• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2008-2009 Gael Guennebaud <gael.guennebaud@inria.fr>
5 //
6 // This Source Code Form is subject to the terms of the Mozilla
7 // Public License v. 2.0. If a copy of the MPL was not distributed
8 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
9 
10 #include "main.h"
11 
array_for_matrix(const MatrixType & m)12 template<typename MatrixType> void array_for_matrix(const MatrixType& m)
13 {
14   typedef typename MatrixType::Index Index;
15   typedef typename MatrixType::Scalar Scalar;
16   typedef Matrix<Scalar, MatrixType::RowsAtCompileTime, 1> ColVectorType;
17   typedef Matrix<Scalar, 1, MatrixType::ColsAtCompileTime> RowVectorType;
18 
19   Index rows = m.rows();
20   Index cols = m.cols();
21 
22   MatrixType m1 = MatrixType::Random(rows, cols),
23              m2 = MatrixType::Random(rows, cols),
24              m3(rows, cols);
25 
26   ColVectorType cv1 = ColVectorType::Random(rows);
27   RowVectorType rv1 = RowVectorType::Random(cols);
28 
29   Scalar  s1 = internal::random<Scalar>(),
30           s2 = internal::random<Scalar>();
31 
32   // scalar addition
33   VERIFY_IS_APPROX(m1.array() + s1, s1 + m1.array());
34   VERIFY_IS_APPROX((m1.array() + s1).matrix(), MatrixType::Constant(rows,cols,s1) + m1);
35   VERIFY_IS_APPROX(((m1*Scalar(2)).array() - s2).matrix(), (m1+m1) - MatrixType::Constant(rows,cols,s2) );
36   m3 = m1;
37   m3.array() += s2;
38   VERIFY_IS_APPROX(m3, (m1.array() + s2).matrix());
39   m3 = m1;
40   m3.array() -= s1;
41   VERIFY_IS_APPROX(m3, (m1.array() - s1).matrix());
42 
43   // reductions
44   VERIFY_IS_MUCH_SMALLER_THAN(m1.colwise().sum().sum() - m1.sum(), m1.squaredNorm());
45   VERIFY_IS_MUCH_SMALLER_THAN(m1.rowwise().sum().sum() - m1.sum(), m1.squaredNorm());
46   VERIFY_IS_MUCH_SMALLER_THAN(m1.colwise().sum() + m2.colwise().sum() - (m1+m2).colwise().sum(), (m1+m2).squaredNorm());
47   VERIFY_IS_MUCH_SMALLER_THAN(m1.rowwise().sum() - m2.rowwise().sum() - (m1-m2).rowwise().sum(), (m1-m2).squaredNorm());
48   VERIFY_IS_APPROX(m1.colwise().sum(), m1.colwise().redux(internal::scalar_sum_op<Scalar,Scalar>()));
49 
50   // vector-wise ops
51   m3 = m1;
52   VERIFY_IS_APPROX(m3.colwise() += cv1, m1.colwise() + cv1);
53   m3 = m1;
54   VERIFY_IS_APPROX(m3.colwise() -= cv1, m1.colwise() - cv1);
55   m3 = m1;
56   VERIFY_IS_APPROX(m3.rowwise() += rv1, m1.rowwise() + rv1);
57   m3 = m1;
58   VERIFY_IS_APPROX(m3.rowwise() -= rv1, m1.rowwise() - rv1);
59 
60   // empty objects
61   VERIFY_IS_APPROX(m1.block(0,0,0,cols).colwise().sum(),  RowVectorType::Zero(cols));
62   VERIFY_IS_APPROX(m1.block(0,0,rows,0).rowwise().prod(), ColVectorType::Ones(rows));
63 
64   // verify the const accessors exist
65   const Scalar& ref_m1 = m.matrix().array().coeffRef(0);
66   const Scalar& ref_m2 = m.matrix().array().coeffRef(0,0);
67   const Scalar& ref_a1 = m.array().matrix().coeffRef(0);
68   const Scalar& ref_a2 = m.array().matrix().coeffRef(0,0);
69   VERIFY(&ref_a1 == &ref_m1);
70   VERIFY(&ref_a2 == &ref_m2);
71 
72   // Check write accessors:
73   m1.array().coeffRef(0,0) = 1;
74   VERIFY_IS_APPROX(m1(0,0),Scalar(1));
75   m1.array()(0,0) = 2;
76   VERIFY_IS_APPROX(m1(0,0),Scalar(2));
77   m1.array().matrix().coeffRef(0,0) = 3;
78   VERIFY_IS_APPROX(m1(0,0),Scalar(3));
79   m1.array().matrix()(0,0) = 4;
