eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version devel
Last Updated 2024-06-18
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version devel
Last Updated 2024-06-18
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version devel
Last Updated 2024-06-18
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version devel
Last Updated 2024-06-18
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version master
Last Updated 2024-06-11
Dev Status MAINTAINED
CI status Continuous Integration : 0 / 0
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

No version for distro ardent. Known supported distros are highlighted in the buttons above.
No version for distro bouncy. Known supported distros are highlighted in the buttons above.
No version for distro crystal. Known supported distros are highlighted in the buttons above.

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version master
Last Updated 2024-06-11
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Package Dependencies

Deps Name
1 catkin

System Dependencies

Dependant Packages

No known dependants.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

No version for distro dashing. Known supported distros are highlighted in the buttons above.

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version devel
Last Updated 2024-06-18
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version master
Last Updated 2024-06-11
Dev Status MAINTAINED
CI status No Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

No version for distro lunar. Known supported distros are highlighted in the buttons above.
No version for distro jade. Known supported distros are highlighted in the buttons above.
No version for distro indigo. Known supported distros are highlighted in the buttons above.
No version for distro hydro. Known supported distros are highlighted in the buttons above.

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version master
Last Updated 2024-06-11
Dev Status MAINTAINED
CI status Continuous Integration
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange

eigenpy package from eigenpy repo

eigenpy

Package Summary

Tags No category tags.
Version 3.7.0
License BSD
Build type CMAKE
Use RECOMMENDED

Repository Summary

Checkout URI https://github.com/stack-of-tasks/eigenpy.git
VCS Type git
VCS Version master
Last Updated 2024-06-11
Dev Status MAINTAINED
CI status Continuous Integration : 0 / 0
Released RELEASED
Tags No category tags.
Contributing Help Wanted (0)
Good First Issues (0)
Pull Requests to Review (0)

Package Description

Bindings between Numpy and Eigen using Boost.Python

Additional Links

Maintainers

  • Justin Carpentier
  • Wolfgang Merkt

Authors

  • Justin Carpentier
  • Nicolas Mansard

EigenPy — Versatile and efficient Python bindings between Numpy and Eigen

License Build Status Conda Downloads Conda Version PyPI version Code style: black Linter: ruff

EigenPy is an open-source framework that allows the binding of the famous Eigen C++ library in Python via Boost.Python.

EigenPy provides: - full memory sharing between Numpy and Eigen, avoiding memory allocation - full support Eigen::Ref avoiding memory allocation - full support of the Eigen::Tensor module - exposition of the Geometry module of Eigen for easy code prototyping - standard matrix decomposion routines of Eigen such as the Cholesky decomposition (SVD and QR decompositions can be added) - full support of SWIG objects - full support of runtime declaration of Numpy scalar types - extended API to expose std::vector types - full support of vectorization between C++ and Python (all the hold objects are properly aligned in memory)

Setup

The installation of EigenPy on your computer is made easy for Linux/BSD, Mac OS X, and Windows environments.

The Conda approach

You simply need this simple line:

conda install eigenpy -c conda-forge

Ubuntu

You can easily install EigenPy from binaries.

Add robotpkg apt repository

  1. Add robotpkg as source repository to apt:
sudo sh -c "echo 'deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -cs) robotpkg' >> /etc/apt/sources.list.d/robotpkg.list"

  1. Register the authentication certificate of robotpkg:
curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -

  1. You need to run at least one apt update to fetch the package descriptions:
sudo apt-get update

Install EigenPy

  1. The installation of EigenPy and its dependencies is made through the line:
sudo apt install robotpkg-py35-eigenpy

where 35 should be replaced by the Python 3, you want to work this (e.g., robotpkg-py36-eigenpy to work with Python 3.6).

Mac OS X

The installation of EigenPy on Mac OS X is made via HomeBrew. You just need to register the tap of the software repository.

brew tap gepetto/homebrew-gepetto

and then install EigenPy for Python 3.x with:

brew install eigenpy

Contributing

Standard matrix decomposion routines of Eigen such as the SVD and QR decompositions can be readily added to EigenPy following the example of the Cholesky decomposition that is already implemented. Feel free to open a PR if you wrap them for your use case.

Credits

The following people have been involved in the development of EigenPy:

If you have taken part in the development of EigenPy, feel free to add your name and contribution here.

Acknowledgments

The development of EigenPy is supported by the Gepetto team @LAAS-CNRS and the Willow team @INRIA.

