Mx ocrmypdf

.gitignore

@@ -82,4 +82,13 @@ scripts/import-for-development
 scripts/nuke
 
 # Static files collected by the collectstatic command
-./static/
+static/
+
+.DS*
+
+# consume
+consume/
+
+src/version.txt
+
+export/

.travis.yml

@@ -8,9 +8,9 @@ sudo: false
 
 matrix:
     include:        
-        - python: "3.5"
         - python: "3.6"
         - python: "3.7-dev"
+        - python: "3.8"
         - env:
             - BUILD_DOCKER=1
             # Variable to add to publish the Docker image:

Dockerfile

@@ -1,67 +1,87 @@
-FROM alpine:3.11
+FROM ubuntu:20.04 as builder
 
-LABEL maintainer="The Paperless Project https://github.com/the-paperless-project/paperless" \
-      contributors="Guy Addadi <[email protected]>, Pit Kleyersburg <[email protected]>, \
-        Sven Fischer <[email protected]>"
+ENV LANG=C.UTF-8
+ENV DEBIAN_FRONTEND=noninteractive
 
-# Copy Pipfiles file, init script and gunicorn.conf
-COPY Pipfile* /usr/src/paperless/
-COPY scripts/docker-entrypoint.sh /sbin/docker-entrypoint.sh
-COPY scripts/gunicorn.conf /usr/src/paperless/
+RUN apt-get update && \
+  apt-get upgrade -y && \
+  apt-get install -y --no-install-recommends \
+  build-essential autoconf automake libtool \
+  python3 \
+  python3-venv \
+  python3-setuptools \
+  python3-wheel \
+  pipenv \
+  python3-pip
+
+# Get build deps
+RUN apt-get install -y --no-install-recommends \
+  libpython3.8-dev \
+  libpq-dev \
+  libmariadb-dev \
+  libpoppler-cpp-dev \
+  libxslt-dev \
+  libxml2-dev
+
+# get dependencies
+WORKDIR /usr/src/paperless
+COPY Pipfile* .
+RUN pipenv lock --keep-outdated --requirements > requirements.txt
+
+ENV VIRTUAL_ENV=/usr/src/paperless
+RUN python3 -m venv $VIRTUAL_ENV
+ENV PATH="$VIRTUAL_ENV/bin:$PATH"
+RUN pip3 install -r requirements.txt
+
+FROM jbarlow83/ocrmypdf
+
+ENV LANG=C.UTF-8
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && \
+  apt-get upgrade -y && \
+  apt-get install -y --no-install-recommends \
+  python3.8 \
+  python3-venv \
+  gnupg \
+  imagemagick \
+  optipng \
+  libpoppler-cpp0v5 \
+  sudo \
+  gettext \
+  mariadb-client \
+  libmariadb3 \
+  libmagic1 \
+  curl && \
+  apt-get clean -y && \
+  rm -rf /var/lib/apt/lists/*
 
 # Set export and consumption directories
-ENV PAPERLESS_EXPORT_DIR=/export \
-    PAPERLESS_CONSUMPTION_DIR=/consume
-
-RUN apk add --no-cache \
-      bash \
-      curl \
-      ghostscript \
-      gnupg \
-      imagemagick \
-      libmagic \
-      libpq \
-      optipng \
-      poppler \
-      python3 \
-      shadow \
-      sudo \
-      tesseract-ocr \
-			tzdata \
-      unpaper && \
-    apk add --no-cache --virtual .build-dependencies \
-      g++ \
-      gcc \
-      jpeg-dev \
-      musl-dev \
-      poppler-dev \
-      postgresql-dev \
-      python3-dev \
-      zlib-dev && \
-# Install python dependencies
-    python3 -m ensurepip && \
-    rm -r /usr/lib/python*/ensurepip && \
-    cd /usr/src/paperless && \
-    pip3 install --upgrade pip pipenv && \
-    pipenv install --system --deploy && \
-# Remove build dependencies
-    apk del .build-dependencies && \
-# Create the consumption directory
-    mkdir -p $PAPERLESS_CONSUMPTION_DIR && \
-# Create user
-    addgroup -g 1000 paperless && \
-    adduser -D -u 1000 -G paperless -h /usr/src/paperless paperless && \
-    chown -Rh paperless:paperless /usr/src/paperless && \
-    mkdir -p $PAPERLESS_EXPORT_DIR && \
-# Avoid setrlimit warnings
-# See: https://gitlab.alpinelinux.org/alpine/aports/issues/11122
-    echo 'Set disable_coredump false' >> /etc/sudo.conf && \
+ENV PAPERLESS_EXPORT_DIR=/export
+ENV PAPERLESS_CONSUMPTION_DIR=/consume
+
+# Create the directories and user
+RUN mkdir -p $PAPERLESS_CONSUMPTION_DIR && \
+  mkdir -p $PAPERLESS_EXPORT_DIR && \
+  addgroup --gid 1000 paperless && \
+  adduser --home /usr/src/paperless --disabled-password --gecos "" --uid 1000 --ingroup paperless paperless
+
+RUN echo 'Defaults env_keep += "VIRTUAL_ENV"' >>/etc/sudoers.d/paperless && \
+  echo 'Defaults secure_path=/usr/src/paperless/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/bin' >> /etc/sudoers.d/paperless
+
+COPY --chown=paperless:paperless --from=builder /usr/src/paperless/ /usr/src/paperless
+
 # Setup entrypoint
-    chmod 755 /sbin/docker-entrypoint.sh
+COPY scripts/docker-entrypoint.sh /sbin/docker-entrypoint.sh
+RUN chmod 755 /sbin/docker-entrypoint.sh
+
+# Copy gunicorn.conf
+COPY scripts/gunicorn.conf.py /usr/src/paperless/
 
 WORKDIR /usr/src/paperless/src
 # Mount volumes and set Entrypoint
 VOLUME ["/usr/src/paperless/data", "/usr/src/paperless/media", "/consume", "/export"]
+
 ENTRYPOINT ["/sbin/docker-entrypoint.sh"]
 CMD ["--help"]
 
@@ -70,5 +90,13 @@ COPY src/ /usr/src/paperless/src/
 COPY data/ /usr/src/paperless/data/
 COPY media/ /usr/src/paperless/media/
 
+# setup venv
+ENV VIRTUAL_ENV=/usr/src/paperless
+RUN python3 -m venv $VIRTUAL_ENV
+ENV PATH="$VIRTUAL_ENV/bin:$PATH"
+
+RUN cd /usr/src/paperless/src && \
+  django-admin compilemessages
+
 # Collect static files
 RUN sudo -HEu paperless /usr/src/paperless/src/manage.py collectstatic --clear --no-input

Pipfile

@@ -4,7 +4,7 @@ verify_ssl = true
 name = "pypi"
 
 [packages]
-django = "<2.1,>=2.0"
+django = "<2.3,>=2.0"
 pillow = "*"
 coveralls = "*"
 dateparser = "*"
@@ -13,7 +13,9 @@ django-crispy-forms = "*"
 django-extensions = "*"
 django-filter = "*"
 djangorestframework = "*"
-factory-boy = "*"
+django-admin-list-filter-dropdown = "*"
+django-mysql = "*"
+factory-boy = "<3.0"
 filemagic = "*"
 fuzzywuzzy = {extras = ["speedup"],version = "==0.15.0"}
 gunicorn = "*"
@@ -38,6 +40,8 @@ psycopg2 = "*"
 djangoql = "*"
 whitenoise = "*"
 brotli = "*"
+pikepdf = ">=1.19"
+mysqlclient = "*"
 
 [dev-packages]
 ipython = "*"