Install Poetry Action

A Github action for installing and configuring Poetry.

The action installs Poetry, adds executables to the runner system path, and sets relevant Poetry config settings.

Inspired by dschep's archived poetry action.

Usage

If all you need is default Poetry, simply add this to your workflow:

- name: Install Poetry
  uses: snok/install-poetry@v1

If you want to set Poetry config settings, or install a specific version, you can specify inputs:

- name: Install and configure Poetry
  uses: snok/install-poetry@v1
  with:
    version: 1.1.8
    virtualenvs-create: true
    virtualenvs-in-project: false
    virtualenvs-path: ~/my-custom-path
    installer-parallel: true

The action is fully tested for MacOS and Ubuntu runners, on Poetry versions >= 1.1.0.

If you're using this with Windows, see the Running on Windows section.

Defaults

The current default settings are:

version: latest
virtualenvs-create: true
virtualenvs-in-project: false
virtualenvs-path: {cache-dir}/virtualenvs
installer-parallel: true

If you want to make further config changes - e.g., to change one of the experimental Poetry config settings, or just to make changes to the Poetry config after invoking the action - you can do so in a subsequent step, like this:

- uses: snok/install-poetry@v1
- run: poetry config experimental.new-installer false

Workflow examples and tips

This section contains a collection of workflow examples to try and help

Give you a starting point for setting up your own workflows
Demonstrate how to implement caching for performance improvements
Clarify the implications of different settings

Some of the examples are a bit long, so here are some links

Testing
Testing (using an OS matrix)
Codecov upload
Running on Windows
Virtualenv variations
Caching the Poetry installation

Testing

A basic example workflow for running your test-suite can be structured like this.

name: test

on: pull_request

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      #----------------------------------------------
      #       check-out repo and set-up python
      #----------------------------------------------
      - name: Check out repository
        uses: actions/checkout@v2
      - name: Set up python
        uses: actions/setup-python@v2
        with:
          python-version: 3.9
      #----------------------------------------------
      #  -----  install & configure poetry  -----
      #----------------------------------------------
      - name: Install Poetry
        uses: snok/install-poetry@v1
        with:
          virtualenvs-create: true
          virtualenvs-in-project: true
          installer-parallel: true

      #----------------------------------------------
      #       load cached venv if cache exists
      #----------------------------------------------
      - name: Load cached venv
        id: cached-poetry-dependencies
        uses: actions/cache@v2
        with:
          path: .venv
          key: venv-${{ runner.os }}-${{ hashFiles('**/poetry.lock') }}
      #----------------------------------------------
      # install dependencies if cache does not exist
      #----------------------------------------------
      - name: Install dependencies
        if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
        run: poetry install --no-interaction --no-root
      #----------------------------------------------
      # install your root project, if required
      #----------------------------------------------
      - name: Install library
        run: poetry install --no-interaction
      #----------------------------------------------
      #              run test suite
      #----------------------------------------------
      - name: Run tests
        run: |
          source .venv/bin/activate
          pytest tests/
          coverage report

Testing using a matrix

A more extensive example for running your test-suite on combinations of multiple operating systems, python versions, or package-versions, can be structured like this.

The linting job has nothing to do with the matrix, and is only included for inspiration.