CHANGELOG

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

Unreleased

3.7.0 - 2024-06-11

Added

  • Added id() helper to retrieve unique object identifier in Python (#477)
  • Expose QR solvers (#478)

3.6.0 - 2024-06-05

Added

  • Added a deprecation call policy shortcut (#466)

Fixed

  • Fix register_symbolic_link_to_registered_type() for multiple successive registrations (#471)

3.5.1 - 2024-04-25

Fixed

  • Allow EigenToPy/EigenFromPy specialization with CL compiler (#462)
  • Fix missing include for boost >= 1.85 (#464)

3.5.0 - 2024-04-14

Added

Fixed

  • Fix unit test build in C++11 (#442)
  • Fix unit test function signature #443)
  • Fix CMake export (#446
  • Fix int management on Windows (#455)
  • Fix long long management on Mac (#455)
  • Allow to run test in the build directory on Windows (#457)

Removed

  • Remove casting when converting from Eigen scalar to Numpy scalar. This should not remove any functionality since Numpy array are created from the Eigen scalar type (#455)

3.4.0 - 2024-02-26

Added

  • Support for Eigen::SparseMatrix types (#426)
  • Support for boost::variant types with VariantConverter (#430)
  • Support for std::variant types with VariantConverter (#431)
  • Support for std::unique_ptr as a return types with StdUniquePtrCallPolicies and boost::python::default_call_policies (#433)
  • Support for std::unique_ptr as an internal reference with ReturnInternalStdUniquePtr (#433)
  • Support for Eigen::Simplicial{LLT,LDLT} and Eigen::Cholmod{Simplicial,Supernodal}{LLT,LDLT} Cholesky de compositions (#438)
  • Switch to ruff for lints, format, and import sort (#441)

Fixed

  • Fix the issue of missing exposition of Eigen types with __int64 scalar type (#426)
  • Fix namespace use in unittest/std_pair.cpp (#429)
  • Fix case of zero-size sparse matrices (#437)

3.3.0 - 2024-01-23

Fixed

  • Fix potential memory leak when returning a list from an std::vector or an std::array (423)

3.2.0 - 2023-12-12

Added

  • Support for C++11 std::array types (#412)
  • Support for std::pair types (#417)

3.1.4 - 2023-11-27

Added

  • Add new helper functions to check Tensor support

Fixed

  • Fix stub generation on Windows

3.1.3 - 2023-11-09

Fixed

  • Install include/eigenpy/registration_class.hpp

3.1.2 - 2023-11-09

Added

  • Support Python 3.12 (#391)

Fixed

  • Add method to std::vector<MatrixXX> binding even if another library had registered it (#393)

Changed

  • CMake minimal version is now 3.10 (#388)

3.1.1 - 2023-07-31

What's Changed

3.1.0 - 2023-06-01

What's Changed

3.0.0 - 2023-04-22

What's Changed

2.9.2 - 2023-02-01

What's Changed

2.9.1 - 2023-01-31

What's Changed

2.9.0 - 2023-01-09

What's Changed

2.8.1 - 2022-12-07

What's Changed

2.8.0 - 2022-12-05

What's Changed

New Contributors

2.7.14 - 2022-09-11

What's Changed

2.7.13 - 2022-09-08

What's Changed

2.7.12 - 2022-08-24

What's Changed

2.7.11 - 2022-08-11

What's Changed

2.7.10 - 2022-07-27

What's Changed

2.7.9 - 2022-07-27

What's Changed

2.7.8 - 2022-07-24

What's Changed

2.7.7 - 2022-07-19

What's Changed

2.7.6 - 2022-05-22

What's Changed

2.7.5 - 2022-05-20

What's Changed

2.7.4 - 2022-05-06

This release fixes a major bug related to Eigen::Ref when using Row Major matrices.

What's Changed

2.7.3 - 2022-05-02

What's Changed

2.7.2 - 2022-04-22

What's Changed

New Contributors

2.7.1 - 2022-04-09

What's Changed

2.7.0 - 2022-04-02

What's Changed

2.6.11 - 2022-02-25

What's Changed

2.6.10 - 2022-02-02

This new release enhances portability for ROS2 and fixes issues with Eigen::RowMajor matrix types.

What's Changed

2.6.9 - 2021-10-29

Mostly a maintenance release that enhances packaging support and provides fixes for Windows.

What's Changed

2.6.8 - 2021-09-05

This new release: - enhances the compatibility with Boost >= 1.77 - fixes Python doc - adds the support of Python stubs

2.6.7 - 2021-08-19

This new release provides additional support in user-type for extracting class objects exposed through boost python.

2.6.6 - 2021-08-13

This new release provides: - extended support of Custom types with Numpy - extended support of Eigen::{LLT,LDLT} for matrix solution - MINRES solver support

2.6.5 - 2021-07-30

This new release extends the support of custom types registration within NumPy arrays.

2.6.4 - 2021-05-25

This new release fixes some bugs when compiling with -march=native. It also fixes a bug of include orders.

2.6.3 - 2021-04-16

This new release enhances the support of EigenPy with Numpy >= 1.2.0

2.6.2 - 2021-03-28

Add support of boolean matrices.

