Wisc-HCI / urdfForRobotScene

Convert URDFs for use in Robot-Scene

urdfForRobotScene

• Note: To make things easy. I would suggest running all of these commands in one specified directory.To start with, at the directory where you have the scripts, you should have(and will eventually have) the following format:

\urdfForRobotScene
  ...scripts.py
  \robot_name (make this folder before you start!)
      \robot_description
        \all_robots.urdf (You should have it in this directory when you run mesh_lookup_populator.py, You may simply paste the entire robot's folder, the script walks through all folders for you)
      \meshes (make this folder before you start!)
        \all your .STL files
      \visual (make this folder before you start!)
         all .js files generated by JSXToJS.py
         all .GLB files generated by StlToGLB.py
      \collision (make this folder before you start!)
         all .js files generated by JSXToJS.py
         all .GLB files generated by StlToGLB.py
      items_tf.json (generated by urdf_parser.py)

At the end of this pipeline, you may remove all .jsx files (NOT .js) and the \meshes folder. Refer back to the example above when you're running the scripts! Make sure you run the commands in the order they are described below! More elaborate instructions in a demo video here: Link

StlToGlb.py: STL to GLB Converter

This script converts all STL files found in a specified directory (and its subdirectories) to GLB format using Blender's Python API.

Requirements

• Blender: This script is intended to be run within Blender's Python environment.
• Python 3.x: Ensure Blender's Python version matches the script requirements.
• Setup blender to be recognized from your command line by adding it to your system's environment path.

Usage

To run this script, you will need to use Blender's command-line interface. The general command format is as follows:

blender --background --python StlToGlb.py -- [base_search_dir] [base_output_dir]

• blender is the command to run Blender.
• --background tells Blender to run without its GUI.
• --python GlbToJSX.py instructs Blender to execute the given Python script.
• -- is used to indicate that the following arguments should be passed to the script.
• [base_search_dir] is the relative path to the directory where the script will search for STL files.
• [base_output_dir] is the relative path to the directory where the script will save the converted GLB files.

Example

blender --background --python GlbToJSX.py -- /path/to/source /path/to/destination

• Note: The path will be relative to where your current working directory is, i.e. The folder at which you are calling blender.

GlbToJSX.py: GLB to JSX Converter

This script converts all .glb files in a specified folder to .jsx components using the gltfjsx tool.

Requirements

• Node.js and npm: Ensure you have Node.js and npm installed to use npx and gltfjsx.
• gltfjsx: This script relies on the gltfjsx tool, which can be installed via npm.

Installation of gltfjsx

Before running the script, you must install gltfjsx if you haven't already. You can install it globally using npm:

npm install -g gltfjsx

Usage

To use this script, navigate to the directory where the script is located and run it with Python, providing the input and output directories as arguments:

python GlbToJSX.py input_folder output_folder

• input_folder is the directory containing the .glb files you want to convert, relative to where your script is.
• output_folder is the directory where you want the .jsx files to be saved, relative.

Example

python GlbToJSX.py ./models ./jsxComponents

This will convert all .glb files found in ./models to .jsx components and save them in the ./jsxComponents directory.

Notes

You may change the settings of the converter at line 25 where:

command = f"npx gltfjsx {full_file_path} -o {output_file_path}"

Usage details are in this link: https://github.com/pmndrs/gltfjsx

JSXToJS.py: JSX to JS Converter

This Python script is designed to convert .jsx files to .js files with the ability to specify scale and rotation for the models. It searches for .glb file references within the .jsx files, modifies their import statements, and applies transformations according to the provided command-line arguments.

Requirements

• Python 3.x installed on your system.

How to Run the Script

Basic Usage

To run this script, navigate to the directory containing the script and use the following command format:

python JSXToJS.py input_dir output_dir [options]

• input_dir is the directory containing your .jsx files.
• output_dir is the directory where you want the converted .js files to be saved.

Options

• --scale: Specifies the scale of the model in the format [x,y,z]. For example, --scale [1,1,1].
• --rotation: Specifies the rotation of the model in the format [x,y,z], where each component is an angle in radians. For example, --rotation [0,0,0].
• Note : If you do not specify these flags, the script ignores them, and the parameters wont be added. to the JS code.

