nullary.cpp
Go to the documentation of this file.
1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2010-2011 Jitse Niesen <jitse@maths.leeds.ac.uk>
5 // Copyright (C) 2016 Gael Guennebaud <gael.guennebaud@inria.fr>
6 //
7 // This Source Code Form is subject to the terms of the Mozilla
8 // Public License v. 2.0. If a copy of the MPL was not distributed
9 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
10 
11 #include "main.h"
12 
13 template<typename MatrixType>
15 {
16  typedef typename MatrixType::Scalar Scalar;
17  Scalar zero = static_cast<Scalar>(0);
18 
19  bool offDiagOK = true;
20  for (Index i = 0; i < A.rows(); ++i) {
21  for (Index j = i+1; j < A.cols(); ++j) {
22  offDiagOK = offDiagOK && (A(i,j) == zero);
23  }
24  }
25  for (Index i = 0; i < A.rows(); ++i) {
26  for (Index j = 0; j < (std::min)(i, A.cols()); ++j) {
27  offDiagOK = offDiagOK && (A(i,j) == zero);
28  }
29  }
30 
31  bool diagOK = (A.diagonal().array() == 1).all();
32  return offDiagOK && diagOK;
33 
34 }
35 
36 template<typename VectorType>
37 void check_extremity_accuracy(const VectorType &v, const typename VectorType::Scalar &low, const typename VectorType::Scalar &high)
38 {
39  typedef typename VectorType::Scalar Scalar;
40  typedef typename VectorType::RealScalar RealScalar;
41 
43  Index size = v.size();
44 
45  if(size<20)
46  return;
47 
48  for (int i=0; i<size; ++i)
49  {
50  if(i<5 || i>size-6)
51  {
52  Scalar ref = (low*RealScalar(size-i-1))/RealScalar(size-1) + (high*RealScalar(i))/RealScalar(size-1);
53  if(std::abs(ref)>1)
54  {
55  if(!internal::isApprox(v(i), ref, prec))
56  std::cout << v(i) << " != " << ref << " ; relative error: " << std::abs((v(i)-ref)/ref) << " ; required precision: " << prec << " ; range: " << low << "," << high << " ; i: " << i << "\n";
57  VERIFY(internal::isApprox(v(i), (low*RealScalar(size-i-1))/RealScalar(size-1) + (high*RealScalar(i))/RealScalar(size-1), prec));
58  }
59  }
60  }
61 }
62 
63 template<typename VectorType>
65 {
66  typedef typename VectorType::Scalar Scalar;
67  typedef typename VectorType::RealScalar RealScalar;
68 
69  const Index size = base.size();
70 
71  Scalar high = internal::random<Scalar>(-500,500);
72  Scalar low = (size == 1 ? high : internal::random<Scalar>(-500,500));
73  if (low>high) std::swap(low,high);
74 
75  // check low==high
76  if(internal::random<float>(0.f,1.f)<0.05f)
77  low = high;
78  // check abs(low) >> abs(high)
79  else if(size>2 && std::numeric_limits<RealScalar>::max_exponent10>0 && internal::random<float>(0.f,1.f)<0.1f)
80  low = -internal::random<Scalar>(1,2) * RealScalar(std::pow(RealScalar(10),std::numeric_limits<RealScalar>::max_exponent10/2));
81 
82  const Scalar step = ((size == 1) ? 1 : (high-low)/(size-1));
83 
84  // check whether the result yields what we expect it to do
85  VectorType m(base);
86  m.setLinSpaced(size,low,high);
87 
89  {
90  VectorType n(size);
91  for (int i=0; i<size; ++i)
92  n(i) = low+i*step;
93  VERIFY_IS_APPROX(m,n);
94 
95  CALL_SUBTEST( check_extremity_accuracy(m, low, high) );
96  }
97 
98  if((!NumTraits<Scalar>::IsInteger) || ((high-low)>=size && (Index(high-low)%(size-1))==0) || (Index(high-low+1)<size && (size%Index(high-low+1))==0))
99  {
100  VectorType n(size);
101  if((!NumTraits<Scalar>::IsInteger) || (high-low>=size))
102  for (int i=0; i<size; ++i)
103  n(i) = size==1 ? low : (low + ((high-low)*Scalar(i))/(size-1));
104  else
105  for (int i=0; i<size; ++i)
106  n(i) = size==1 ? low : low + Scalar((double(high-low+1)*double(i))/double(size));
107  VERIFY_IS_APPROX(m,n);
108 
109  // random access version
110  m = VectorType::LinSpaced(size,low,high);
111  VERIFY_IS_APPROX(m,n);
112  VERIFY( internal::isApprox(m(m.size()-1),high) );
113  VERIFY( size==1 || internal::isApprox(m(0),low) );
114  VERIFY_IS_EQUAL(m(m.size()-1) , high);
116  CALL_SUBTEST( check_extremity_accuracy(m, low, high) );
117  }
118 
119  VERIFY( m(m.size()-1) <= high );
120  VERIFY( (m.array() <= high).all() );
121  VERIFY( (m.array() >= low).all() );
122 
123 
124  VERIFY( m(m.size()-1) >= low );
125  if(size>=1)
126  {
127  VERIFY( internal::isApprox(m(0),low) );
128  VERIFY_IS_EQUAL(m(0) , low);
129  }
130 
131  // check whether everything works with row and col major vectors
132  Matrix<Scalar,Dynamic,1> row_vector(size);
133  Matrix<Scalar,1,Dynamic> col_vector(size);
134  row_vector.setLinSpaced(size,low,high);
135  col_vector.setLinSpaced(size,low,high);
136  // when using the extended precision (e.g., FPU) the relative error might exceed 1 bit
137  // when computing the squared sum in isApprox, thus the 2x factor.
