Skip to content
/ ofiuco Public

Bazel rules for use poetry lock files with cross-platform Python packages support

License

Notifications You must be signed in to change notification settings

oxidase/ofiuco

Repository files navigation

⛎ Ofiuco Rules for Bazel

Overview

The repository defines Bazel installation rules for multi-platform Python lock files. The major difference to pip rules in rules_python is that Python packages are installed as py_library targets and not as external repositories. This allows to use platform information of resolved Python toolchains and build cross-platform Python artifacts.

Minimum requirements:

  • Bazel 6.x and rules_python with registered Python >= 3.11 toolchain.

Getting started

Import ofiuco as a module

To import ofiuco in your project, you first need to add it to your MODULE.bazel file

bazel_dep(name = "rules_python", version = "0.33.2")

python = use_extension("@rules_python//python/extensions:python.bzl", "python")
python.toolchain(python_version = "3.12")
use_repo(python, "python_3_12")

bazel_dep(name = "ofiuco", version = "0.3.7")

poetry = use_extension("@ofiuco//python:extensions.bzl", "poetry")
poetry.parse(
    name = "poetry",
    lock = "@//path/to:poetry.lock",
)
use_repo(poetry, "poetry")

and Python dependencies can be used as

py_library(
    name = "mylib",
    srcs = ["mylib.py"],
    deps = [
        "@poetry//:package"
    ]
)

Update lock files

A lock file in the workspace can be updated using a host Python interpreter as

python3 -m pip install poetry

poetry update

or using a pre-defined target

load("@ofiuco//python:poetry.bzl", "poetry_update")

poetry_update(
    name = "update_lock",
    toml = "pyproject.toml",
    lock = "poetry.lock",
)

In both cases the host interpreter is used in the latter case poetry package with dependencies is installed as an external repository.

Use in a pre-bzlmod setup

Minimal example which uses the system Python run-time could be as in examples/workspace/WORKSPACE.

Multi-version and multi-repository example which uses Python interpreters is in examples/workspace_multi directory. The test test_multiple_repos_import checks the modules imports priority which is defined by the order of dependencies in deps section. For example, in the following case

            [
                "@repo2//:pytest",
                "@repo1//:pytest",
            ],

pytest will be loaded from the repo2 repository.

Note Mixing different repositories in one deps block may lead to side-effects related to using incompatible versions of transitive dependencies.