Example Command:

python JSXToJS.py ../collision_jsx ../models/collision --scale [1,1,1] --rotation [0,0,0]

mesh_lookup_populator.py : Populates a dictionary MeshLookup.js for robot-scene

This script is designed to assist in automating the process of importing mesh files from a Universal Robot Description Format (URDF) file into a JavaScript project. It scans a specified directory for JavaScript (.js) mesh loader files, generates import statements for them, and creates a lookup table that maps mesh filenames found in the URDF file to their corresponding JavaScript imports. This facilitates the dynamic loading of mesh files in web applications or other JavaScript-based projects.

This python script scans a urdf file for all entries of It then automatically writes the imports by appending the filename with .js, then puts the filename as a key and the resulting import as a value. For example,

<mesh filename="package://p_grip_description/meshes/p_grip_2F/back_cover.STL">

results in

import back_cover from "./MeshLoaders/./Lio/visual/back_cover.js"
const MeshLookupTable = {
 "package://p_grip_description/meshes/p_grip_2F/back_cover.STL": back_cover,
}

Requirements

• Python 3.x installed on your system.

Basic Usage

To run this script, navigate to the directory containing the script: Accepts two commands-line arguments:

• The base directory containing the visual subdirectory with the mesh loader JavaScript files.
• The path to the URDF file containing the mesh filenames.

The command to run the script looks like this:

python mesh_lookup_populator.py <base_directory of visual/collision meshes> <base_dir for urdf files> <robot name>

Example:

Assuming your directories look like this:

/project
    /visual
       all mesh .js files generated by JSX to JS Converter
    /collision
       all mesh .js files generated by JSX to JS Converter
    robot.urdf

Your command should be:

python mesh_lookup_populator.py /path/to/project/dir_with_visual_collision /path/to/project/sub_dirs_of_urdfs robot_name

You will then find a file named MeshLookup.js at /path/to/project/dir_with_visual_collision/MeshLookup.js Note: This script is capable of generating both visual and collision Meshes. Simply change line 19 of script to include collisions. Note: Names are unique due to robot_name to prevent conflicts when loading in multiple robots.

urdf_parser.py:Populates a .json file containing all tfs(joints) and items(links) of the robot.

This Python script is designed to convert the contents of a Universal Robot Description Format (URDF) file into a structured JSON format. It specifically extracts joint and link information, including positions, rotations (expressed as quaternions), scales, and colors. The script handles URDF properties for dynamic value replacement and supports conversion from Euler angles to quaternions for rotation representation.

Features:

• Parses URDF files to extract joints and links information.
• Handles dynamic replacement of values defined by xacro:property within the URDF. Dynamically evaluates expressions within braces during runtime. Note that this might cause a security risk due to code injections, so make sure to only evaluate URDFs from trusted sources!
• Converts rotations from Euler angles to quaternions.
• Generates a JSON file containing structured information about the robot's configuration, suitable for use in simulations, visualizations, or further processing.

Requirements:

• Python 3.x installed on your system.
• a urdf file for your robot.

Basic Usage:

To use the script, provide the source URDF file path and the destination JSON file path as command-line arguments:

python urdf_parser.py <source_urdf_file_path> <destination_json_file> <main_urdf_filename>

Parameters:

Parameters

• <source_urdf_file_path>: Path to the URDF file to be parsed.
• <destination_json_file>: Path where the output JSON file will be saved.
• <main_urdf_file>: name of the main file that would define your robot's base link. The script will compare file names and flag a file if it matches.

Example Command:

python urdf_parser.py path/to/robot.urdf path/to/tfs_items.json

Output format:

The output JSON file will contain two main sections: tfs for joint transformations and items for link descriptions. Each section includes detailed information such as position, rotation, scale, and color. Rotation data is provided in quaternion format to facilitate usage in 3D environments.

What next? Go into: the MeshTest directory to test your robot and see if it works. Follow the readme in there!