138  VERIFY( row_vector.isApprox(col_vector.transpose(), Scalar(2)*NumTraits<Scalar>::epsilon()));
139 
140  Matrix<Scalar,Dynamic,1> size_changer(size+50);
141  size_changer.setLinSpaced(size,low,high);
142  VERIFY( size_changer.size() == size );
143 
144  typedef Matrix<Scalar,1,1> ScalarMatrix;
145  ScalarMatrix scalar;
146  scalar.setLinSpaced(1,low,high);
147  VERIFY_IS_APPROX( scalar, ScalarMatrix::Constant(high) );
148  VERIFY_IS_APPROX( ScalarMatrix::LinSpaced(1,low,high), ScalarMatrix::Constant(high) );
149 
150  // regression test for bug 526 (linear vectorized transversal)
151  if (size > 1 && (!NumTraits<Scalar>::IsInteger)) {
152  m.tail(size-1).setLinSpaced(low, high);
153  VERIFY_IS_APPROX(m(size-1), high);
154  }
155 
156  // regression test for bug 1383 (LinSpaced with empty size/range)
157  {
158  Index n0 = VectorType::SizeAtCompileTime==Dynamic ? 0 : VectorType::SizeAtCompileTime;
159  low = internal::random<Scalar>();
160  m = VectorType::LinSpaced(n0,low,low-1);
161  VERIFY(m.size()==n0);
162 
163  if(VectorType::SizeAtCompileTime==Dynamic)
164  {
165  VERIFY_IS_EQUAL(VectorType::LinSpaced(n0,0,Scalar(n0-1)).sum(),Scalar(0));
166  VERIFY_IS_EQUAL(VectorType::LinSpaced(n0,low,low-1).sum(),Scalar(0));
167  }
168 
169  m.setLinSpaced(n0,0,Scalar(n0-1));
170  VERIFY(m.size()==n0);
171  m.setLinSpaced(n0,low,low-1);
172  VERIFY(m.size()==n0);
173 
174  // empty range only:
175  VERIFY_IS_APPROX(VectorType::LinSpaced(size,low,low),VectorType::Constant(size,low));
176  m.setLinSpaced(size,low,low);
177  VERIFY_IS_APPROX(m,VectorType::Constant(size,low));
178 
180  {
181  VERIFY_IS_APPROX( VectorType::LinSpaced(size,low,Scalar(low+size-1)), VectorType::LinSpaced(size,Scalar(low+size-1),low).reverse() );
182 
183  if(VectorType::SizeAtCompileTime==Dynamic)
184  {
185  // Check negative multiplicator path:
186  for(Index k=1; k<5; ++k)
187  VERIFY_IS_APPROX( VectorType::LinSpaced(size,low,Scalar(low+(size-1)*k)), VectorType::LinSpaced(size,Scalar(low+(size-1)*k),low).reverse() );
188  // Check negative divisor path:
189  for(Index k=1; k<5; ++k)
190  VERIFY_IS_APPROX( VectorType::LinSpaced(size*k,low,Scalar(low+size-1)), VectorType::LinSpaced(size*k,Scalar(low+size-1),low).reverse() );
191  }
192  }
193  }
194 }
195 
196 template<typename MatrixType>
198 {
199  using std::abs;
200  const Index rows = m.rows();
201  const Index cols = m.cols();
202  typedef typename MatrixType::Scalar Scalar;
203  typedef typename MatrixType::RealScalar RealScalar;
204 
205  Scalar s1;
206  do {
207  s1 = internal::random<Scalar>();
208  } while(abs(s1)<RealScalar(1e-5) && (!NumTraits<Scalar>::IsInteger));
209 
210  MatrixType A;
211  A.setIdentity(rows, cols);
213  VERIFY(equalsIdentity(MatrixType::Identity(rows, cols)));
214 
215 
216  A = MatrixType::Constant(rows,cols,s1);
217  Index i = internal::random<Index>(0,rows-1);
218  Index j = internal::random<Index>(0,cols-1);
219  VERIFY_IS_APPROX( MatrixType::Constant(rows,cols,s1)(i,j), s1 );
220  VERIFY_IS_APPROX( MatrixType::Constant(rows,cols,s1).coeff(i,j), s1 );
