ros/urdf_parser_py

Name: urdf_parser_py

Owner: ROS core stacks

Description: Standalone URDF parser for Python.

Created: 2016-02-22 16:53:33.0

Updated: 2018-01-31 22:58:39.0

Pushed: 2018-05-18 00:17:54.0

Homepage: null

Size: 214

Language: Python

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

urdf_parser_py

Development Setup

You must manually run setup.py. For catkin development, you can install to $ws/../build/lib/pythonX.Y/dist-packages via

devel_prefix=$(cd $(catkin_find --first-only)/.. && pwd)
cd ../urdf_parser_py
python setup.py install --install-layout deb --prefix $devel_prefix

Not yet sure how to get it to generate catkin-like development installs, which uses __init__.py to point to the development source directory.

Authors
Reflection (or just Serialization?)

This an attempt to generalize the structure of the URDF via reflection to make it easier to extend. This concept is taken from Gazebo's SDF structure, and was done with SDF in mind to a) make an SDF parser and b) make a simple converter between URDF and SDF.

Changes
Todo
  1. Develop a Python SDF API in a sdf module.
    • Maybe make the package itself be robot_model_py so that the respective modules would be robot_model_py.urdf_parser and robot_model_py.sdf_parser?
    • Parse Gazebo's SDF definition files at some point? For speed's sake, parse it and have it generate code to use?
  2. Make a direct, two-way URDF <-> SDF converter.
    • Gazebo has the ability to load URDFs and save SDFs, but it lumps everything together and (I think) adds some “noise” from OpenDE for positions.
  3. Make the names a little clearer, especially the fact that from_xml and to_xml write to a node, but do not create a new one.
  4. Figure out good policy for handling default methods. If saving to XML, write out default values, or leave them out for brevity (and to leave it open for change)? Might be best to add that as an option.
  5. Find a lightweight package that can handle the reflection aspect more elegantly. Enthought traits? IPython's spinoff of traits?

This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.