Setup CheckMAITE
Supported environment
- OS: Linux x86_64 (CI-tested). macOS x86_64 and arm64 are supported but not CI-tested. The conda lock is Linux x86_64 only.
- Python: 3.11, 3.12, 3.13, and 3.14 with
uvor pip (requires-python = ">=3.11, <3.15"). conda supports 3.11, 3.12, and 3.13 (python = ">=3.11,<3.14"): conda-forgenrtkrequires asimsimdrelease that has no 3.14 build. The committedconda-lock.ymlis resolved for 3.13.checkmaite-pluginsis<3.12on all paths. - GPU: CPU is the supported baseline. CUDA is optional for PyTorch.
- Hardware, architectures, and internet access: see System requirements.
- Container: CheckMAITE is also available as a batch container for Linux
AMD64, with CPU and NVIDIA CUDA variants, built from the repository
Dockerfile. It needs no Python environment on the host. See Run CheckMAITE in a container.
Clone the project repository
Ensure you have the right permissions, and clone the project repository:
git clone https://gitlab.jatic.net/jatic/orchestration-interoperability/checkmaite.git
On macOS, install OpenMP before creating the environment (dataeval → LightGBM
needs libomp.dylib):
brew install libomp
Create environment and install the package
Both uv and conda are supported; uv is recommended.
Option 1: uv environment
To set up a uv environment, install
uv, and build the
environment with:
uv sync
Optional extras:
uv sync --extra reporting # PDF report export (markdown + xhtml2pdf)
uv sync --extra cloud # cloud storage dependencies (aws+gcs+azure)
JATIC tools which are not under active development are distributed through the
separate checkmaite-plugins package.
Use these tools with caution because they may contain bugs or may be
incompatible with the environment in the future.
checkmaite-plugins supports Python <3.12 and is installed directly from
GitLab instead of through a
checkmaite extra so that the PyPI package metadata remains valid.
uv pip install "checkmaite-plugins[unsupported] @ git+https://gitlab.jatic.net/jatic/orchestration-interoperability/checkmaite-plugins.git@main"
If you are developing locally across both repositories:
pip install -e /path/to/checkmaite
pip install -e "/path/to/checkmaite-plugins[unsupported]"
Option 2: conda environment
To set up a conda environment, install
conda
on your machine and build the environment with:
# create env with conda-lock installed
conda create -n checkmaite "conda-lock>=3"
# activate the environment
conda activate checkmaite
# use conda-lock to install dependencies
conda-lock install -n checkmaite conda-lock.yml
# finally, install the `checkmaite` package
pip install . --no-deps