221  VERIFY_IS_APPROX( A(i,j), s1 );
222 }
223 
225 {
226  CALL_SUBTEST_1( testMatrixType(Matrix2d()) );
227  CALL_SUBTEST_2( testMatrixType(MatrixXcf(internal::random<int>(1,300),internal::random<int>(1,300))) );
228  CALL_SUBTEST_3( testMatrixType(MatrixXf(internal::random<int>(1,300),internal::random<int>(1,300))) );
229 
230  for(int i = 0; i < g_repeat*10; i++) {
231  CALL_SUBTEST_4( testVectorType(VectorXd(internal::random<int>(1,30000))) );
232  CALL_SUBTEST_5( testVectorType(Vector4d()) ); // regression test for bug 232
233  CALL_SUBTEST_6( testVectorType(Vector3d()) );
234  CALL_SUBTEST_7( testVectorType(VectorXf(internal::random<int>(1,30000))) );
235  CALL_SUBTEST_8( testVectorType(Vector3f()) );
236  CALL_SUBTEST_8( testVectorType(Vector4f()) );
237  CALL_SUBTEST_8( testVectorType(Matrix<float,8,1>()) );
238  CALL_SUBTEST_8( testVectorType(Matrix<float,1,1>()) );
239 
240  CALL_SUBTEST_9( testVectorType(VectorXi(internal::random<int>(1,10))) );
241  CALL_SUBTEST_9( testVectorType(VectorXi(internal::random<int>(9,300))) );
242  CALL_SUBTEST_9( testVectorType(Matrix<int,1,1>()) );
243  }
244 
245 #ifdef EIGEN_TEST_PART_6
246  // Assignment of a RowVectorXd to a MatrixXd (regression test for bug #79).
247  VERIFY( (MatrixXd(RowVectorXd::LinSpaced(3, 0, 1)) - RowVector3d(0, 0.5, 1)).norm() < std::numeric_limits<double>::epsilon() );
248 #endif
249 
250 #ifdef EIGEN_TEST_PART_9
251  // Check possible overflow issue
252  {
253  int n = 60000;
254  ArrayXi a1(n), a2(n);
255  a1.setLinSpaced(n, 0, n-1);
256  for(int i=0; i<n; ++i)
257  a2(i) = i;
258  VERIFY_IS_APPROX(a1,a2);
259  }
260 #endif
261 
262 #ifdef EIGEN_TEST_PART_10
263  // check some internal logic
264  VERIFY(( internal::has_nullary_operator<internal::scalar_constant_op<double> >::value ));
265  VERIFY(( !internal::has_unary_operator<internal::scalar_constant_op<double> >::value ));
266  VERIFY(( !internal::has_binary_operator<internal::scalar_constant_op<double> >::value ));
267  VERIFY(( internal::functor_has_linear_access<internal::scalar_constant_op<double> >::ret ));
268 
269  VERIFY(( !internal::has_nullary_operator<internal::scalar_identity_op<double> >::value ));
270  VERIFY(( !internal::has_unary_operator<internal::scalar_identity_op<double> >::value ));
271  VERIFY(( internal::has_binary_operator<internal::scalar_identity_op<double> >::value ));
272  VERIFY(( !internal::functor_has_linear_access<internal::scalar_identity_op<double> >::ret ));
273 
274  VERIFY(( !internal::has_nullary_operator<internal::linspaced_op<float,float> >::value ));
275  VERIFY(( internal::has_unary_operator<internal::linspaced_op<float,float> >::value ));
276  VERIFY(( !internal::has_binary_operator<internal::linspaced_op<float,float> >::value ));
277  VERIFY(( internal::functor_has_linear_access<internal::linspaced_op<float,float> >::ret ));
278 
279  // Regression unit test for a weird MSVC bug.
280  // Search "nullary_wrapper_workaround_msvc" in CoreEvaluators.h for the details.
281  // See also traits<Ref>::match.
