FullPivHouseholderQR.hpp
Go to the documentation of this file.
1 /*
2  * Copyright 2024 INRIA
3  */
4 
5 #ifndef __eigenpy_decompositions_full_piv_houselholder_qr_hpp__
6 #define __eigenpy_decompositions_full_piv_houselholder_qr_hpp__
7 
8 #include "eigenpy/eigenpy.hpp"
10 
11 #include <Eigen/QR>
12 
13 namespace eigenpy {
14 
15 template <typename _MatrixType>
17  : public boost::python::def_visitor<
18  FullPivHouseholderQRSolverVisitor<_MatrixType> > {
19  typedef _MatrixType MatrixType;
20  typedef typename MatrixType::Scalar Scalar;
21  typedef typename MatrixType::RealScalar RealScalar;
22  typedef Eigen::Matrix<Scalar, Eigen::Dynamic, 1, MatrixType::Options>
24  typedef Eigen::Matrix<Scalar, Eigen::Dynamic, Eigen::Dynamic,
25  MatrixType::Options>
27  typedef Eigen::FullPivHouseholderQR<MatrixType> Solver;
28  typedef Solver Self;
29 
30  template <class PyClass>
31  void visit(PyClass &cl) const {
32  cl.def(bp::init<>(bp::arg("self"),
33  "Default constructor.\n"
34  "The default constructor is useful in cases in which the "
35  "user intends to perform decompositions via "
36  "HouseholderQR.compute(matrix)"))
37  .def(bp::init<Eigen::DenseIndex, Eigen::DenseIndex>(
38  bp::args("self", "rows", "cols"),
39  "Default constructor with memory preallocation.\n"
40  "Like the default constructor but with preallocation of the "
41  "internal data according to the specified problem size. "))
42  .def(bp::init<MatrixType>(
43  bp::args("self", "matrix"),
44  "Constructs a QR factorization from a given matrix.\n"
45  "This constructor computes the QR factorization of the matrix "
46  "matrix by calling the method compute()."))
47 
48  .def("absDeterminant", &Self::absDeterminant, bp::arg("self"),
49  "Returns the absolute value of the determinant of the matrix of "
50  "which *this is the QR decomposition.\n"
51  "It has only linear complexity (that is, O(n) where n is the "
52  "dimension of the square matrix) as the QR decomposition has "
53  "already been computed.\n"
54  "Note: This is only for square matrices.")
55  .def("logAbsDeterminant", &Self::logAbsDeterminant, bp::arg("self"),
56  "Returns the natural log of the absolute value of the determinant "
57  "of the matrix of which *this is the QR decomposition.\n"
58  "It has only linear complexity (that is, O(n) where n is the "
59  "dimension of the square matrix) as the QR decomposition has "
60  "already been computed.\n"
61  "Note: This is only for square matrices. This method is useful to "
62  "work around the risk of overflow/underflow that's inherent to "
63  "determinant computation.")
64  .def("dimensionOfKernel", &Self::dimensionOfKernel, bp::arg("self"),
65  "Returns the dimension of the kernel of the matrix of which *this "
66  "is the QR decomposition.")
67  .def("isInjective", &Self::isInjective, bp::arg("self"),
68  "Returns true if the matrix associated with this QR decomposition "
69  "represents an injective linear map, i.e. has trivial kernel; "
70  "false otherwise.\n"
71  "\n"
72  "Note: This method has to determine which pivots should be "
73  "considered nonzero. For that, it uses the threshold value that "
74  "you can control by calling setThreshold(threshold).")
75  .def("isInvertible", &Self::isInvertible, bp::arg("self"),
76  "Returns true if the matrix associated with the QR decomposition "
77  "is invertible.\n"
78  "\n"
79  "Note: This method has to determine which pivots should be "
80  "considered nonzero. For that, it uses the threshold value that "
81  "you can control by calling setThreshold(threshold).")
82  .def("isSurjective", &Self::isSurjective, bp::arg("self"),
83  "Returns true if the matrix associated with this QR decomposition "
84  "represents a surjective linear map; false otherwise.\n"
85  "\n"
86  "Note: This method has to determine which pivots should be "
87  "considered nonzero. For that, it uses the threshold value that "
88  "you can control by calling setThreshold(threshold).")
89  .def("maxPivot", &Self::maxPivot, bp::arg("self"),
90  "Returns the absolute value of the biggest pivot, i.e. the "
91  "biggest diagonal coefficient of U.")
92  .def("nonzeroPivots", &Self::nonzeroPivots, bp::arg("self"),
93  "Returns the number of nonzero pivots in the QR decomposition. "
94  "Here nonzero is meant in the exact sense, not in a fuzzy sense. "
95  "So that notion isn't really intrinsically interesting, but it is "
96  "still useful when implementing algorithms.")
97  .def("rank", &Self::rank, bp::arg("self"),
98  "Returns the rank of the matrix associated with the QR "
99  "decomposition.\n"
100  "\n"
101  "Note: This method has to determine which pivots should be "
102  "considered nonzero. For that, it uses the threshold value that "
103  "you can control by calling setThreshold(threshold).")
104 
105  .def("setThreshold",
106  (Self & (Self::*)(const RealScalar &)) & Self::setThreshold,
107  bp::args("self", "threshold"),
108  "Allows to prescribe a threshold to be used by certain methods, "
109  "such as rank(), who need to determine when pivots are to be "
110  "considered nonzero. This is not used for the QR decomposition "
111  "itself.\n"
112  "\n"
113  "When it needs to get the threshold value, Eigen calls "
114  "threshold(). By default, this uses a formula to automatically "
115  "determine a reasonable threshold. Once you have called the "
116  "present method setThreshold(const RealScalar&), your value is "
117  "used instead.\n"
118  "\n"
119  "Note: A pivot will be considered nonzero if its absolute value "
120  "is strictly greater than |pivot| ⩽ threshold×|maxpivot| where "
121  "maxpivot is the biggest pivot.",
122  bp::return_self<>())
123  .def("threshold", &Self::threshold, bp::arg("self"),
124  "Returns the threshold that will be used by certain methods such "
125  "as rank().")
