m-calibrate 0.0.3

A multi-purpose calibration tool.

Multi Calibrate

A multi-purpose calibration tool.

Get Started

pip install m-calibrate

mcal run <my_config.yml>

Local setup

conda create --name mcal python
conda activate mcal

# Editable install
python -m pip install -e .
# Dev dependencies
python -m pip install -e '.[dev]'
# Docs dependencies
python -m pip install -e '.[docs]'
# All dependencies
pip install -e '.[all]'

Running tests

python -m pytest --cov mcal
python -m pytest --cov mcal --slow # With slow tests

# Run full test suite across all versions
# Note: Tox will run slow tests
tox
tox -m single_version

Dev Kubernetes

Install the following:

• Kind
• Kubectl
• Helm

Setup cluster

# Create kind cluster
mcal dev cluster create
# Configure KUBECONFIG to use created cluster
$(mcal dev cluster setup)

Apply needed configurations

mcal dev cluster apply MetricsServer
mcal dev cluster apply NRI
mcal dev cluster apply DaskOperator

Delete cluster after finished

mcal dev cluster delete-all

Releasing

Update the version in pyproject.toml

version='X.Y.Z'

Create a git tag and push

git tag vX.Y.Z
git push --tags

Then create a release via github.

If you mess up and need to edit things

Remove old tag and re-tag

git tag -d vX.Y.Z
git tag vX.Y.Z

git push -f --tags

Delete previous github release and re-create.

TODO:

• Pixie?
• Prometheus