Re: [eigen] [PATCH] adding Reverse expression |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] [PATCH] adding Reverse expression
- From: Ricard Marxer Piñón <email@xxxxxxxxxxxxxxxx>
- Date: Tue, 3 Feb 2009 01:41:11 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:sender:received:in-reply-to :references:date:x-google-sender-auth:message-id:subject:from:to :content-type; bh=sYHvgNmQu84VjNNEkgekkzo3jO2FhX37HXY6qJ5Z7+E=; b=JOmAbPAMlj/6g/wAwatX1SUNm24ET2uLINip/G3HcU9pyn4+Nm+tbQW1n4sljnTs+t zXhOb7bLrs16HFRqxU013mkrgyaQWtLZssRAVz8LzcvLDeAKl0Bc57taUrZkIXlSsHUA nv4xaE3/gWSGovQ4e17NeKrRgx16WeN8hWriY=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:content-type; b=EnhVU1CDZy+wvzovYd+l37twG+badrt3wyeq7KAbmhd0yDXmwmHYOBKqrBX4p16nOD C0C3fKXjvHxKgHJUchkEoDr7tRaznZiNsotmO9ZE9J0bqtDeaf/HNF4yB8OzxKuMlE12 6SYYIZqPYZaTONWgUZ43z6oyw69A8FFEJMMgo=
On Tue, Feb 3, 2009 at 12:50 AM, Benoit Jacob
<jacob.benoit.1@xxxxxxxxx> wrote:
Thanks for your patch; comments below:
2009/2/3 Ricard Marxer Piñón <email@xxxxxxxxxxxxxxxx>:
> Here is a patch to add the Reverse _expression_ this is useful for flipping
> matrices, rows or columns.
I can see that you declare a class Reverse, but don't actually define
it. You probably forgot to "svn add Reverse.h" ?
ups yep. I had some trouble because I still had all the Toeplitz stuff and was trying to isolate the reverse stuff (aaah I wish I could get git svn to work)
Anyway here is the full diff.
Your PartialRedux::reverse() method returns a ExpressionType which has
2 drawbacks:
1) it requires ExpressionType to be writable, e.g. one can't use your
reverse() like this,
m3 = (m1+m2).rowwise().reverse();
2) more importantly, your PartialRedux::reverse() is not lazy, so when one does
v2 = v1.reverse();
the arrays are traversed twice: first a temporary vector is created by
reverse(), then this temporary vector is copied into v2.
This is of course inefficient. It is exactly in order to overcome
that, that we introduct _expression_ templates ;)
Yes the PartialRedux reverse is very drafty, I just wanted the API to be as I wanted it to be. I will try to get it right with a bit more work.
Cheers,
Benoit
--
ricard
http://www.ricardmarxer.com
http://www.caligraft.com
Index: doc/snippets/PartialRedux_reverse.cpp
===================================================================
--- doc/snippets/PartialRedux_reverse.cpp (revision 0)
+++ doc/snippets/PartialRedux_reverse.cpp (revision 0)
@@ -0,0 +1,4 @@
+MatrixXi m = MatrixXi::Random(3,4);
+cout << "Here is the matrix m:" << endl << m << endl;
+cout << "Here is the rowwise reverse of m:" << endl << m.rowwise().reverse() << endl;
+cout << "Here is the colwise reverse of m:" << endl << m.colwise().reverse() << endl;
Index: doc/snippets/MatrixBase_reverse.cpp
===================================================================
--- doc/snippets/MatrixBase_reverse.cpp (revision 0)
+++ doc/snippets/MatrixBase_reverse.cpp (revision 0)
@@ -0,0 +1,8 @@
+MatrixXi m = MatrixXi::Random(3,4);
+cout << "Here is the matrix m:" << endl << m << endl;
+cout << "Here is the reverse of m:" << endl << m.reverse() << endl;
+cout << "Here is the coefficient (1,0) in the reverse of m:" << endl
+ << m.reverse()(1,0) << endl;
+cout << "Let us overwrite this coefficient with the value 4." << endl;
+m.reverse()(1,0) = 4;
+cout << "Now the matrix m is:" << endl << m << endl;
Index: Eigen/src/Core/Reverse.h
===================================================================
--- Eigen/src/Core/Reverse.h (revision 0)
+++ Eigen/src/Core/Reverse.h (revision 0)
@@ -0,0 +1,169 @@
+// This file is part of Eigen, a lightweight C++ template library
+// for linear algebra. Eigen itself is part of the KDE project.
+//
+// Copyright (C) 2006-2008 Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
+// Copyright (C) 2006-2008 Ricard Marxer <email@xxxxxxxxxxxxxxxx>
+//
+// Eigen is free software; you can redistribute it and/or
+// modify it under the terms of the GNU Lesser General Public
+// License as published by the Free Software Foundation; either
+// version 3 of the License, or (at your option) any later version.
