Skip to content

DHARPA-Project/kiara

Repository files navigation

PyPI status PyPI version PyPI pyversions Build Status Coverage Status Code style

kiara

A data-centric workflow orchestration framework.

Description

Kiara is the data orchestration engine developed by the DHARPA project. It uses a modular approach to let users re-use tried and tested data orchestration pipelines, as well as create new ones from existing building blocks. It also helps you manage your research data, and augment it with automatically-, semi-automatically-, and manually- created metadata. Most of this is not yet implemented.

Development

Requirements

  • Python (version >=3.6 -- some make targets only work for Python >=3.7, but kiara itself should work on 3.6)
  • pip, virtualenv
  • git
  • make
  • direnv (optional)

Prepare development environment

git clone https://github.com/DHARPA-Project/kiara.git
cd kiara
python3 -m venv .venv
source .venv/bin/activate
make init

If you use direnv, you can alternatively do:

git clone https://github.com/DHARPA-Project/kiara.git
cd kiara
cp .envrc.disabled .envrc
direnv allow
make init

Note: you might want to adjust the Python version in .envrc (should not be necessary in most cases though)

make targets

  • init: init development project (install project & dev dependencies into virtualenv, as well as pre-commit git hook)
  • update-modules: update default kiara modules package from git
  • flake: run flake8 tests
  • mypy: run mypy tests
  • test: run unit tests
  • docs: create static documentation pages (under build/site)
  • serve-docs: serve documentation pages (incl. auto-reload) for getting direct feedback when working on documentation
  • clean: clean build directories

For details (and other, minor targets), check the Makefile.

Running tests

> make test
# or
> make coverage

Copyright & license

This project is MPL v2.0 licensed, for the license text please check the LICENSE file in this repository.

  • Copyright (c) 2021, 2022 DHARPA project
  • Copyright (c) 2021, 2022 Markus Binsteiner