2.6.1 - 2021-01-20

This new release fixes a bug related to Quaternion initialization in Python.

2.6.0 - 2021-01-04

This new release provides extended support for arbitrary scalar types.

2.5.0 - 2020-08-25

This new release provides a support of Eigen::Ref to Python with shared memory.

2.4.4 - 2020-08-18

This new release fixes a bug encountered when trying to expose non-square matrix. It also improves the packaging support on Conda.

2.4.3 - 2020-07-20

This new release fixes some packaging issues and removes some CMake warnings.

2.4.2 - 2020-07-17

This new release provides extended support to NumPy 1.19 and more.

2.4.1 - 2020-06-09

This new release improves the packaging of the project by removing any dependency to pkg-config when searching for dependencies. In addition, the Windows compatibility is enhanced.

2.4.0 - 2020-05-25

This new release enables: - the exposition of user types inside NumPy - a enhance compatibility with Windows

2.3.2 - 2020-04-23

This new release enforces the compatibility with: - ROS and Python3 - Ubuntu 20.04 - Eigen >= 3.3.90

2.3.1 - 2020-04-08

This new release fixes some packaging issues for OS X systems.

2.3.0 - 2020-04-03

This new release comes with new features: - the Geometry module has now full shared memory with Python

and comes with additional bug fixes: - the project can still be used with former CMake style - the compilations issues with Boost 1.71 has been fixed - on OS X systems, the project can be relocated

2.2.2 - 2020-03-30

This release fixes a packaging with former versions of ROS.

2.2.1 - 2020-03-27

This new release of EigenPy: - removes boring compilation warnings in C++11 - fixes a bug in the Python documentation of the Geometry module - fixes the CMake export of the project with respect to ROS

2.2.0 - 2020-03-18

This new release fully introduces the support of sharing of memory from Eigen to Numpy. This feature can be disabled thanks to "eigenpy.sharedMemory(False)" to recover the previous behavior.

In addition, eigenpy::Refhas been removed, as Eigen::Ref is now fully supported.

2.1.2 - 2020-02-28

This release fixes a packaging bug with older version of CMake.

2.1.1 - 2020-02-27

This new release fixes the support with Windows systems and fixes some bugs with the CMake packaging of the project.

2.1.0 - 2020-02-25

This new release adds the full support for: - sharing the memory between Eigen and Numpy - provides a complete way of exposing Eigen::Ref (with allocation only when needed) - improves the whole efficiency of the code

2.0.3 - 2020-02-20

This new release fixes: - improve the search of Python - fix a potential memory leak when using Eigen::MatrixBase objects

2.0.2 - 2020-02-06

This new release provides some additional supports on Eigen::Matrix to np.array conversion. It also reduces the memory overload when compiling EigenPy.

2.0.1 - 2020-01-31

This new release improves the compatibility of EigenPy with Win32 systems.

2.0.0 - 2020-01-30

This is the new release of EigenPy, which some new features:

  • full support of np.array which becomes the default conversion format. The end-user is still able to switch to numpy.matrix by calling eigenpy.switchToNumpyMatrix()
  • we provide a full exposition of the Eigen decompositions (LLT, LDLT, EigenSolvers, etc.)
  • fixes have been done that were appearing in some very particular cases

1.6.13 - 2020-01-10

This new release uniformizes the build of unitary tests and also provides a new function to set the random seed of the std random generator.

1.6.12 - 2019-12-10

This release fixes the compatibility on Win32 systems and provides a fix for a bug introduced in 1.6.10.

1.6.11 - 2019-12-09

This new release fixes an important bug introduced in 1.6.10.

1.6.10 - 2019-12-09

This new release: - makes the project fully compatible with CMake package policy - improves the documentation of the Python bindings - fixes the convertibility of Eigen base classes

1.6.9 - 2019-11-25

Missing update of the package.xml file.

1.6.8 - 2019-11-25

This new release improves the compatibility version between numpy.array and Eigen::Matrix conversions. It also fixes some compilations issues on Win32 systems.

1.6.7 - 2019-11-15

This new release fixes some compilations of EigenPy on the ROS build farm. Thanks to [@wxmerkt](https://github.com/wxmerkt] for the fix.

1.6.6 - 2019-11-13

This new patched release improves the packaging of the project with respect to external forges.

1.6.5 - 2019-11-08

This new release fixes the export of the Eigen project for CMake targets relying on EigenPy. It also provides some fixes aroung Geometry classes.

1.6.4 - 2019-11-07

This new release improves the packaging of the project with respect to ROS and also fixes some constructors issues with respect to the Quaternion class.

1.6.3 - 2019-10-29

This release fix issues introduced recently in the packaging which was preventing the project to correctly export the main library.

1.6.2 - 2019-10-24

This is a maintenance release that improves the packaging for Windows and ROS. It also fixes some bugs in the bindings of Quaternion.