+//
+// Alternatively, you can redistribute it and/or
+// modify it under the terms of the GNU General Public License as
+// published by the Free Software Foundation; either version 2 of
+// the License, or (at your option) any later version.
+//
+// Eigen is distributed in the hope that it will be useful, but WITHOUT ANY
+// WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+// FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License or the
+// GNU General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public
+// License and a copy of the GNU General Public License along with
+// Eigen. If not, see <http://www.gnu.org/licenses/>.
+
+#ifndef EIGEN_REVERSE_H
+#define EIGEN_REVERSE_H
+
+/** \class Reverse
+ *
+ * \brief Expression of the reverse of a vector
+ *
+ * \param MatrixType the type of the object of which we are taking the reverse
+ *
+ * This class represents an expression of the reverse of a vector.
+ * It is the return type of MatrixBase::reverse()
+ * and most of the time this is the only way it is used.
+ *
+ * \sa MatrixBase::reverse()
+ */
+template<typename MatrixType>
+struct ei_traits<Reverse<MatrixType> >
+{
+ typedef typename MatrixType::Scalar Scalar;
+ typedef typename ei_nested<MatrixType>::type MatrixTypeNested;
+ typedef typename ei_unref<MatrixTypeNested>::type _MatrixTypeNested;
+ enum {
+ RowsAtCompileTime = MatrixType::RowsAtCompileTime,
+ ColsAtCompileTime = MatrixType::ColsAtCompileTime,
+ MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime,
+ MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime,
+
+ // TODO: check how to correctly set the new flags
+ Flags = ((int(_MatrixTypeNested::Flags) ^ RowMajorBit)
+ & ~(LowerTriangularBit | UpperTriangularBit))
+ | (int(_MatrixTypeNested::Flags)&UpperTriangularBit ? LowerTriangularBit : 0)
+ | (int(_MatrixTypeNested::Flags)&LowerTriangularBit ? UpperTriangularBit : 0),
+
+ // TODO: should add two index type AddCosts (due to the rows - i - 1) or only one, and add the cost of calling .rows() and .cols()
+ CoeffReadCost = _MatrixTypeNested::CoeffReadCost
+ };
+};
+
+template<typename MatrixType> class Reverse
+ : public MatrixBase<Reverse<MatrixType> >
+{
+ public:
+
+ EIGEN_GENERIC_PUBLIC_INTERFACE(Reverse)
+
+ inline Reverse(const MatrixType& matrix) : m_matrix(matrix) {}
+
+ EIGEN_INHERIT_ASSIGNMENT_OPERATORS(Reverse)
+
+ inline int rows() const { return m_matrix.rows(); }
+ inline int cols() const { return m_matrix.cols(); }
+ inline int nonZeros() const { return m_matrix.nonZeros(); }
+ inline int stride(void) const { return m_matrix.stride(); }
+
+ inline Scalar& coeffRef(int row, int col)
+ {
+ return m_matrix.const_cast_derived().coeffRef(m_matrix.rows() - row - 1, m_matrix.cols() - col - 1);
+ }
+
+ inline const Scalar coeff(int row, int col) const
+ {
+ return m_matrix.coeff(m_matrix.rows() - row - 1, m_matrix.cols() - col - 1);
+ }
+
+ inline const Scalar coeff(int index) const
+ {
+ return m_matrix.coeff(m_matrix.rows() * m_matrix.cols() - index - 1);
+ }
+
+ inline Scalar& coeffRef(int index)
+ {
+ return m_matrix.const_cast_derived().coeffRef(m_matrix.rows() * m_matrix.cols() - index - 1);
+ }
+
+ // TODO: We must reverse the packet reading and writing, which is currently not done here, I think
+ template<int LoadMode>
+ inline const PacketScalar packet(int row, int col) const
+ {
+ return m_matrix.template packet<LoadMode>(m_matrix.rows() - row - 1, m_matrix.cols() - col - 1);
+ }
+
+ template<int LoadMode>
+ inline void writePacket(int row, int col, const PacketScalar& x)
+ {
+ m_matrix.const_cast_derived().template writePacket<LoadMode>(m_matrix.rows() - row - 1, m_matrix.cols() - col - 1, x);
+ }
+
+ template<int LoadMode>
+ inline const PacketScalar packet(int index) const
+ {
+ return m_matrix.template packet<LoadMode>(m_matrix.rows() * m_matrix.cols() - index - 1);
+ }
+
+ template<int LoadMode>
+ inline void writePacket(int index, const PacketScalar& x)
+ {
+ m_matrix.const_cast_derived().template writePacket<LoadMode>(m_matrix.rows() * m_matrix.cols() - index - 1, x);
+ }
+
+ protected:
+ const typename MatrixType::Nested m_matrix;
+};
+
+/** \returns an expression of the reverse of *this.
