Skip to content

Latest commit

 

History

History
64 lines (46 loc) · 1.52 KB

development.md

File metadata and controls

64 lines (46 loc) · 1.52 KB
Raw

Development

This section describes how to install the software within a development sandbox on your workstation, and how to run the software tests.

Install Sandbox

Acquire source code and install development sandbox. The authors recommend to use a Python virtualenv.

git clone https://github.com/crate/mlflow-cratedb
cd mlflow-cratedb
python3 -m venv .venv
source .venv/bin/activate
pip install --editable='.[examples,develop,docs,test]'

Run Software Tests

Run linters and software tests, skipping slow tests:

poe check-fast

Exclusively run "slow" tests.

pytest -m slow

Build OCI Images

If you are aiming to make changes to the software or to the Dockerfile, you can build a local OCI image from the working tree on your workstation.

Run the following commands from the root directory of the project:

export DOCKER_BUILDKIT=1
export COMPOSE_DOCKER_CLI_BUILD=1
export BUILDKIT_PROGRESS=plain

docker build --tag=local/mlflow-cratedb --file=release/oci-server/Dockerfile .
docker build --tag=local/ml-runtime --file=release/oci-runtime/Dockerfile .

Run a Release

git tag v0.x.x
git push && git push --tags

Build sdist and wheel packages, and upload them to PyPI.

poe release

Navigate to the release page on GitHub, and announce the corresponding release, adding the changelog items from the file instead of the items automatically generated by GitHub.