Dockerfile

# Include global ARGs at the dockerfile top
ARG ARCH="x86_64"
ARG WORKDIR_ROOT="/var/task"


FROM python:3.12-bookworm AS builder_global

ARG ARCH
ARG WORKDIR_ROOT
ARG POETRY_NO_INTERACTION
ARG POETRY_VIRTUALENVS_IN_PROJECT
ARG POETRY_VIRTUALENVS_CREATE
ARG POETRY_CACHE_DIR
ARG ZLIB1G="http://ftp.it.debian.org/debian/pool/main/z/zlib/zlib1g_1.3.dfsg-3+b1_amd64.deb"
ENV PYTHONPATH="${WORKDIR_ROOT}:${PYTHONPATH}:/usr/local/lib/python3/dist-packages"
ENV MPLCONFIGDIR=/tmp/matplotlib
ARG USER="999"


RUN echo "ARCH: $ARCH, ARG POETRY_CACHE_DIR: ${POETRY_CACHE_DIR}, ENV PYTHONPATH: $PYTHONPATH, USER: $USER ..."
# RUN groupadd -g 999 python && useradd -r -u 999 -g python python

# Set working directory to function root directory
WORKDIR ${WORKDIR_ROOT}
COPY --chown=python:python requirements.txt ${WORKDIR_ROOT}/

# avoid segment-geospatial exception caused by missing libGL.so.1 library
RUN echo "BUILDER: check libz.s* before start" && ls -l /usr/lib/${ARCH}-linux-gnu/libz.so*
RUN apt update && apt install -y curl ffmpeg libgl1 python3-pip && apt clean
COPY --chown=python:python ./dockerfiles/apt_preferences /etc/apt/preferences
COPY --chown=python:python ./dockerfiles/debian.sources /etc/apt/sources.list.d/debian.sources
RUN apt update && apt install -t trixie zlib1g -y && apt clean
RUN echo "BUILDER: check libz.s* after install from trixie" && ls -l /usr/lib/${ARCH}-linux-gnu/libz.so*

RUN ls -l /etc/apt/sources* /etc/apt/preferences*

# poetry installation path is NOT within ${WORKDIR_ROOT}: not needed for runtime docker image
RUN python3 -m venv ${WORKDIR_ROOT}/.venv
ENV PATH="${WORKDIR_ROOT}/.venv/bin:$PATH"
RUN . ${WORKDIR_ROOT}/.venv/bin/activate && python -m pip install -r ${WORKDIR_ROOT}/requirements.txt

# USER 999


FROM python:3.12-slim-bookworm AS runtime

RUN groupadd -g 999 python && useradd -r -u 999 -g python python

ARG ARCH
ARG WORKDIR_ROOT
ENV PYTHONPATH="${WORKDIR_ROOT}:${WORKDIR_ROOT}/.venv:${PYTHONPATH}:/usr/local/lib/python3/dist-packages"
ENV MPLCONFIGDIR=/tmp/matplotlib
ENV IS_DOCKER_CONTAINER="YES"

ENV VIRTUAL_ENV=${WORKDIR_ROOT}/.venv PATH="${WORKDIR_ROOT}/.venv/bin:$PATH"

RUN apt update && apt install -y ffmpeg && apt clean
RUN echo "COPY --chown=python:python --from=builder_global /usr/lib/${ARCH}-linux-gnu/libGL.so* /usr/lib/${ARCH}-linux-gnu/"
RUN echo "RUNTIME: check libz.s* before upgrade" && ls -l /usr/lib/${ARCH}-linux-gnu/libz.so*
RUN echo "RUNTIME: remove libz.s* to force upgrade" && rm /usr/lib/${ARCH}-linux-gnu/libz.so*
COPY --chown=python:python --from=builder_global /usr/lib/${ARCH}-linux-gnu/libz.so* /usr/lib/${ARCH}-linux-gnu/
COPY --chown=python:python --from=builder_global /lib/${ARCH}-linux-gnu/libexpat.so* /lib/${ARCH}-linux-gnu/
RUN echo "RUNTIME: check libz.s* after copy" && ls -l /usr/lib/${ARCH}-linux-gnu/libz.so*
COPY --chown=python:python --from=builder_global ${WORKDIR_ROOT}/.venv ${WORKDIR_ROOT}/.venv
RUN echo "check ffmpeg files..."
RUN ls -ld /usr/share/ffmpeg || echo "ffpeg folder not found!"
RUN ls -l /usr/bin/ff* || echo "ffpeg bin not found!"
RUN ls -l /usr/share/ffmpeg || echo "ffpeg folder share not found!"
RUN . ${WORKDIR_ROOT}/.venv && which python && pip list

RUN echo "new WORKDIR_ROOT after hidden venv COPY --chown=python:python => ${WORKDIR_ROOT}"
RUN ls -ld ${WORKDIR_ROOT}/
RUN ls -lA ${WORKDIR_ROOT}/

COPY --chown=python:python . ${WORKDIR_ROOT}/.

RUN python --version
RUN pip list
RUN echo "PATH: ${PATH}."
RUN echo "WORKDIR_ROOT: ${WORKDIR_ROOT}."
RUN ls -l ${WORKDIR_ROOT}
RUN ls -ld ${WORKDIR_ROOT}
RUN python -c "import sys; print(sys.path)"
RUN python -c "import epitran"
RUN python -c "import flask"
RUN python -c "import pandas"
RUN python -c "from torch import Tensor"
RUN python -c "import gunicorn"
RUN df -h
RUN ls -l ${WORKDIR_ROOT}/app.py
RUN ls -l ${WORKDIR_ROOT}/static/

USER 999
ENV PATH="${WORKDIR_ROOT}:${WORKDIR_ROOT}/.venv/bin:$PATH"
RUN echo "PATH: $PATH ..."
RUN echo "PYTHONPATH: $PYTHONPATH ..."
RUN echo "MPLCONFIGDIR: $MPLCONFIGDIR ..."

CMD ["gunicorn", "--bind", "0.0.0.0:3000", "app:app"]