1 // This file is part of Eigen, a lightweight C++ template library 2 // for linear algebra. 3 // 4 // Copyright (C) 2008-2011 Gael Guennebaud <gael.guennebaud (at) 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 "sparse.h" 11 12 template<typename Scalar,typename StorageIndex> void sparse_vector(int rows, int cols) 13 { 14 double densityMat = (std::max)(8./(rows*cols), 0.01); 15 double densityVec = (std::max)(8./(rows), 0.1); 16 typedef Matrix<Scalar,Dynamic,Dynamic> DenseMatrix; 17 typedef Matrix<Scalar,Dynamic,1> DenseVector; 18 typedef SparseVector<Scalar,0,StorageIndex> SparseVectorType; 19 typedef SparseMatrix<Scalar,0,StorageIndex> SparseMatrixType; 20 Scalar eps = 1e-6; 21 22 SparseMatrixType m1(rows,rows); 23 SparseVectorType v1(rows), v2(rows), v3(rows); 24 DenseMatrix refM1 = DenseMatrix::Zero(rows, rows); 25 DenseVector refV1 = DenseVector::Random(rows), 26 refV2 = DenseVector::Random(rows), 27 refV3 = DenseVector::Random(rows); 28 29 std::vector<int> zerocoords, nonzerocoords; 30 initSparse<Scalar>(densityVec, refV1, v1, &zerocoords, &nonzerocoords); 31 initSparse<Scalar>(densityMat, refM1, m1); 32 33 initSparse<Scalar>(densityVec, refV2, v2); 34 initSparse<Scalar>(densityVec, refV3, v3); 35 36 Scalar s1 = internal::random<Scalar>(); 37 38 // test coeff and coeffRef 39 for (unsigned int i=0; i<zerocoords.size(); ++i) 40 { 41 VERIFY_IS_MUCH_SMALLER_THAN( v1.coeff(zerocoords[i]), eps ); 42 //VERIFY_RAISES_ASSERT( v1.coeffRef(zerocoords[i]) = 5 ); 43 } 44 { 45 VERIFY(int(nonzerocoords.size()) == v1.nonZeros()); 46 int j=0; 47 for (typename SparseVectorType::InnerIterator it(v1); it; ++it,++j) 48 { 49 VERIFY(nonzerocoords[j]==it.index()); 50 VERIFY(it.value()==v1.coeff(it.index())); 51 VERIFY(it.value()==refV1.coeff(it.index())); 52 } 53 } 54 VERIFY_IS_APPROX(v1, refV1); 55 56 // test coeffRef with reallocation 57 { 58 SparseVectorType v4(rows); 59 DenseVector v5 = DenseVector::Zero(rows); 60 for(int k=0; k<rows; ++k) 61 { 62 int i = internal::random<int>(0,rows-1); 63 Scalar v = internal::random<Scalar>(); 64 v4.coeffRef(i) += v; 65 v5.coeffRef(i) += v; 66 } 67 VERIFY_IS_APPROX(v4,v5); 68 } 69 70 v1.coeffRef(nonzerocoords[0]) = Scalar(5); 71 refV1.coeffRef(nonzerocoords[0]) = Scalar(5); 72 VERIFY_IS_APPROX(v1, refV1); 73 74 VERIFY_IS_APPROX(v1+v2, refV1+refV2); 75 VERIFY_IS_APPROX(v1+v2+v3, refV1+refV2+refV3); 76 77 VERIFY_IS_APPROX(v1*s1-v2, refV1*s1-refV2); 78 79 VERIFY_IS_APPROX(v1*=s1, refV1*=s1); 80 VERIFY_IS_APPROX(v1/=s1, refV1/=s1); 81 82 VERIFY_IS_APPROX(v1+=v2, refV1+=refV2); 83 VERIFY_IS_APPROX(v1-=v2, refV1-=refV2); 84 85 VERIFY_IS_APPROX(v1.dot(v2), refV1.dot(refV2)); 86 VERIFY_IS_APPROX(v1.dot(refV2), refV1.dot(refV2)); 87 88 VERIFY_IS_APPROX(m1*v2, refM1*refV2); 89 VERIFY_IS_APPROX(v1.dot(m1*v2), refV1.dot(refM1*refV2)); 90 { 91 int i = internal::random<int>(0,rows-1); 92 VERIFY_IS_APPROX(v1.dot(m1.col(i)), refV1.dot(refM1.col(i))); 93 } 94 95 96 VERIFY_IS_APPROX(v1.squaredNorm(), refV1.squaredNorm()); 97 98 VERIFY_IS_APPROX(v1.blueNorm(), refV1.blueNorm()); 99 100 // test aliasing 101 VERIFY_IS_APPROX((v1 = -v1), (refV1 = -refV1)); 102 VERIFY_IS_APPROX((v1 = v1.transpose()), (refV1 = refV1.transpose().eval())); 103 VERIFY_IS_APPROX((v1 += -v1), (refV1 += -refV1)); 104 105 // sparse matrix to sparse vector 106 SparseMatrixType mv1; 107 VERIFY_IS_APPROX((mv1=v1),v1); 108 VERIFY_IS_APPROX(mv1,(v1=mv1)); 109 VERIFY_IS_APPROX(mv1,(v1=mv1.transpose())); 110 111 // check copy to dense vector with transpose 112 refV3.resize(0); 113 VERIFY_IS_APPROX(refV3 = v1.transpose(),v1.toDense()); 114 VERIFY_IS_APPROX(DenseVector(v1),v1.toDense()); 115 116 // test conservative resize 117 { 118 std::vector<StorageIndex> inc; 119 if(rows > 3) 120 inc.push_back(-3); 121 inc.push_back(0); 122 inc.push_back(3); 123 inc.push_back(1); 124 inc.push_back(10); 125 126 for(std::size_t i = 0; i< inc.size(); i++) { 127 StorageIndex incRows = inc[i]; 128 SparseVectorType vec1(rows); 129 DenseVector refVec1 = DenseVector::Zero(rows); 130 initSparse<Scalar>(densityVec, refVec1, vec1); 131 132 vec1.conservativeResize(rows+incRows); 133 refVec1.conservativeResize(rows+incRows); 134 if (incRows > 0) refVec1.tail(incRows).setZero(); 135 136 VERIFY_IS_APPROX(vec1, refVec1); 137 138 // Insert new values 139 if (incRows > 0) 140 vec1.insert(vec1.rows()-1) = refVec1(refVec1.rows()-1) = 1; 141 142 VERIFY_IS_APPROX(vec1, refVec1); 143 } 144 } 145 146 } 147 148 void test_sparse_vector() 149 { 150 for(int i = 0; i < g_repeat; i++) { 151 int r = Eigen::internal::random<int>(1,500), c = Eigen::internal::random<int>(1,500); 152 if(Eigen::internal::random<int>(0,4) == 0) { 153 r = c; // check square matrices in 25% of tries 154 } 155 EIGEN_UNUSED_VARIABLE(r+c); 156 157 CALL_SUBTEST_1(( sparse_vector<double,int>(8, 8) )); 158 CALL_SUBTEST_2(( sparse_vector<std::complex<double>, int>(r, c) )); 159 CALL_SUBTEST_1(( sparse_vector<double,long int>(r, c) )); 160 CALL_SUBTEST_1(( sparse_vector<double,short>(r, c) )); 161 } 162 } 163 164
