Piton
View Source Piton is a library which will help you to run your Python code.
You can implement your own Piton.Port and run your python code but I highly recommend to use Piton.Pool,
a pool which will allow to run Python code in parallel, a way of avoiding the GIL, and it will protect you from
python exceptions.
Requirements
- Elixir ~> 1.19
- Python 3.x
- Erlang/OTP 27
Installation
Add piton to your list of dependencies in mix.exs:
def deps do
[{:piton, "~> 0.5.0"}]
endHow to use it
Define your own port
The Easiest one
defmodule MySimplePort do use Piton.Port endA port with some wrapper functions which will help you to call the python function: YOUR_MODULE.execute(pid, python_module, python_function, list_of_arguments)
defmodule MyCustomPort do use Piton.Port def start_link(), do: MyCustomPort.start_link([path: Path.expand("python_folder"), python: "python"], [name: __MODULE__]) def fun(n), do: MyCustomPort.execute(__MODULE__, :functions, :fun, [n]) endA port prepared to be run by
Piton.PoolThey have to have a function start() and it has not to be linked.defmodule MyPoolPort do use Piton.Port def start(), do: MyPoolPort.start([path: Path.expand("python_folder"), python: "python"], []) def fun(pid, n), do: MyPoolPort.execute(pid, :functions, :fun, [n]) end
Run a Pool
Pay attention to the number of Pythons you want to run in parallel. It does not exist an optimal number, maybe it is the number of cores, maybe half or maybe double. Test it with your application.
{:ok, pool} = Piton.Pool.start_link([module: MyPoolPort, pool_number: pool_number], [])Call a Port (No pool)
iex> MyCustomPort.execute(pid_of_the_port, python_module, python_function, list_of_arguments_of_python_function)Call a Pool
iex> Piton.Pool.execute(pid_of_the_pool, elixir_function, list_of_arguments_of_elixir_function)Tasks
Some Mix.Tasks have been included in order to facilitate the integration of a python project
Mix.Tasks.Piton.Venv: It creates a Python Virtual Environment.Mix.Tasks.Piton.Pip: It upgrades the Python pip.Mix.Tasks.Piton.Requirements: It gets the dependencies of the Python project.
Test
Before running tests, create a Python virtual environment:
python3 -m venv test/venv
Then run the tests:
mix test
CI/CD
This project uses GitHub Actions for continuous integration and deployment:
Continuous Integration (CI)
Every push and pull request triggers automated tests.
Workflow: .github/workflows/ci.yml
The CI workflow:
- ✅ Runs on Elixir 1.19.2 and OTP 27.1
- ✅ Compiles code with warnings as errors
- ✅ Runs full test suite
- ✅ Checks code formatting
- ✅ Caches dependencies for faster builds
Triggers: All branches on push and all pull requests
Automatic Publishing to Hex.pm
When you create a new GitHub release, the package is automatically published to Hex.pm.
Workflow: .github/workflows/publish.yml
The workflow:
- ✅ Runs on Elixir 1.19.2 and OTP 27.1
- ✅ Installs dependencies with caching
- ✅ Runs tests before publishing
- ✅ Automatically publishes to Hex.pm on release
To publish a new version:
- Update the version in
mix.exs - Update
CHANGELOG.mdwith changes - Commit and push changes
- Create a new GitHub release (e.g.,
v0.5.0) - The package will be automatically published to Hex.pm
Required Secret: HEX_KEY - Your Hex.pm API key must be added to GitHub repository secrets.
Name
Pitón is only Python in Spanish :stuck_out_tongue_winking_eye: :snake: