Docker with uv
最近の開発で、Pythonのパッケージ管理ツールを uv に切り替えて、それに伴うDockerイメージの最適化方法を知りたいです。
人気のあるレポジトリ FastAPIテンプレート
にあるDockerfileを勉強材料にしました。
これからステップ・バイ・ステップで、Dockerfileの分析を行います。
1
2
| FROM python:3.10
ENV PYTHONUNBUFFERED=1
|
PYTHONUNBUFFEREDを空以外の値に設定すると、Python標準出力stdout・標準エラーstderrストリームのバッファリングを行わず、
すぐにターミナルに書き出される。
これはなにかのエラーになるとき、デバッグしやすくなります。
1
2
3
4
5
6
7
8
9
10
11
12
13
| WORKDIR /app/
# Install uv
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#installing-uv
COPY --from=ghcr.io/astral-sh/uv:0.5.11 /uv /uvx /bin/
# Place executables in the environment at the front of the path
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#using-the-environment
ENV PATH="/app/.venv/bin:$PATH"
# Compile bytecode
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#compiling-bytecode
ENV UV_COMPILE_BYTECODE=1
|
UV_COMPILE_BYTECODEを設定することにより、パッケージをインストールする時に、.pyをバイトコード.pycファイルにビルドされる。
起動時間を短縮することができそうです。
1
2
3
| # uv Cache
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#caching
ENV UV_LINK_MODE=copy
|
公式サイトにより、この設定はハードリンクが使えない場合にuvが警告しないようにするものです。
1
2
3
4
5
6
| # Install dependencies
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#intermediate-layers
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=bind,source=uv.lock,target=uv.lock \
--mount=type=bind,source=pyproject.toml,target=pyproject.toml \
uv sync --frozen --no-install-project
|
Dockerのキャッシュを利用するため、uv syncを実行するとき、パッケージのみインストールし、プロジェクトファイルを無視する。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
| ENV PYTHONPATH=/app
COPY ./scripts /app/scripts
COPY ./pyproject.toml ./uv.lock ./alembic.ini /app/
COPY ./app /app/app
COPY ./tests /app/tests
# Sync the project
# Ref: https://docs.astral.sh/uv/guides/integration/docker/#intermediate-layers
RUN --mount=type=cache,target=/root/.cache/uv \
uv sync
CMD ["fastapi", "run", "--workers", "4", "app/main.py"]
|
すべてのパッケージがインストールされた後で、プロジェクト関連ファイルをコピーし、再度uv syncで同期する。
最後の部分をカスタマイズしたら、自分のプロジェクトに利用することができます。