rules_python_poetry
Bazel rules to install Python dependencies from a Poetry project. Works with native Python rules for Bazel.
Getting started
Add the following to your WORKSPACE
file:
load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
http_archive(
name = "rules_python",
url = "https://github.com/bazelbuild/rules_python/releases/download/0.1.0/rules_python-0.1.0.tar.gz",
sha256 = "b6d46438523a3ec0f3cead544190ee13223a52f6a6765a29eae7b7cc24cc83a0",
)
http_archive(
name = "rules_python_poetry",
url = "https://github.com/martinxsliu/rules_python_poetry/archive/v0.1.0.tar.gz",
sha256 = "8f0abc58a8fcf75341b4615c6b7d9bb254119577629f45c2b1bb60f60f31b301",
strip_prefix = "rules_python_poetry-0.1.0"
)
load("@rules_python_poetry//:defs.bzl", "poetry_install_toolchain", "poetry_install")
# Optional, if you want to use a specific version of Poetry (1.0.10 is the default).
poetry_install_toolchain(poetry_version = "1.1.4")
poetry_install(
name = "my_deps",
pyproject_toml = "//path/to:pyproject.toml",
poetry_lock = "//path/to:poetry.lock",
dev = True, # Optional
)
Under the hood, poetry_install
uses Poetry to export a requirements.txt
file which is then passed to rule_python
's pip_install
repository rule. You can consume dependencies the same way as you would with pip_install
, e.g.:
load("@my_deps//:requirements.bzl", "requirement")
py_library(
name = "my_lib",
srcs = ["my_lib.py"],
deps = [
":my_other_lib",
requirement("some_pip_dep"),
requirement("another_pip_dep[some_extra]"),
],
)
Poetry dependencies
Poetry allows you to specify dependencies from different types of sources that are not automatically fetched and installed by the poetry_install
rule. You will have to manually declare these dependencies.
See tests/multi/app
for examples.
Local directory dependency
A dependency on a local directory, for example if you have multiple projects within a monorepo that depend on each other.
[tool.poetry.dependencies]
foo = {path = "../libs/foo"}
If the local dependency has a py_library
target, you can include it in the deps
attribute.
py_library(
name = "my_lib",
srcs = ["my_lib.py"],
deps = [
"//path/to/libs:foo",
],
)
Local file dependency
A dependency on a local tarball, for example if you have vendored packages.
[tool.poetry.dependencies]
foo = {path = "../vendor/foo-1.2.3.tar.gz"}
There are some options available. The first is to extract the archive and vendor the extracted files. Then add a py_library
that can be included as a deps
, like the local directory dependency.
The second is to use the py_archive
repository rule to declare the archive as an external repository in your WORKSPACE
file, e.g.:
load("@rules_python_poetry//:defs.bzl", "py_archive")
py_archive(
name = "foo",
archive = "//path/to/vendor:foo-1.2.3.tar.gz",
strip_prefix = "foo-1.2.3",
)
The py_archive
rule defines a target named :py_library
that can be referenced like so:
py_library(
name = "my_lib",
srcs = ["my_lib.py"],
deps = [
"@foo//:py_library",
],
)
URL dependency
A dependency on a remote archive.
[tool.poetry.dependencies]
foo = {url = "https://example.com/packages/foo-1.2.3.tar.gz"}
You can use the py_archive
repository rule to declare the remote archive as an external repository in your WORKSPACE
file, e.g.:
load("@rules_python_poetry//:defs.bzl", "py_archive")
py_archive(
name = "foo",
url = "https://example.com/packages/foo-1.2.3.tar.gz",
sha256 = "...",
strip_prefix = "foo-1.2.3",
)
The py_archive
rule defines a target named :py_library
that can be referenced like so:
py_library(
name = "my_lib",
srcs = ["my_lib.py"],
deps = [
"@foo//:py_library",
],
)
Git dependency
Git dependencies are not currently supported. You can work around this by using a URL dependency instead of a git key.