🚀 iDynTree 4.0.0 Released

[traversaro]

A new major release of iDynTree was released last week and can be found at https://github.com/robotology/idyntree/releases/tag/v4.0.0 .

Installation

If you use conda, iDynTree 4 is the first release that is already available in conda-forge, so you can install iDynTree directly as:

conda install -c conda-forge idyntree

As conda-forge does not support MATLAB extensions, the MATLAB bindings of iDynTree are now installed in a separe conda package, that continues to be part of the robotology channel:

conda install -c conda-forge -c robotology idyntree-matlab-bindings

If you do not use conda but you prefer to use dependencies installed via apt or vcpkg, the library can be compiled from source, on its own as any other CMake project or via the robotology-superbuild.

iDynTree 4 Changes

The changes in this release can be found in the following. The specific changes in 4.0.0 are quite limited, as several improvements were added incrementally in iDynTree 3.* minor releases, feel free to check their release notes directly at iDynTree release page.

idyntree-yarp-tools

A main difference of iDynTree 4 is that the tools that depend on YARP (urdf2dh, yarprobotstatepublisher, idyntree-sole-gui, idyntree-plotter) have been moved in idyntree-yarp-tools and their use in iDynTree has been deprecated. The option IDYNTREE_COMPILES_YARP_TOOLS has been introduced to disable their compilation. This option is set by default to OFF in iDynTree 4, and will be removed in iDynTree 5.

Beside the tools moved there from iDynTree, the idyntree-yarp-tools repository also contains idyntree-yarp-visualizer, a new simple YARP-based visualizer for the robot state. idyntree-yarp-visualizer is similar to the iCubGui, but with the advantage that robot models can be loaded from URDF file. If you are curious about idyntree-yarp-visualizer, check its docs at https://github.com/robotology/idyntree-yarp-tools/blob/main/src/modules/idyntree-yarp-visualizer/README.md .

Migration to idyntree-yarp-tools

conda

If you conda, to install the idyntree-yarp-tools that used to be installed directly with idyntree, you just need to install the idyntree-yarp-tools from the robotology channel:

conda install -c conda-forge -c robotology idyntree-yarp-tools

robotology-superbuild

If you use the robotology-superbuild, idyntree-yarp-tools is already included in the ROBOTOLOGY_ENABLE_DYNAMICS profile, so you do not need to change anything.
If you use a released version of the robotology-superbuild, note that this change will be included in the 2021.08 release.

Detailed ChangeLog

Added

• Add the visualization of labels in the visualizer (Allow the visualization of labels idyntree#879)

Removed

• Remove headers iDynTree/Core/AngularForceVector3.h, iDynTree/Core/AngularMotionVector3.h, include/iDynTree/Core/ForceVector3.h, iDynTree/Core/LinearForceVector3.h, include/iDynTree/Core/LinearMotionVector3.h, include/iDynTree/Core/MotionVector3.h. They were deprecated in iDynTree 2.0 (Add again deprecated headers idyntree#708, Remove deprecated headers idyntree#885).
• The method ModelVisualization::getWorldModelTransform() was removed, it was deprecated in iDynTree 3.0.1 .

Acknowledgement

Thanks a lot to the contributors (in code, documentation, reviews or suggestions) to this iDynTree release.

[traversaro]

fyi @mfussi66 @randaz81 @elandini84

[traversaro]

Shortly after iDynTree 4.0.0, also iDynTree 4.1.0 and 4.2.0 have been released. In these releases, @GiulioRomualdi added what I think it is a really cool feature: visualization of iDynTree Models (so robots, or any multibody structure) inline in Jupyter Notebook cells, whose content can be modified directly from the Jupyter Notebook using interactive IPyWidgets.

You can see an example of the use of this new visualizer (based on MeshCat, https://github.com/rdeits/meshcat-python):

idyntree_jupyter2.mp4

You can also find an example notebook, with instructions how to quickly install the necessary software using conda in Linux/macOS/Windows in https://github.com/robotology/idyntree/blob/v4.2.0/examples/python/MeshcatVisualizerExample.ipynb . If instead you just want to try the visualizer without installing anything on your system, you can try an example Notebook that works on Google Colab, available in https://colab.research.google.com/github/GiulioRomualdi/idyntree-meshcat-notebook/blob/main/idyntree_meshcat.ipynb (however in Google Colab the visualizer is rendered in a new tab, not inline in Jupyter Cell).

Thanks a lot to @GiulioRomualdi for this new features, that may be extremly useful for reproducible science!