name: test

on: pull_request

jobs:
  linting:
    runs-on: ubuntu-latest
    steps:
      #----------------------------------------------
      #       check-out repo and set-up python
      #----------------------------------------------
      - uses: actions/checkout@v2
      - uses: actions/setup-python@v2
      #----------------------------------------------
      #        load pip cache if cache exists
      #----------------------------------------------
      - uses: actions/cache@v2
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pip
          restore-keys: ${{ runner.os }}-pip
      #----------------------------------------------
      #          install and run linters
      #----------------------------------------------
      - run: python -m pip install black flake8 isort
      - run: |
          flake8 .
          black . --check
          isort .
  test:
    needs: linting
    strategy:
      fail-fast: true
      matrix:
        os: [ "ubuntu-latest", "macos-latest" ]
        python-version: [ "3.6", "3.7", "3.8", "3.9" ]
        django-version: [ "2.2", "3.0", "3.1" ]
    runs-on: ${{ matrix.os }}
    steps:
      #----------------------------------------------
      #       check-out repo and set-up python
      #----------------------------------------------
      - name: Check out repository
        uses: actions/checkout@v2
      - name: Set up python ${{ matrix.python-version }}
        uses: actions/setup-python@v2
        with:
          python-version: ${{ matrix.python-version }}
      #----------------------------------------------
      #  -----  install & configure poetry  -----
      #----------------------------------------------
      - name: Install Poetry
        uses: snok/install-poetry@v1
        with:
          virtualenvs-create: true
          virtualenvs-in-project: true
      #----------------------------------------------
      #       load cached venv if cache exists
      #----------------------------------------------
      - name: Load cached venv
        id: cached-poetry-dependencies
        uses: actions/cache@v2
        with:
          path: .venv
          key: venv-${{ runner.os }}-${{ hashFiles('**/poetry.lock') }}
      #----------------------------------------------
      # install dependencies if cache does not exist
      #----------------------------------------------
      - name: Install dependencies
        if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
        run: poetry install --no-interaction --no-root
      #----------------------------------------------
      # install your root project, if required
      #----------------------------------------------
      - name: Install library
        run: poetry install --no-interaction
      #----------------------------------------------
      #    add matrix specifics and run test suite
      #----------------------------------------------
      - name: Install django ${{ matrix.django-version }}
        run: |
          source .venv/bin/activate
          pip install "Django==${{ matrix.django-version }}"
      - name: Run tests
        run: |
          source .venv/bin/activate
          pytest tests/
          coverage report

Codecov upload

This section contains a simple codecov upload. See the codecov action for more information.

name: coverage

on:
  push:
    branches:
      - master

jobs:
  codecov:
    runs-on: ubuntu-latest
    steps:
      #----------------------------------------------
      #       check-out repo and set-up python
      #----------------------------------------------
      - uses: actions/checkout@v2
      - uses: actions/setup-python@v2
        with:
          python-version: 3.9
      #----------------------------------------------
      #  -----  install & configure poetry  -----
      #----------------------------------------------
      - name: Install Poetry
        uses: snok/install-poetry@v1
        with:
          virtualenvs-create: true
          virtualenvs-in-project: true
      #----------------------------------------------
      #       load cached venv if cache exists
      #----------------------------------------------
      - name: Load cached venv
        id: cached-poetry-dependencies
        uses: actions/cache@v2
        with:
          path: .venv
          key: venv-${{ runner.os }}-${{ hashFiles('**/poetry.lock') }}
      #----------------------------------------------
      # install dependencies if cache does not exist
      #----------------------------------------------
      - name: Install dependencies
        if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
        run: poetry install --no-interaction --no-root
      #----------------------------------------------
      # install your root project, if required
      #----------------------------------------------
      - name: Install library
        run: poetry install --no-interaction
      #----------------------------------------------
      #    run test suite and output coverage file
      #----------------------------------------------
      - name: Test with pytest
        run: poetry run pytest --cov=<project-dir> --cov-report=xml
      #----------------------------------------------
      #             upload coverage stats
      # (requires CODECOV_TOKEN in repository secrets)
      #----------------------------------------------
      - name: Upload coverage
        uses: codecov/codecov-action@v1
        with:
          token: ${{ secrets.CODECOV_TOKEN }}  # Only required for private repositories
          file: ./coverage.xml
          fail_ci_if_error: true

Running on Windows

Running this action on Windows is supported, but two things are important to note:

You need to set the job-level default shell to bash
```
defaults:
  run:
    shell: bash
```
If you are running an OS matrix, and want to activate your venv in-project you have to deal with different folder structures on different operating systems. To make this work, you can do this
```
- run: |
    source .venv/scripts/activate
    pytest --version
  if: runner.os == 'Windows'
- run: |
    source .venv/bin/activate
    pytest --version
  if: runner.os != 'Windows'
```
but we think this is an annoying way to have to structure our workflows, so we set a custom environment variable, $VENV which will point to the OS-specific venv activation script, whether you're running UNIX or Windows. This means you can do this instead
```
- run: |
    source $VENV
    pytest --version
```

