v0.2.0

Fixed

• Only run AWS upload job if secrets are defined.

Added

• Upload to PyPI option.

Changed

• Delete conda recipe if upload to conda is not desired.
• index_package cookiecutter parameter split into upload_conda_package and upload_pypi_package.
• Uploading to Anaconda/PyPI defaults to "n".
• Documentation split into separate pages for setting up and updating projects.

v0.1.0

These initial changes are all relative to the original cookiecutter PyPackage repository, of which this is a fork.
The list is not exhaustive as many changes have been made.

Added

• GitHub actions that rely on City Modelling Lab reusable actions.
• Schema for Cookiecutter input argument config (in cookiecutter.json) + tests added to CI to validate the config and the schema itself.
• Pre-commit hooks and pre-commit CI
• Templates for different types of GitHub issues.
• Cookiecutter contributing guidelines and changelog.
• Dockerfile to generate basic Docker image from project.
• Memory profiling test template.
• Option to include recipe to build project with conda, ready to trigger the appropriate reusable actions to have it built and uploaded to an Anaconda channel (e.g. city-modelling-lab).
• Option to have a Jupyter Notebook directory (under examples) which will be automatically linted, formatted, tested, and rendered in the documentation.

Changed

• Documentation based on Markdown and MKDocs instead of ReStructured text and Sphinx (for improved readability and simpler configuration).
• Python project config moved from setup.py to pyproject.toml.
• Differentiating between repository name (for github), package name (for indexing online), and module name (for importing in python).
  This extends the original use of project_name and project_slug, the latter being difficult to understand.

Removed

• Use of Travis CI.
• Use of Tox.
• Upload to PyPi (with a plan to reintroduce this).
• Possibility to use Argparse for CLI (i.e., Click is now the only option).