0, 0.0, 0.0, 0.0]) |
---|
26 | |
---|
27 | ## print A |
---|
28 | ## print xe |
---|
29 | ## print b |
---|
30 | ## print b.shape |
---|
31 | |
---|
32 | ## x = conjugate_gradient(A,b,x,iprint=1) |
---|
33 | |
---|
34 | ## assert allclose(x,xe) |
---|
35 | |
---|
36 | def test_sparse_solve(self): |
---|
37 | """Small Sparse Matrix""" |
---|
38 | |
---|
39 | A = [[2.0, -1.0, 0.0, 0.0 ], |
---|
40 | [-1.0, 2.0, -1.0, 0.0], |
---|
41 | [0.0, -1.0, 2.0, -1.0], |
---|
42 | [0.0,0.0, -1.0, 2.0]] |
---|
43 | |
---|
44 | A = sparse.dok_matrix(A) |
---|
45 | |
---|
46 | xe = [0.0, 1.0, 2.0, 3.0] |
---|
47 | b = A*xe |
---|
48 | x = [0.0, 0.0, 0.0, 0.0] |
---|
49 | |
---|
50 | x = conjugate_gradient(A,b,x,iprint=0) |
---|
51 | |
---|
52 | assert allclose(x,xe) |
---|
53 | |
---|
54 | |
---|
55 | def test_solve_large(self): |
---|
56 | """Standard 1d laplacian """ |
---|
57 | |
---|
58 | A = sparse.dok_matrix() |
---|
59 | |
---|
60 | n = 50 |
---|
61 | for i in arange(0,n): |
---|
62 | A[i,i] = 1.0 |
---|
63 | if i > 0 : |
---|
64 | A[i,i-1] = -0.5 |
---|
65 | if i < n-1 : |
---|
66 | A[i,i+1] = -0.5 |
---|
67 | |
---|
68 | xe = ones( (n,), Float) |
---|
69 | |
---|
70 | b = A*xe |
---|
71 | x = conjugate_gradient(A,b,b,tol=1.0e-5,iprint=0) |
---|
72 | |
---|
73 | assert allclose(x,xe) |
---|
74 | |
---|
75 | def test_solve_large_2d(self): |
---|
76 | """Standard 2d laplacian""" |
---|
77 | |
---|
78 | A = sparse.dok_matrix() |
---|
79 | |
---|
80 | n = 20 |
---|
81 | m = 10 |
---|
82 | for i in arange(0,n): |
---|
83 | for j in arange(0,m): |
---|
84 | I = j+m*i |
---|
85 | A[I,I] = 4.0 |
---|
86 | if i > 0 : |
---|
87 | A[I,I-m] = -1.0 |
---|
88 | if i < n-1 : |
---|
89 | A[I,I+m] = -1.0 |
---|
90 | if j > 0 : |
---|
91 | A[I,I-1] = -1.0 |
---|
92 | if j < m-1 : |
---|
93 | A[I,I+1] = -1.0 |
---|
94 | |
---|
95 | xe = ones( (n*m,), Float) |
---|
96 | |
---|
97 | b = A*xe |
---|
98 | x = conjugate_gradient(A,b,b,iprint=0) |
---|
99 | |
---|
100 | assert allclose(x,xe) |
---|
101 | |
---|
102 | |
---|
103 | def test_vector_shape_error(self): |
---|
104 | """Raise VectorShapeError""" |
---|
105 | |
---|
106 | A = [[2.0, -1.0, 0.0, 0.0 ], |
---|
107 | [-1.0, 2.0, -1.0, 0.0], |
---|
108 | [0.0, -1.0, 2.0, -1.0], |
---|
109 | [0.0,0.0, -1.0, 2.0]] |
---|
110 | |
---|
111 | A = sparse.dok_matrix(A) |
---|
112 | |
---|
113 | xe = [[0.0,2.0], [1.0,3.0], [2.0,4.0], [3.0,2.0]] |
---|
114 | b = A*xe |
---|
115 | x = [0.0, 0.0, 0.0, 0.0] |
---|
116 | |
---|
117 | try: |
---|
118 | x = conjugate_gradient(A,b,b,iprint=0) |
---|
119 | except: |
---|
120 | pass |
---|
121 | else: |
---|
122 | msg = 'Should have raised exception' |
---|
123 | raise msg |
---|
124 | |
---|
125 | |
---|
126 | #------------------------------------------------------------- |
---|
127 | if __name__ == "__main__": |
---|
128 | suite = unittest.makeSuite(TestCase,'test') |
---|
129 | runner = unittest.TextTestRunner() #(verbosity=2) |
---|
130 | runner.run(suite) |
---|
131 | |
---|
132 | |
---|
133 | |
---|
134 | |
---|
135 | |
---|
136 | |
---|