rc-cli 0.2.8

RoseCloud CLI

RoseCloud CLI

The RoseCloud CLI tool. It is intended to be used to setup and manage a RoseCloud workspace for instructors and teaching assistants.

Requirements

• python 3.5+

Development

Setup

With pipenv

1. Install pipenv: sudo pip install pipenv.
2. Run pipenv install. This sets up a python virtual environment for you. All subsequent commands should be run via pipenv run .... If you need to add another dependency, run pipenv install <package_name>.

Link CLI

1. Run pipenv install -e .. This will create a symlink to this project. All changes here will show up in the pipenv.
2. Run pipenv run rc --help.

With pip

1. Run pip install -r requirements.txt This will install all dependencies to your machine. This is if you do not want to mess around with virtual environments and are fine with adding a few dependencies globally to your machine.

Link CLI

1. Run pip install -e .
2. Run rc --help

Things to Keep in Mind

• Update requirements.txt once in a while. Recommend using pigar
• Update setup.py dependencies once in a while.
• Update Docker image once in a while by running
  • docker build -t docker.csse.rose-hulman.edu/rosebuild/rc-cli/pypi-deployer:MAJOR.MINOR.PATCH . --no-cache
  • docker push docker.csse.rose-hulman.edu/rosebuild/rc-cli/pypi-deployer:MAJOR.MINOR.PATCH
• We need test
  • Click supports isolated filesystem and internal invocation

Publishing

CI

You can read below for how to do it manually. To deploy, make sure the repository is cleaned (all files committed and pushed) and add a tag following semantic versioning.

To add a tag:

git fetch --tags --force // fetch all tags
git tag // display a list of tags
git tag MAJOR.MINOR.PATCH // fill in the MAJOR, MINOR, and PATCH
git push --tags // push all uncommitted tags

This will trigger a CI pipeline that will automatically deploy to pypi. You can see if your changes made it here or you can check the CI/CD pipeline on Gitlab.

NOTE We are using setuptools_scm to fetch git tags and use it as our version. setuptools_scm finds the latest tag and attempts to compute the latest PATCH version based on the number of commits (and unstaged changes) since the commit of the tag. This is why it is important that you tag after all files are committed and pushed.

Setup

First, you need to install PyPi's publication tool twine. Run,

pip install twine

You will also need wheel. Run,

pip install wheel

Lastly, ensure you have the latest setuptools by running

pip install --upgrade setuptools

You will need to also setup a PyPi account (link) and a test PyPi account (link).

For ease of use, you can create a file in your home directory called .pypirc with the following. For linux and osx, this would be ~/.pypirc. Twine will use this to request rather than requesting for credentials.

[pypi]
username = <username>
password = <password>

[testpypi]
username = <username>
password = <password>

Short Version

The details are below. If you do not want to do the below or want to remember what you need to do, do this:

1. Update setup.py's version.
2. Run python deploy.py.

Generating Dist Files

Before you begin, make sure python has version 3.5+. The following command will generate dist files that are specific to your python version and some of the features used in this project are python 3.5+ specific.

To generate dist files, you need to run

python setup.py bdist_wheel

Upload New Distribution

If you are uploading changes, you should update the version in the setup.py following semantic versioning principles.

Run the following to upload latest changes to official PyPi servers

twine upload dist/*

You can now install it using

pip install rc-cli

Alternatively, to test without messing with production, run this

twine upload --repository-url https://test.pypi.org/legacy/ dist/*

To install from test,

pip install --index-url https://test.pypi.org/rc-cli rc-cli

More info here on PyPi distribution link.

More info here on test PyPi link.