80   VERIFY_IS_APPROX(m1(0,0),Scalar(4));
81 }
82 
comparisons(const MatrixType & m)83 template<typename MatrixType> void comparisons(const MatrixType& m)
84 {
85   using std::abs;
86   typedef typename MatrixType::Index Index;
87   typedef typename MatrixType::Scalar Scalar;
88   typedef typename NumTraits<Scalar>::Real RealScalar;
89 
90   Index rows = m.rows();
91   Index cols = m.cols();
92 
93   Index r = internal::random<Index>(0, rows-1),
94         c = internal::random<Index>(0, cols-1);
95 
96   MatrixType m1 = MatrixType::Random(rows, cols),
97              m2 = MatrixType::Random(rows, cols),
98              m3(rows, cols);
99 
100   VERIFY(((m1.array() + Scalar(1)) > m1.array()).all());
101   VERIFY(((m1.array() - Scalar(1)) < m1.array()).all());
102   if (rows*cols>1)
103   {
104     m3 = m1;
105     m3(r,c) += 1;
106     VERIFY(! (m1.array() < m3.array()).all() );
107     VERIFY(! (m1.array() > m3.array()).all() );
108   }
109 
110   // comparisons to scalar
111   VERIFY( (m1.array() != (m1(r,c)+1) ).any() );
112   VERIFY( (m1.array() > (m1(r,c)-1) ).any() );
113   VERIFY( (m1.array() < (m1(r,c)+1) ).any() );
114   VERIFY( (m1.array() == m1(r,c) ).any() );
115   VERIFY( m1.cwiseEqual(m1(r,c)).any() );
116 
117   // test Select
118   VERIFY_IS_APPROX( (m1.array()<m2.array()).select(m1,m2), m1.cwiseMin(m2) );
119   VERIFY_IS_APPROX( (m1.array()>m2.array()).select(m1,m2), m1.cwiseMax(m2) );
120   Scalar mid = (m1.cwiseAbs().minCoeff() + m1.cwiseAbs().maxCoeff())/Scalar(2);
121   for (int j=0; j<cols; ++j)
122   for (int i=0; i<rows; ++i)
123     m3(i,j) = abs(m1(i,j))<mid ? 0 : m1(i,j);
124   VERIFY_IS_APPROX( (m1.array().abs()<MatrixType::Constant(rows,cols,mid).array())
125                         .select(MatrixType::Zero(rows,cols),m1), m3);
126   // shorter versions:
127   VERIFY_IS_APPROX( (m1.array().abs()<MatrixType::Constant(rows,cols,mid).array())
128                         .select(0,m1), m3);
129   VERIFY_IS_APPROX( (m1.array().abs()>=MatrixType::Constant(rows,cols,mid).array())
130                         .select(m1,0), m3);
131   // even shorter version:
132   VERIFY_IS_APPROX( (m1.array().abs()<mid).select(0,m1), m3);
133 
134   // count
135   VERIFY(((m1.array().abs()+1)>RealScalar(0.1)).count() == rows*cols);
136 
137   // and/or
138   VERIFY( ((m1.array()<RealScalar(0)).matrix() && (m1.array()>RealScalar(0)).matrix()).count() == 0);
139   VERIFY( ((m1.array()<RealScalar(0)).matrix() || (m1.array()>=RealScalar(0)).matrix()).count() == rows*cols);
140   RealScalar a = m1.cwiseAbs().mean();
141   VERIFY( ((m1.array()<-a).matrix() || (m1.array()>a).matrix()).count() == (m1.cwiseAbs().array()>a).count());
142 
143   typedef Matrix<typename MatrixType::Index, Dynamic, 1> VectorOfIndices;
144 
145   // TODO allows colwise/rowwise for array
146   VERIFY_IS_APPROX(((m1.array().abs()+1)>RealScalar(0.1)).matrix().colwise().count(), VectorOfIndices::Constant(cols,rows).transpose());
147   VERIFY_IS_APPROX(((m1.array().abs()+1)>RealScalar(0.1)).matrix().rowwise().count(), VectorOfIndices::Constant(rows, cols));
148 }
149 
lpNorm(const VectorType & v)150 template<typename VectorType> void lpNorm(const VectorType& v)
151 {
152   using std::sqrt;
153   typedef typename VectorType::RealScalar RealScalar;
154   VectorType u = VectorType::Random(v.size());
155 
156   if(v.size()==0)
