Dockerfile-cluster-base

ARG PYTHON_VERSION=unset

FROM python:${PYTHON_VERSION}-slim

ARG CMAKE_VERSION="3.31.3"
ARG DASK_VERSION=unset
ARG PYTHON_VERSION=unset

ENV \
    DEBIAN_FRONTEND=noninteractive \
    LANG=C.UTF-8 \
    LC_ALL=C.UTF-8

RUN --mount=type=bind,source=bin,target=/build-bin \
<<EOF
apt-get update
apt-get install -y --no-install-recommends \
    build-essential \
    curl \
    libomp-dev

/build-bin/install-cmake "${CMAKE_VERSION}"

python -m pip install \
    --no-cache-dir \
    --prefer-binary \
        blosc \
        bokeh \
        "dask==${DASK_VERSION}" \
        'dask-ml>=2024.4.4' \
        "distributed==${DASK_VERSION}" \
        lz4 \
        numpy \
        'pandas>=2.0.0' \
        scikit-learn

# remove unnecessary files
find \
    /usr/local/lib/python${PYTHON_VERSION}/site-packages \
    -type f \
    \( \
    -name '*.c'                                 \
    -o -name '*.cc'                             \
    -o -name '*.cpp'                            \
    -o -name '*.h'                              \
    -o -name '*.hpp'                            \
    -o -wholename '*bokeh/sampledata/*'         \
    -o -wholename '*dask/*tests/*'              \
    -o -wholename '*joblib/test/*'              \
    -o -wholename '*llvmlite/tests/*'           \
    -o -wholename '*numba/*tests/*'             \
    -o -wholename '*numpy/*tests/*'             \
    -o -wholename '*pandas/tests*'              \
    -o -wholename '*pandas/*/tests/*'           \
    -o -wholename '*psutil/tests/*'             \
    -o -wholename 'pyarrow/_pyarrow_cpp_tests*' \
    -o -wholename '*scikit-learn/tests*'        \
    -o -wholename '*scikit-learn/*/tests*'      \
    -o -wholename '*sklearn/tests*'             \
    -o -wholename '*sklearn/*/tests*'           \
    -o -wholename '*scipy/*/tests*'             \
    -o -wholename '*sparse/*/tests/*'           \
    -o -wholename '*toolz/tests/*'              \
    -o -wholename '*tornado/test/*'             \
    -o -wholename '*zict/tests/*'               \
    -o -wholename '*/__pycache__/*'             \
    \) \
    -exec rm '{}' '+'

find \
    /usr/local/lib/python${PYTHON_VERSION}/site-packages \
    -type d \
    -wholename '*__pycache__*' \
    -exec rm -rf '{}' '+'

# clean apt-get files
apt-get clean
apt-get purge -y --auto-remove
rm -rf /var/lib/apt/lists/*

# clean other files
rm -rf ~/.cache
EOF