282  {
283  MatrixXf A = MatrixXf::Random(3,3);
284  Ref<const MatrixXf> R = 2.0*A;
285  VERIFY_IS_APPROX(R, A+A);
286 
287  Ref<const MatrixXf> R1 = MatrixXf::Random(3,3)+A;
288 
289  VectorXi V = VectorXi::Random(3);
290  Ref<const VectorXi> R2 = VectorXi::LinSpaced(3,1,3)+V;
291  VERIFY_IS_APPROX(R2, V+Vector3i(1,2,3));
292 
293  VERIFY(( internal::has_nullary_operator<internal::scalar_constant_op<float> >::value ));
294  VERIFY(( !internal::has_unary_operator<internal::scalar_constant_op<float> >::value ));
295  VERIFY(( !internal::has_binary_operator<internal::scalar_constant_op<float> >::value ));
296  VERIFY(( internal::functor_has_linear_access<internal::scalar_constant_op<float> >::ret ));
297 
298  VERIFY(( !internal::has_nullary_operator<internal::linspaced_op<int,int> >::value ));
299  VERIFY(( internal::has_unary_operator<internal::linspaced_op<int,int> >::value ));
300  VERIFY(( !internal::has_binary_operator<internal::linspaced_op<int,int> >::value ));
301  VERIFY(( internal::functor_has_linear_access<internal::linspaced_op<int,int> >::ret ));
302  }
303 #endif
304 }
Matrix3f m
def step(data, isam, result, truth, currPoseIndex)
Definition: visual_isam.py:82
SCALAR Scalar
Definition: bench_gemm.cpp:33
void test_nullary()
Definition: nullary.cpp:224
#define min(a, b)
Definition: datatypes.h:19
EIGEN_DONT_INLINE Scalar zero()
Definition: svd_common.h:271
ArrayXcf v
Definition: Cwise_arg.cpp:1
int n
Rot2 R(Rot2::fromAngle(0.1))
MatrixXf MatrixType
Holds information about the various numeric (i.e. scalar) types allowed by Eigen. ...
Definition: NumTraits.h:150
Matrix< SCALARA, Dynamic, Dynamic > A
Definition: bench_gemm.cpp:35
void check_extremity_accuracy(const VectorType &v, const typename VectorType::Scalar &low, const typename VectorType::Scalar &high)
Definition: nullary.cpp:37
static double epsilon
Definition: testRot3.cpp:39
Scalar Scalar int size
Definition: benchVecAdd.cpp:17
#define VERIFY_IS_APPROX(a, b)
#define VERIFY_IS_EQUAL(a, b)
Definition: main.h:331
bool equalsIdentity(const MatrixType &A)
Definition: nullary.cpp:14
static int g_repeat
Definition: main.h:144
EIGEN_DEFAULT_DENSE_INDEX_TYPE Index
The Index type as used for the API.
Definition: Meta.h:33
void testMatrixType(const MatrixType &m)
Definition: nullary.cpp:197
Point2(* f)(const Point3 &, OptionalJacobian< 2, 3 >)
Array< double, 1, 3 > e(1./3., 0.5, 2.)
const mpreal sum(const mpreal tab[], const unsigned long int n, int &status, mp_rnd_t mode=mpreal::get_default_rnd())
Definition: mpreal.h:2381
Q R1(Eigen::AngleAxisd(1, Q_z_axis))
NumTraits< Scalar >::Real RealScalar
Definition: bench_gemm.cpp:34
A matrix or vector expression mapping an existing expression.
Definition: Ref.h:192
#define CALL_SUBTEST(FUNC)
Definition: main.h:342
#define VERIFY(a)
Definition: main.h:325
void reverse(const MatrixType &m)
DenseIndex ret
Definition: level1_impl.h:59
Annotation indicating that a class derives from another given type.
Definition: attr.h:42
const int Dynamic
Definition: Constants.h:21
Jet< T, N > pow(const Jet< T, N > &f, double g)
Definition: jet.h:570
EIGEN_DEVICE_FUNC bool isApprox(const Scalar &x, const Scalar &y, const typename NumTraits< Scalar >::Real &precision=NumTraits< Scalar >::dummy_precision())
#define abs(x)
Definition: datatypes.h:17
mxArray * scalar(mxClassID classid)
Definition: matlab.h:85
Q R2(Eigen::AngleAxisd(2, Vector3(0, 1, 0)))
void swap(mpfr::mpreal &x, mpfr::mpreal &y)
Definition: mpreal.h:2986
std::ptrdiff_t j
void testVectorType(const VectorType &base)
Definition: nullary.cpp:64


gtsam
Author(s):
autogenerated on Sat May 8 2021 02:43:04