Major overhaul of the Dockerfile

Switched from Ubuntu to debian:stretch because stretch has more recent
versions of our binary packages and starts smaller.  In particular,
stretch has both pillow==2.9.0 and reportlab==3.2.0 available as system
packages which saves the considerable hassle of install a toolchain.

Instead, a pyvenv is set up with access to system's site-packages (note:
needs two steps), making the binary-dependent packages available.  Then
the remaining packages are installed into the pyvenv with --no-cache-dir
to avoid saving files. And there we are.

Image is still very large (>500 MB), but programs like reportlab require
font rendering capabilities so they pull in large portions of the Linux
graphics stack. Not much will shrink that.

Dockerfile

@@ -1,7 +1,7 @@
 # OCRmyPDF
 #
 # VERSION               3.0.0
-FROM      ubuntu:14.04
+FROM      debian:stretch
 MAINTAINER James R. Barlow <jim@purplerock.ca>
 
 # Add unprivileged user
@@ -9,13 +9,8 @@ RUN useradd docker \
   && mkdir /home/docker \
   && chown docker:docker /home/docker
 
-# Update system
+# Update system and install our dependencies
 RUN apt-get update && apt-get install -y --no-install-recommends \
-  bc \
-  curl \
-  wget \
-  zlib1g-dev \
-  libjpeg-dev \
   ghostscript \
   tesseract-ocr \
   tesseract-ocr-deu tesseract-ocr-spa tesseract-ocr-eng tesseract-ocr-fra \
@@ -24,39 +19,29 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
   poppler-utils \
   python3 \
   python3-pip \
-  python3-pytest \
-  python3-dev \
-  gcc
+  python3-venv \
+  python3-reportlab \
+  python3-pil
 
-
-# Ubuntu 14.04's ensurepip is broken, so it cannot create py3 virtual envs
-# Use elaborate workaround: create a venv without pip, activate that environment,
-# download a script to install pip into the environment
-# http://www.thefourtheye.in/2014/12/Python-venv-problem-with-ensurepip-in-Ubuntu.html
-
-RUN python3 -m venv appenv --without-pip
-RUN . /appenv/bin/activate; \
-  wget -O - -o /dev/null https://bootstrap.pypa.io/get-pip.py | python
-
-COPY ./*requirements.txt /application/
-
-# Build wheels separately so build is cached by Docker
-RUN . /appenv/bin/activate; \
-  pip install wheel \
-  && pip wheel --wheel-dir=/wheelhouse -r /application/requirements.txt \
-  && pip wheel --wheel-dir=/wheelhouse -r /application/test_requirements.txt
+# Set up a Python virtualenv and take all of the system packages, so we can
+# rely on the platform packages rather than importing GCC and compiling them
+RUN pyvenv /appenv \
+  && pyvenv --system-site-packages /appenv
 
 COPY . /application/
 
-# Install application using our built wheels
+# Install application and dependencies
+# In this arrangement Pillow and reportlab will be provided by the system
 RUN . /appenv/bin/activate; \
-  pip install --no-index --find-links=/wheelhouse /application \
-  && pip install --no-index --find-links=/wheelhouse -r /application/test_requirements.txt
+  pip install --upgrade pip \
+  && pip install --no-cache-dir /application \
+  && pip install --no-cache-dir -r /application/test_requirements.txt
 
 USER docker
 WORKDIR /home/docker
 
 ENV DEFAULT_RUFFUS_HISTORY_FILE=/tmp/.{basename}.ruffus_history.sqlite
+ENV OCRMYPDF_TEST_OUTPUT=/tmp/test-output
 
 # Must use array form of ENTRYPOINT
 # Non-array form does not append other arguments, because that is "intuitive"

setup.py

@@ -158,7 +158,7 @@ if command.startswith('install') or \
     )
     check_external_program(
         program='unpaper',
-        need_version='6.1',
+        need_version='0.4.2',
         package='unpaper',
         optional=True
     )

test_requirements.txt

@@ -1,2 +1,2 @@
 img2pdf>=0.1.5
-pytest>=2.7.2
+pytest>=2.7.2

tests/test_main.py

@@ -7,7 +7,6 @@ import os
 import shutil
 from contextlib import suppress
 import sys
-from unittest.mock import patch, create_autospec
 import pytest
 from ocrmypdf.pageinfo import pdf_get_all_pageinfo
 import PyPDF2 as pypdf
@@ -22,7 +21,9 @@ TESTS_ROOT = os.path.abspath(os.path.dirname(__file__))
 PROJECT_ROOT = os.path.dirname(TESTS_ROOT)
 OCRMYPDF = os.path.join(PROJECT_ROOT, 'OCRmyPDF.sh')
 TEST_RESOURCES = os.path.join(PROJECT_ROOT, 'tests', 'resources')
-TEST_OUTPUT = os.path.join(PROJECT_ROOT, 'tests', 'output')
+TEST_OUTPUT = os.environ.get(
+    'OCRMYPDF_TEST_OUTPUT',
+    default=os.path.join(PROJECT_ROOT, 'tests', 'output'))
 TEST_BINARY_PATH = os.path.join(TEST_OUTPUT, 'fakebin')