Curobo Service

The Curobo Service is a FastAPI http server which calls out to the curobo motion planning library. You can hit the curobo service endpoints with whatever language you like.

Install and run the Curobo Service in a virtual environment

• Navigate to the directory where you would like to clone the repos into and set it as an environment variable.

  cd <repos directory>
  export REPOS_DIR=$(pwd)
  

• Clone the curobo service repo into your repos directory.

  cd $REPOS_DIR
  git clone https://github.com/NVlabs/curobo_service.git
  cd curobo_service
  

• Set up a virtual environment and install the project into it.

  python3 -m venv .venv
  source .venv/bin/activate
  pip install -e .
  

• Navigate to your repos directory and clone curobo into it.

  cd $REPOS_DIR
  git clone https://github.com/NVlabs/curobo.git
  cd curobo
  

• Install curobo and dependencies into the virtual environment.

  pip install torch==2.4.0
  pip install -e .
  

• Return to your curobo_service directory and run the tests.

  cd $REPOS_DIR/curobo_service
  make test
  

• Run the service:

  make run
  

You can verify that the service is running by visiting http://127.0.0.1:10000/docs in your browser.

How the Curobo Service Works

Curobo service uses fastapi, which in turn uses Pydantic for data validation. One file src/nvidia/srl/curobo_service/main.py contains all the endpoints for the service.

A different file src/nvidia/srl/curobo_service/data_models.py contains the pydantic data models (validated dataclasses) for the service.

The server is run with uvicorn in src/nvidia/srl/curobo_service/main.py. See uvicorn for more information.