For context, a full os-matrix using windows-latest could be set up like this:

name: test

on: pull_request

jobs:
  test-windows:
    strategy:
      matrix: [ "ubuntu-latest", "macos-latest", "windows-latest" ]
    defaults:
      run:
        shell: bash
    runs-on: ${{ matrix.os }}
    steps:
      - name: Check out repository
        uses: actions/checkout@v2
      - name: Set up python
        uses: actions/setup-python@v2
        with:
          python-version: 3.9
      - name: Install Poetry
        uses: snok/install-poetry@v1
        with:
          virtualenvs-create: true
          virtualenvs-in-project: true
      - name: Load cached venv
        id: cached-pip-wheels
        uses: actions/cache@v2
        with:
          path: ~/.cache
          key: venv-${{ runner.os }}-${{ hashFiles('**/poetry.lock') }}
      - name: Install dependencies
        run: poetry install --no-interaction --no-root
      - name: Install library
        run: poetry install --no-interaction
      - run: |
          source $VENV
          pytest --version

Caching on Windows runners

For some reason, caching your venv does not seem to work as expected on Windows runners. You can see an example of what happens here, where a workflow stalls and runs for over 3 hours before it was manually cancelled.

If you do want to cache your dependencies on a Windows runner, you should look into caching your pip wheels instead of your venv; this seems to work fine.

Virtualenv variations

All of the examples we've added use these Poetry settings

- name: Install Poetry
  uses: snok/install-poetry@v1
  with:
    virtualenvs-create: true
    virtualenvs-in-project: true

While this should work for most, and we generally prefer creating our virtualenvs in-project to make the caching step as simple as possible, there are valid reasons for not wanting to construct a venv in your project directory.

There are two other relevant scenarios:

Creating a venv, but not in the project dir

If you're using the default settings, the venv location changes from .venv to using {cache-dir}/virtualenvs. You can also change the path to whatever you'd like. Generally though, this can make things a little tricky, because the directory will be vary depending on the OS, making it harder to write OS agnostic workflows.

A solution to this is to bypass this issue completely by taking advantage of Poetry's poetry run command.

Using the last two steps in the Matrix testing example as an example, this is how we have otherwise documented installing a matrix-specific dependency and running the test suite:
```
- name: Install django ${{ matrix.django-version }}
  run: |
    source .venv/bin/activate
    pip install "Django==${{ matrix.django-version }}"
- name: Run tests
  run: |
    source .venv/bin/activate
    pytest tests/
    coverage report
```
With a remote venv you can do this instead:
```
- name: Install django ${{ matrix.django-version }}
  run: poetry add "Django==${{ matrix.django-version }}"
- name: Run tests
  run: |
    poetry run pytest tests/
    poetry run coverage report
```
We have never needed to cache remote venvs in our workflows. If you have, feel free to submit a PR explaining how it's done.
Skipping venv creation

If you want to skip venv creation, all the original examples are made valid by simply removing the venv activation line: source .venv/bin/activate.

To enable caching in this case, you will want to set up something resembling the the linting job caching step in the Matrix testing; caching your pip wheels rather than your installed dependencies.

Since you're not caching your whole venv, you will need to re-install dependencies every time you run the job; caching will, however, still save you the time it would take to download the wheels (and it will reduce the strain on PyPi).

Caching the Poetry installation

In addition to caching your python dependencies you might find it useful to cache the Poetry installation itself. This should cut ~10 seconds of your total runtime and roughly 95% of this action's runtime.

name: test

on: pull_request

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - name: Check out repository
        uses: actions/checkout@v2
      - name: Set up python
        uses: actions/setup-python@v2
        with:
          python-version: 3.9
      - name: Load cached Poetry installation
        uses: actions/cache@v2
        with:
          path: ~/.local  # the path depends on the OS
          key: poetry-0  # increment to reset cache
      - name: Install Poetry
        uses: snok/install-poetry@v1

The directory to cache will depend on the operating system of the runner.

Contributing

Contributions are always welcome; submit a PR!

License

install-poetry is licensed under an MIT license. See the license file for details.

Showing your support

Leave a ⭐️ if this project helped you!

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

GitHub Action

Install Poetry Action