# syntax=docker/dockerfile:labs

FROM python:3.11-alpine AS builder

ARG PAPERLESS_VERSION=2.6.3

WORKDIR /opt

RUN set -euxo pipefail &&\
    apk --no-cache add \
      git \
      make \
      cmake \
      build-base \
      gfortran \
      mariadb-dev \
      libpq-dev \
      freetype-dev \
      imagemagick-dev \
      libxml2-dev \
      openblas-dev \
      qpdf-dev \
      tar \
      xz \
      curl \
      ca-certificates \
      rdfind \
    &&\
    curl -sSLO https://github.com/paperless-ngx/paperless-ngx/releases/download/v${PAPERLESS_VERSION}/paperless-ngx-v${PAPERLESS_VERSION}.tar.xz &&\
    tar xvJf paperless-ngx-v${PAPERLESS_VERSION}.tar.xz &&\
    rm paperless-ngx-v${PAPERLESS_VERSION}.tar.xz &&\
    mv paperless-ngx paperless &&\
    cd paperless &&\
    python3 -m venv venv &&\
    source venv/bin/activate &&\
    pip --no-cache-dir install -r requirements.txt &&\
    ./src/manage.py collectstatic --no-input &&\
    mkdir -p /usr/share/nltk_data &&\
    python3 -c "import nltk; nltk.download('stopwords', download_dir='/usr/share/nltk_data')" &&\
    python3 -c "import nltk; nltk.download('punkt', download_dir='/usr/share/nltk_data')" &&\
    rdfind /opt

FROM python:3.11-alpine
MAINTAINER Daniel Berteaud <dbd@ehtrace.com>

ARG CARETAKERD_VERSION=1.0.8

ENV LANG=fr_FR.utf8 \
    TZ=Europe/Paris \
    PAPERLESS_MODE=all-in-one \
    PATH=/opt/paperless/venv/bin:${PATH} \
    PAPERLESS_REDIS=redis://127.0.0.1:6379/0 \
    PAPERLESS_DBENGINE=postgresql \
    PAPERLESS_DBHOST=127.0.0.1 \
    PAPERLESS_DBPORT=5432 \
    PAPERLESS_DBNAME=paperless \
    PAPERLESS_DBUSER=paperless \
    PAPERLESS_DBPASS=paperless \
    PAPERLESS_CONSUMPTION_DIR=/input \
    PAPERLESS_CONSUMER_POLLING=60 \
    PAPERLESS_DATA_DIR=/data \
    PAPERLESS_MEDIA_ROOT=/data \
    PAPERLESS_TRASH_DIR=/data/trash \
    PAPERLESS_ADMIN_USER=admin \
    PAPERLESS_ADMIN_PASSWORD=password \
    PAPERLESS_STATICDIR=/opt/paperless/static \
    PAPERLESS_FILENAME_FORMAT={created_year}/{created_month}/{title} \
    PAPERLESS_SECRET_KEY=changeme \
    PAPERLESS_URL=https://paperless.example.org \
    PAPERLESS_OCR_LANGUAGE=fra \
    PAPERLESS_TIME_ZONE=Europe/Paris \
    PAPERLESS_CONSUMER_DELETE_DUPLICATES=true \
    PAPERLESS_CONSUMER_RECURSIVE=true \
    PAPERLESS_CONSUMER_SUBDIRS_AS_TAGS=true \
    PAPERLESS_CONSUMER_POLLING=60 \
    PAPERLESS_WEBSERVER_WORKERS=1 \
    PAPERLESS_BIND_ADDR=0.0.0.0 \
    PAPERLESS_PORT=8994 \
    PAPERLESS_ADMIN_USER=admin \
    PAPERLESS_ADMIN_PASSWORD=admin \
    PAPERLESS_ADMIN_EMAIL=admin@localhost

ADD https://git.lapiole.org/nomad/base_tools.git#master /
COPY --from=builder /opt /opt
COPY --from=builder /usr/share/nltk_data /usr/share/nltk_data

RUN set -euxo pipefail &&\
    apk --no-cache add \
      tini \
      curl \
      unpaper \
      font-liberation \
      tesseract-ocr \
      tesseract-ocr-data-osd \
      tesseract-ocr-data-eng \
      tesseract-ocr-data-fra \
      optipng \
      libpq \
      zbar \
      poppler-utils \
      gnupg \
      imagemagick \
      ghostscript \
      qpdf \
      leptonica \
      libxml2 \
      libmagic \
      pngquant \
      zlib \
      openblas \
    &&\
    addgroup --gid 8994 paperless &&\
    adduser --system --ingroup paperless --disabled-password --uid 8994 --home /opt/paperless --shell /sbin/nologin paperless &&\
    mkdir -p /data /input &&\
    chown paperless:paperless /data /input

COPY root/ /
COPY --from=danielberteaud/alpine:24.3-1 /usr/local/bin/minit /usr/local/bin/minit

EXPOSE ${PAPERLESS_PORT}
USER paperless
ENTRYPOINT ["tini", "--", "/entrypoint.sh"]
CMD ["paperless"]