157   {
158     VERIFY_IS_APPROX(u.template lpNorm<Infinity>(), RealScalar(0));
159     VERIFY_IS_APPROX(u.template lpNorm<1>(), RealScalar(0));
160     VERIFY_IS_APPROX(u.template lpNorm<2>(), RealScalar(0));
161     VERIFY_IS_APPROX(u.template lpNorm<5>(), RealScalar(0));
162   }
163   else
164   {
165     VERIFY_IS_APPROX(u.template lpNorm<Infinity>(), u.cwiseAbs().maxCoeff());
166   }
167 
168   VERIFY_IS_APPROX(u.template lpNorm<1>(), u.cwiseAbs().sum());
169   VERIFY_IS_APPROX(u.template lpNorm<2>(), sqrt(u.array().abs().square().sum()));
170   VERIFY_IS_APPROX(numext::pow(u.template lpNorm<5>(), typename VectorType::RealScalar(5)), u.array().abs().pow(5).sum());
171 }
172 
cwise_min_max(const MatrixType & m)173 template<typename MatrixType> void cwise_min_max(const MatrixType& m)
174 {
175   typedef typename MatrixType::Index Index;
176   typedef typename MatrixType::Scalar Scalar;
177 
178   Index rows = m.rows();
179   Index cols = m.cols();
180 
181   MatrixType m1 = MatrixType::Random(rows, cols);
182 
183   // min/max with array
184   Scalar maxM1 = m1.maxCoeff();
185   Scalar minM1 = m1.minCoeff();
186 
187   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, minM1), m1.cwiseMin(MatrixType::Constant(rows,cols, minM1)));
188   VERIFY_IS_APPROX(m1, m1.cwiseMin(MatrixType::Constant(rows,cols, maxM1)));
189 
190   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, maxM1), m1.cwiseMax(MatrixType::Constant(rows,cols, maxM1)));
191   VERIFY_IS_APPROX(m1, m1.cwiseMax(MatrixType::Constant(rows,cols, minM1)));
192 
193   // min/max with scalar input
194   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, minM1), m1.cwiseMin( minM1));
195   VERIFY_IS_APPROX(m1, m1.cwiseMin(maxM1));
196   VERIFY_IS_APPROX(-m1, (-m1).cwiseMin(-minM1));
197   VERIFY_IS_APPROX(-m1.array(), ((-m1).array().min)( -minM1));
198 
199   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, maxM1), m1.cwiseMax( maxM1));
200   VERIFY_IS_APPROX(m1, m1.cwiseMax(minM1));
201   VERIFY_IS_APPROX(-m1, (-m1).cwiseMax(-maxM1));
202   VERIFY_IS_APPROX(-m1.array(), ((-m1).array().max)(-maxM1));
203 
204   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, minM1).array(), (m1.array().min)( minM1));
205   VERIFY_IS_APPROX(m1.array(), (m1.array().min)( maxM1));
206 
207   VERIFY_IS_APPROX(MatrixType::Constant(rows,cols, maxM1).array(), (m1.array().max)( maxM1));
208   VERIFY_IS_APPROX(m1.array(), (m1.array().max)( minM1));
209 
210 }
211 
resize(const MatrixTraits & t)212 template<typename MatrixTraits> void resize(const MatrixTraits& t)
213 {
214   typedef typename MatrixTraits::Index Index;
215   typedef typename MatrixTraits::Scalar Scalar;
216   typedef Matrix<Scalar,Dynamic,Dynamic> MatrixType;
217   typedef Array<Scalar,Dynamic,Dynamic> Array2DType;
218   typedef Matrix<Scalar,Dynamic,1> VectorType;
219   typedef Array<Scalar,Dynamic,1> Array1DType;
220 
221   Index rows = t.rows(), cols = t.cols();
222 
223   MatrixType m(rows,cols);
224   VectorType v(rows);
225   Array2DType a2(rows,cols);
226   Array1DType a1(rows);
227 
228   m.array().resize(rows+1,cols+1);
229   VERIFY(m.rows()==rows+1 && m.cols()==cols+1);
230   a2.matrix().resize(rows+1,cols+1);
231   VERIFY(a2.rows()==rows+1 && a2.cols()==cols+1);
232   v.array().resize(cols);
233   VERIFY(v.size()==cols);
234   a1.matrix().resize(cols);
