Merge pull request #54 from ifad/feature/docker

Add Docker support for local development

.gitignore

@@ -7,3 +7,6 @@ tmp/*
 spec/output.log
 .byebug_history
 storage/
+
+# Docker
+docker/colore/variables.env

README.md

@@ -381,6 +381,33 @@ ARGS="$@"
 exec java -jar /usr/share/java/tika-app.jar $ARGS
 ```
 
+## Docker
+
+A multi-container Docker application defined in `docker-compose.yml` is
+provided to speed up Colore installation on development machines.
+
+1. Copy `docker/colore/variables.env.example` to `docker/colore/variables.env`
+2. Run `docker compose up`
+3. Have fun with Colore up and running on `http://localhost:9240/`
+
+Note that some features of Colore, like synchronous HTML to PDF conversion, may
+need to access stylesheets and JavaScripts on the host machine. In such cases,
+please make sure that:
+
+1. The web application server allows for concurrent HTTP connections
+2. The web application generates absolute urls to stylesheets and JavaScripts
+   (`config.asset_host` on a Rails application)
+3. The container is capable of reaching assets on the host machine (it is
+   possible to use `host.docker.internal` on macOS)
+
+### Run tests in the container
+
+It is possible to run specs in the container by executing:
+
+```
+$ docker compose run colore bundle exec rspec
+```
+
 ## Security notes
 
 Colore does not implement any authentication mechanism. It is up to you to

docker-compose.yml

@@ -0,0 +1,58 @@
+services:
+  colore:
+    platform: linux/amd64
+    build:
+      context: .
+      dockerfile: ./docker/colore/Dockerfile
+      args:
+        RACK_ENV: development
+    container_name: colore
+    depends_on:
+      - redis
+      - sidekiq
+    env_file:
+      - ./docker/colore/variables.env
+    environment:
+      RACK_ENV: development
+    networks:
+      - colore
+    ports:
+      - 9240:9240
+    restart: on-failure
+    volumes:
+      - storage:/colore/storage
+  redis:
+    container_name: redis
+    image: redis:5.0-alpine
+    networks:
+      - colore
+    restart: on-failure
+    volumes:
+      - redis:/data
+  sidekiq:
+    platform: linux/amd64
+    build:
+      context: .
+      dockerfile: ./docker/colore/Dockerfile
+      args:
+        RACK_ENV: development
+    command: bundle exec sidekiq -C config/sidekiq.yml -r ./lib/sidekiq_app.rb
+    container_name: sidekiq
+    depends_on:
+      - redis
+    env_file:
+      - ./docker/colore/variables.env
+    environment:
+      RACK_ENV: development
+    networks:
+      - colore
+    restart: on-failure
+    volumes:
+      - storage:/colore/storage
+
+volumes:
+  storage:
+  redis:
+
+networks:
+  colore:

docker/colore/Dockerfile

@@ -0,0 +1,91 @@
+FROM ruby:2.6.10
+
+RUN apt update && apt install -y \
+  build-essential \
+  imagemagick \
+  libmagic-dev \
+  tesseract-ocr
+
+# Needed to get the latest libreoffice
+# Ref: https://wiki.debian.org/LibreOffice#Using_Debian_backports
+RUN echo 'deb http://deb.debian.org/debian bullseye-backports main contrib non-free' >> /etc/apt/sources.list
+RUN apt update && apt install -y -t bullseye-backports libreoffice
