This is the API library of AVstack. It relies on both avstack-core and having access to integrated datasets/simulators. AVstack was published and presented at ICCPS 2023 - find the paper here that accompanies the repository.
The AVstack API unlocks consistent interfacing to some of the most popular and useful autonomous vehicle driving datasets. The goal of the AVstack API is to standardize interfacing between data source providers. This enables better transfer between datasets and/or simulators and helps unify disparate AV conventions.
See avstack-core
First, clone the repositry and submodules.
git clone --recurse-submodules https://github.com/avstack-lab/lib-avstack-api.git
Dependencies are managed with poetry. This uses the pyproject.toml file to create a poetry.lock file.
You must already have downloaded avstack-core and placed it in a location commensurate with this repository's pyproject.toml file. For example, if pyproject.toml says
lib-avstack-core = { path = "../lib-avstack-core", develop = true }
then lib-avstack-core must be placed in the same folder as lib-avstack-api.
AVstack is also compatible with autonomous vehicle simulators.
See the carla-sandbox to get started with a development sandbox.
We have provided a few Jupyter notebooks in the notebooks/ folder. Please refrain from committing large jupyter notebooks in pull requests. Either ignore them or use something like this approach.
To prevent the inflation of the repository due to e.g., image data in the example notebooks, please configure a pre-commit hook to clear the output of the jupyter notebooks. To make this process easier, we've included a hooks directory. By running:
git config core.hooksPath hooks
in the project root, you can use our pre-made pre-commit hook.
See CONTRIBUTING.md for further details.
Copyright 2023 Spencer Hallyburton
AVstack specific code is distributed under the MIT License.