+ *
+ * Example: \include MatrixBase_reverse.cpp
+ * Output: \verbinclude MatrixBase_reverse.out
+ *
+ */
+template<typename Derived>
+inline Reverse<Derived>
+MatrixBase<Derived>::reverse()
+{
+ return derived();
+}
+
+/** This is the const version of reverse(). */
+template<typename Derived>
+inline const Reverse<Derived>
+MatrixBase<Derived>::reverse() const
+{
+ return derived();
+}
+
+/** This is the "in place" version of reverse: it reverses \c *this.
+ *
+ * In most cases it is probably better to simply use the reversed expression
+ * of a matrix. However, when reversing the matrix data itself is really needed,
+ * then this "in-place" version is probably the right choice because it provides
+ * the following additional features:
+ * - less error prone: doing the same operation with .reverse() requires special care:
+ * \code m = m.reverse().eval(); \endcode
+ * - no temporary object is created (currently there is one created but could be avoided using swap)
+ * - it allows future optimizations (cache friendliness, etc.)
+ *
+ * \sa reverse() */
+template<typename Derived>
+inline void MatrixBase<Derived>::reverseInPlace()
+{
+ this = this.reverse().eval();
+}
+
+
+#endif // EIGEN_REVERSE_H
Index: Eigen/src/Core/MatrixBase.h
===================================================================
--- Eigen/src/Core/MatrixBase.h (revision 920471)
+++ Eigen/src/Core/MatrixBase.h (working copy)
@@ -359,8 +359,11 @@
const Eigen::Transpose<Derived> transpose() const;
void transposeInPlace();
const AdjointReturnType adjoint() const;
+
+ Eigen::Reverse<Derived> reverse();
+ const Eigen::Reverse<Derived> reverse() const;
+ void reverseInPlace();
-
RowXpr row(int i);
const RowXpr row(int i) const;
Index: Eigen/src/Core/util/ForwardDeclarations.h
===================================================================
--- Eigen/src/Core/util/ForwardDeclarations.h (revision 920471)
+++ Eigen/src/Core/util/ForwardDeclarations.h (working copy)
@@ -40,6 +40,7 @@
int _DirectAccessStatus = ei_traits<MatrixType>::Flags&DirectAccessBit ? DirectAccessBit
: ei_traits<MatrixType>::Flags&SparseBit> class Block;
template<typename MatrixType> class Transpose;
+template<typename MatrixType> class Reverse;
template<typename MatrixType> class Conjugate;
template<typename NullaryOp, typename MatrixType> class CwiseNullaryOp;
template<typename UnaryOp, typename MatrixType> class CwiseUnaryOp;
Index: Eigen/src/Array/PartialRedux.h
===================================================================
--- Eigen/src/Array/PartialRedux.h (revision 920471)
+++ Eigen/src/Array/PartialRedux.h (working copy)
@@ -258,6 +258,34 @@
const PartialReduxExpr<ExpressionType, ei_member_count<int>, Direction> count() const
{ return _expression(); }
+
+ /** \returns a matrix expression
+ * where each column (or row) are reversed.
+ *
+ * Example: \include PartialRedux_reverse.cpp
+ * Output: \verbinclude PartialRedux_reverse.out
+ *
+ * \sa MatrixBase::reverse() */
+ const ExpressionType reverse() const
+ {
+ const int rows = _expression().rows();
+ const int cols = _expression().cols();
+
+ ExpressionType mat(rows, cols);
+
+ if(Direction==Vertical) {
+ for (int i = 0; i < cols; i++) {
+ mat.col(i) = _expression().col(i).reverse();
+ }
+ } else {
+ for (int i = 0; i < rows; i++) {
+ mat.row(i) = _expression().row(i).reverse();
+ }
+ }
+
+ return mat;
+ }
+
/** \returns a 3x3 matrix expression of the cross product
* of each column or row of the referenced expression with the \a other vector.
*
Index: Eigen/Core
===================================================================
--- Eigen/Core (revision 920471)
+++ Eigen/Core (working copy)
@@ -136,6 +136,7 @@
#include "src/Core/Block.h"
#include "src/Core/Minor.h"
#include "src/Core/Transpose.h"
+#include "src/Core/Reverse.h"
#include "src/Core/DiagonalMatrix.h"
#include "src/Core/DiagonalCoeffs.h"
#include "src/Core/Sum.h"