+
+# Please keep using version 0.12.3
+# With newer versions of wkhtmltopdf, wkhtmltopdf/wkhtmltopdf#1524 and
+# wkhtmltopdf/wkhtmltopdf#3241 will affect Colore's PDF output
+# TODO: implement PDF comparison specs and update this library
+ARG WKHTMLTOPDF_VERSION=0.12.3
+ARG WKHTMLTOPDF_MD5=6e991e1a1f3293ab673afa015703ef86
+
+RUN wget --quiet https://github.com/wkhtmltopdf/wkhtmltopdf/releases/download/${WKHTMLTOPDF_VERSION}/wkhtmltox-${WKHTMLTOPDF_VERSION}_linux-generic-amd64.tar.xz -O wkhtmltox.tar.xz && \
+    echo "${WKHTMLTOPDF_MD5} wkhtmltox.tar.xz" > MD5SUMS && \
+    md5sum -c MD5SUMS && \
+    tar -xf wkhtmltox.tar.xz && \
+    mv wkhtmltox/bin/wkhtmltopdf /usr/local/bin && \
+    rm -rf wkhtmltox wkhtmltox.tar.xz MD5SUMS
+
+ARG TIKA_VERSION=2.9.2
+
+RUN wget --quiet https://archive.apache.org/dist/tika/KEYS -O tika-keys && \
+  wget --quiet https://archive.apache.org/dist/tika/${TIKA_VERSION}/tika-app-${TIKA_VERSION}.jar.asc -O tika-app.jar.asc && \
+  wget --quiet https://archive.apache.org/dist/tika/${TIKA_VERSION}/tika-app-${TIKA_VERSION}.jar -O tika-app.jar && \
+  gpg --import tika-keys && \
+  gpg --verify tika-app.jar.asc tika-app.jar && \
+  rm -f tika-app.jar.asc tika-keys && \
+  mv tika-app.jar /usr/share/java/tika-app.jar && \
+  echo '#!/bin/sh\n\nARGS="$@"\n\n[ $# -eq 0 ] && ARGS="--help"\n\nexec java -jar /usr/share/java/tika-app.jar $ARGS\n' > /usr/local/bin/tika && \
+  chmod +x /usr/local/bin/tika
+
+ENV APP_DIR /colore
+
+EXPOSE 9240
+
+WORKDIR $APP_DIR
+
+# Create application user and add permissions
+RUN groupadd --gid 1000 colore && \
+    useradd --uid 1000 --gid colore --shell /bin/bash --create-home colore && \
+    chown -R colore:colore $APP_DIR && \
+    chmod -R 755 $APP_DIR
+
+# Switch to application user
+USER colore
+
+RUN mkdir -p log
+RUN mkdir -p tmp/pids
+
+ARG RACK_ENV
+ENV RACK_ENV=${RACK_ENV:-production}
+
+RUN gem install bundler -v 2.4.22
+
+# Copy command always runs as root
+COPY --chown=colore:colore Gemfile .
+COPY --chown=colore:colore Gemfile.lock .
+
+RUN echo $RACK_ENV
+RUN \
+  if [ "$RACK_ENV" = "development" ] || [ "$RACK_ENV" = "test" ]; then \
+    bundle install --retry 10; \
+  else \
+    bundle config set deployment 'true' && bundle --retry 10; \
+  fi
+
+COPY --chown=colore:colore bin ./bin
+COPY --chown=colore:colore config ./config
+COPY --chown=colore:colore lib ./lib
+COPY --chown=colore:colore spec ./spec
+COPY --chown=colore:colore config.ru .
+COPY --chown=colore:colore Rakefile .
+COPY --chown=colore:colore unicorn.rb .
+
+# Create the shared volume so it will be owned by the application user
+RUN mkdir -p storage
+
+COPY --chown=colore:colore docker/colore/docker-entrypoint.sh ./bin
+RUN chmod +x ./bin/docker-entrypoint.sh
+
+ENTRYPOINT ["/colore/bin/docker-entrypoint.sh"]
+
+CMD ["bundle", "exec", "unicorn"]

docker/colore/docker-entrypoint.sh

@@ -0,0 +1,8 @@
+#!/bin/sh
+set -e
+
+pwd
+
+rm -f tmp/pids/*.pid
+
+exec "$@"

docker/colore/variables.env.example

@@ -0,0 +1,4 @@
+LANG=en_US.UTF-8
+LANGUAGE=en_US.UTF-8
+LC_ALL=C.UTF-8
+REDIS_URL=redis://redis:6379/4