1.6.1 - 2019-10-16

This new release provides default exposition of Eigen vector and matrix which are common. On the packaging side, the project is now fully compatible with the new CMake rules for defining and calling the PROJECT master function.

1.6.0 - 2019-09-19

This new release makes a step towards removing the conversion of Eigen::Matrix objects into numpy.matrix objects. This new release now warms by default, saying that you need to make an explicit choice between numpy.matrix and numpy.array.

Future major releases will enforce the default case with numpy.array.

1.5.8 - 2019-09-09

This new release mostly removes a useless print when loading a converter.

1.5.7 - 2019-07-19

This new release adds the support of Windows OS (thanks to @seanyen. It also provides some fixes with respect to recent versions of Boost.Python (>= 1.70.0).

1.5.6 - 2019-07-16

This release fixes some compatibility issues with anaconda.

1.5.5 - 2019-07-15

This new release improves the packaging with respect to Boost >= 1.70.0. This is needed for complete integration inside Anaconda.

1.5.4 - 2019-07-13

This new release mostly improves the packaging of the project, also with ROS.

1.5.3 - 2019-06-28

This new release fixes a serious bug with duplicates symbols.

1.5.2 - 2019-06-26

This new release fixes two important bugs: - a type already registered by another library will have a symbolic link in the current scope - when using the array convention, the Eigen vector objects become flatten numpy.array

1.5.1 - 2019-04-16

This new release fixes a serious bug related to Python3 which was occurring when exiting the interpreter. This new release makes also official the new BSD license.

1.5.0 - 2018-10-29

This new release allows now to support the conversion from C++ to Python either as numpy.matrix or as numpy.array.

The use can switch between both features by calling in Python

eigenpy.switchToNumpyArray()

to convert Eigen::Matrix to numpy.array. Or

eigenpy.switchToNumpyMatrix()

to select conversion from Eigen::Matrix to numpy.matrix.

Thanks to [@jviereck](https://github.com/jviereck] for raising the lost of performances induced by numpy.matrix.

1.4.5 - 2018-08-29

Fix compatibility with Eigen 3.3.5

1.4.4 - 2018-07-19

This is a minor release. It allows to register Eigen::MatrixBase<MatType> from Python to Eigen converters.

1.4.3 - 2018-05-12

This release fixes several compatibility issues at the Cmake level for both Python 3 and recent versions of Boost >= 1.67.0.

1.4.2 - 2018-05-02

This new release allows the binding of Eigen::Ref, facilitates the conversion between compatible scalars (eigenpy won't warn if the numpy matrix has integer scalar type). This new release also improves the compatibility with Python 3.x.

1.4.1 - 2018-02-26

This is mostly a maintenance release where some fixes have been done with respect to BSD systems. We also fixed an issue concerning the catching of exceptions: they are now verbose.

1.4.0 - 2018-01-14

This new release introduces an independent eigenpy library that exposes Eigen solvers in Python. It also allows to use directly NumPy memory through famous Eigen::Ref that are just memory mapping.

1.3.3 - 2017-06-09

It is mostly a maintenance release with suppression of warnings and use of correct types.

1.3.2 - 2016-11-21

Summary

This new release allows the check of registration of any class. This allows to not define twice the same symbols. Thanks to the macro defined in memory.hpp, one can now relies on aligned vector and matrices. The unaligned equivalent type is no more required. This release also improves the exposed API of Quaternions and AngleAxis Eigen classes. The install of documentation can be avoided by setting cmake option INSTALL_DOCUMENTATION to OFF.

Bug Fix

  • the Python function _import_array() must be called first before any class to PyArray functions.

1.3.1 - 2016-09-23

Summary

This new release fixes several bugs encounter in the linkage of EigenPy with other libs integrating Python interpreter. It also improve the packaging of the module by removing the dependencies to the Python lib and making Boost Python defined with undefined symbols (useful for OS X - the library is no more static on OS X).

Bug Fix

  • UnalignedEquivalent struct take now an optional Scalar template
  • PyMatrixType is now a singleton, avoiding bad initialisation during dynamic loading of shared lib.

1.3.0 - 2016-02-03

Summary

Alignment of Eigen and Numpy objects is now properly handled. One can now easily expose C++ struct containing Eigen objects in Python avoiding any unnecessary conversion and data are aligned in memory.

Bug Fix

  • Fix issue in the creation of row and column vectors.
  • The library has to be static under OS X to properly expose symbols.

1.2.0 - 2014-11-13

1.1.0 - 2014-09-16

1.0.1 - 2014-07-18

1.0.0 - 2014-07-18

Wiki Tutorials

See ROS Wiki Tutorials for more details.

Source Tutorials

Not currently indexed.

Launch files

No launch files found

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged eigenpy at Robotics Stack Exchange