126 
127  .def("matrixQR", &Self::matrixQR, bp::arg("self"),
128  "Returns the matrix where the Householder QR decomposition is "
129  "stored in a LAPACK-compatible way.",
130  bp::return_value_policy<bp::copy_const_reference>())
131 
132  .def(
133  "compute",
134  (Solver & (Solver::*)(const Eigen::EigenBase<MatrixType> &matrix)) &
135  Solver::compute,
136  bp::args("self", "matrix"),
137  "Computes the QR factorization of given matrix.",
138  bp::return_self<>())
139 
140  .def("inverse", inverse, bp::arg("self"),
141  "Returns the inverse of the matrix associated with the QR "
142  "decomposition..")
143 
144  .def("solve", &solve<MatrixXs>, bp::args("self", "B"),
145  "Returns the solution X of A X = B using the current "
146  "decomposition of A where B is a right hand side matrix.");
147  }
148 
149  static void expose() {
150  static const std::string classname =
151  "FullPivHouseholderQR" + scalar_name<Scalar>::shortname();
152  expose(classname);
153  }
154 
155  static void expose(const std::string &name) {
156  bp::class_<Solver>(
157  name.c_str(),
158  "This class performs a rank-revealing QR decomposition of a matrix A "
159  "into matrices P, P', Q and R such that:\n"
160  "PAP′=QR\n"
161  "by using Householder transformations. Here, P and P' are permutation "
162  "matrices, Q a unitary matrix and R an upper triangular matrix.\n"
163  "\n"
164  "This decomposition performs a very prudent full pivoting in order to "
165  "be rank-revealing and achieve optimal numerical stability. The "
166  "trade-off is that it is slower than HouseholderQR and "
167  "ColPivHouseholderQR.",
168  bp::no_init)
170  .def(IdVisitor<Solver>());
171  }
172 
173  private:
174  template <typename MatrixOrVector>
175  static MatrixOrVector solve(const Solver &self, const MatrixOrVector &vec) {
176  return self.solve(vec);
177  }
178  static MatrixXs inverse(const Self &self) { return self.inverse(); }
179 };
180 
181 } // namespace eigenpy
182 
183 #endif // ifndef __eigenpy_decompositions_full_piv_houselholder_qr_hpp__
eigenpy::FullPivHouseholderQRSolverVisitor::expose
static void expose()
Definition: FullPivHouseholderQR.hpp:149
scalar-name.hpp
eigenpy::FullPivHouseholderQRSolverVisitor::Scalar
MatrixType::Scalar Scalar
Definition: FullPivHouseholderQR.hpp:20
eigenpy::FullPivHouseholderQRSolverVisitor::VectorXs
Eigen::Matrix< Scalar, Eigen::Dynamic, 1, MatrixType::Options > VectorXs
Definition: FullPivHouseholderQR.hpp:23
test_matrix.vec
vec
Definition: test_matrix.py:180
eigenpy::FullPivHouseholderQRSolverVisitor
Definition: FullPivHouseholderQR.hpp:16
eigenpy::FullPivHouseholderQRSolverVisitor::expose
static void expose(const std::string &name)
Definition: FullPivHouseholderQR.hpp:155
eigenpy::FullPivHouseholderQRSolverVisitor::solve
static MatrixOrVector solve(const Solver &self, const MatrixOrVector &vec)
Definition: FullPivHouseholderQR.hpp:175
eigenpy
Definition: alignment.hpp:14
eigenpy::FullPivHouseholderQRSolverVisitor::MatrixXs
Eigen::Matrix< Scalar, Eigen::Dynamic, Eigen::Dynamic, MatrixType::Options > MatrixXs
Definition: FullPivHouseholderQR.hpp:26
eigenpy::scalar_name::shortname
static std::string shortname()
eigenpy::FullPivHouseholderQRSolverVisitor::Solver
Eigen::FullPivHouseholderQR< MatrixType > Solver
Definition: FullPivHouseholderQR.hpp:27
eigenpy::FullPivHouseholderQRSolverVisitor::Self
Solver Self
Definition: FullPivHouseholderQR.hpp:28
eigenpy::IdVisitor
Add the Python method id to retrieving a unique id for a given object exposed with Boost....
Definition: id.hpp:18
setup.name
name
Definition: setup.in.py:179
eigenpy.hpp
eigenpy::FullPivHouseholderQRSolverVisitor::inverse
static MatrixXs inverse(const Self &self)
Definition: FullPivHouseholderQR.hpp:178
eigenpy::FullPivHouseholderQRSolverVisitor::visit
void visit(PyClass &cl) const
Definition: FullPivHouseholderQR.hpp:31
eigenpy::FullPivHouseholderQRSolverVisitor::RealScalar
MatrixType::RealScalar RealScalar
Definition: FullPivHouseholderQR.hpp:21
eigenpy::FullPivHouseholderQRSolverVisitor::MatrixType
_MatrixType MatrixType
Definition: FullPivHouseholderQR.hpp:19


eigenpy
Author(s): Justin Carpentier, Nicolas Mansard
autogenerated on Sat Nov 2 2024 02:14:45