235   VERIFY(a1.size()==cols);
236 }
237 
238 template<int>
regression_bug_654()239 void regression_bug_654()
240 {
241   ArrayXf a = RowVectorXf(3);
242   VectorXf v = Array<float,1,Dynamic>(3);
243 }
244 
245 // Check propagation of LvalueBit through Array/Matrix-Wrapper
246 template<int>
regrrssion_bug_1410()247 void regrrssion_bug_1410()
248 {
249   const Matrix4i M;
250   const Array4i A;
251   ArrayWrapper<const Matrix4i> MA = M.array();
252   MA.row(0);
253   MatrixWrapper<const Array4i> AM = A.matrix();
254   AM.row(0);
255 
256   VERIFY((internal::traits<ArrayWrapper<const Matrix4i> >::Flags&LvalueBit)==0);
257   VERIFY((internal::traits<MatrixWrapper<const Array4i> >::Flags&LvalueBit)==0);
258 
259   VERIFY((internal::traits<ArrayWrapper<Matrix4i> >::Flags&LvalueBit)==LvalueBit);
260   VERIFY((internal::traits<MatrixWrapper<Array4i> >::Flags&LvalueBit)==LvalueBit);
261 }
262 
test_array_for_matrix()263 void test_array_for_matrix()
264 {
265   for(int i = 0; i < g_repeat; i++) {
266     CALL_SUBTEST_1( array_for_matrix(Matrix<float, 1, 1>()) );
267     CALL_SUBTEST_2( array_for_matrix(Matrix2f()) );
268     CALL_SUBTEST_3( array_for_matrix(Matrix4d()) );
269     CALL_SUBTEST_4( array_for_matrix(MatrixXcf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
270     CALL_SUBTEST_5( array_for_matrix(MatrixXf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
271     CALL_SUBTEST_6( array_for_matrix(MatrixXi(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
272   }
273   for(int i = 0; i < g_repeat; i++) {
274     CALL_SUBTEST_1( comparisons(Matrix<float, 1, 1>()) );
275     CALL_SUBTEST_2( comparisons(Matrix2f()) );
276     CALL_SUBTEST_3( comparisons(Matrix4d()) );
277     CALL_SUBTEST_5( comparisons(MatrixXf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
278     CALL_SUBTEST_6( comparisons(MatrixXi(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
279   }
280   for(int i = 0; i < g_repeat; i++) {
281     CALL_SUBTEST_1( cwise_min_max(Matrix<float, 1, 1>()) );
282     CALL_SUBTEST_2( cwise_min_max(Matrix2f()) );
283     CALL_SUBTEST_3( cwise_min_max(Matrix4d()) );
284     CALL_SUBTEST_5( cwise_min_max(MatrixXf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
285     CALL_SUBTEST_6( cwise_min_max(MatrixXi(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
286   }
287   for(int i = 0; i < g_repeat; i++) {
288     CALL_SUBTEST_1( lpNorm(Matrix<float, 1, 1>()) );
289     CALL_SUBTEST_2( lpNorm(Vector2f()) );
290     CALL_SUBTEST_7( lpNorm(Vector3d()) );
291     CALL_SUBTEST_8( lpNorm(Vector4f()) );
292     CALL_SUBTEST_5( lpNorm(VectorXf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
293     CALL_SUBTEST_4( lpNorm(VectorXcf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
294   }
295   CALL_SUBTEST_5( lpNorm(VectorXf(0)) );
296   CALL_SUBTEST_4( lpNorm(VectorXcf(0)) );
297   for(int i = 0; i < g_repeat; i++) {
298     CALL_SUBTEST_4( resize(MatrixXcf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
299     CALL_SUBTEST_5( resize(MatrixXf(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
300     CALL_SUBTEST_6( resize(MatrixXi(internal::random<int>(1,EIGEN_TEST_MAX_SIZE), internal::random<int>(1,EIGEN_TEST_MAX_SIZE))) );
301   }
302   CALL_SUBTEST_6( regression_bug_654<0>() );
303   CALL_SUBTEST_6( regrrssion_bug_1410<0>() );
304 }
305