Add document file preview generation with LibreOffice [SCI-3145]

Dockerfile

@@ -1,8 +1,11 @@
 FROM ruby:2.5.5
 MAINTAINER BioSistemika <info@biosistemika.com>
 
+RUN echo deb "http://http.debian.net/debian stretch-backports main" >> /etc/apt/sources.list
+
 # additional dependecies
 # libSSL-1.0 is required by wkhtmltopdf binary
+# libreoffice for file preview generation
 RUN curl -sL https://deb.nodesource.com/setup_8.x | bash - && \
   apt-get update -qq && \
   apt-get install -y \
@@ -14,6 +17,7 @@ RUN curl -sL https://deb.nodesource.com/setup_8.x | bash - && \
   unison \
   sudo graphviz --no-install-recommends \
   libfile-mimeinfo-perl && \
+  apt-get install -y --no-install-recommends -t stretch-backports libreoffice && \
   npm install -g yarn && \
   rm -rf /var/lib/apt/lists/*
 

Dockerfile.production

@@ -1,8 +1,11 @@
 FROM ruby:2.5.5
 MAINTAINER BioSistemika <info@biosistemika.com>
 
+RUN echo deb "http://http.debian.net/debian stretch-backports main" >> /etc/apt/sources.list
+
 # additional dependecies
 # libSSL-1.0 is required by wkhtmltopdf binary
+# libreoffice for file preview generation
 RUN curl -sL https://deb.nodesource.com/setup_8.x | bash - && \
   apt-get update -qq && \
   apt-get install -y \
@@ -16,6 +19,7 @@ RUN curl -sL https://deb.nodesource.com/setup_8.x | bash - && \
   default-jre-headless \
   sudo graphviz --no-install-recommends \
   libfile-mimeinfo-perl && \
+  apt-get install -y --no-install-recommends -t stretch-backports libreoffice && \
   npm install -g yarn && \
   rm -rf /var/lib/apt/lists/*
 

Gemfile

@@ -79,8 +79,8 @@ gem 'underscore-rails'
 gem 'wicked_pdf', '~> 1.1.0'
 gem 'wkhtmltopdf-heroku'
 
-gem 'aws-sdk', '~> 2'
-gem 'paperclip', '~> 5.3' # File attachment, image attachment library
+gem 'aws-sdk-s3'
+gem 'paperclip', '~> 6.1' # File attachment, image attachment library
 gem 'delayed_job_active_record'
 gem 'devise-async',
     git: 'https://github.com/mhfs/devise-async.git',

Gemfile.lock

@@ -102,13 +102,19 @@ GEM
       rails (>= 3.1)
     awesome_print (1.8.0)
     aws-eventstream (1.0.3)
-    aws-sdk (2.11.264)
-      aws-sdk-resources (= 2.11.264)
-    aws-sdk-core (2.11.264)
-      aws-sigv4 (~> 1.0)
+    aws-partitions (1.172.0)
+    aws-sdk-core (3.54.2)
+      aws-eventstream (~> 1.0, >= 1.0.2)
+      aws-partitions (~> 1.0)
+      aws-sigv4 (~> 1.1)
       jmespath (~> 1.0)
-    aws-sdk-resources (2.11.264)
-      aws-sdk-core (= 2.11.264)
+    aws-sdk-kms (1.21.0)
+      aws-sdk-core (~> 3, >= 3.53.0)
+      aws-sigv4 (~> 1.1)
+    aws-sdk-s3 (1.42.0)
+      aws-sdk-core (~> 3, >= 3.53.0)
+      aws-sdk-kms (~> 1)
+      aws-sigv4 (~> 1.1)
     aws-sigv4 (1.1.0)
       aws-eventstream (~> 1.0, >= 1.0.2)
     backports (3.14.0)
@@ -334,7 +340,7 @@ GEM
     overcommit (0.47.0)
       childprocess (~> 0.6, >= 0.6.3)
       iniparse (~> 1.4)
-    paperclip (5.3.0)
+    paperclip (6.1.0)
       activemodel (>= 4.2.0)
       activesupport (>= 4.2.0)
       mime-types
@@ -550,7 +556,7 @@ DEPENDENCIES
   auto_strip_attributes (~> 2.1)
   autosize-rails
   awesome_print
-  aws-sdk (~> 2)
+  aws-sdk-s3
   base62
   bcrypt (~> 3.1.10)
   better_errors
@@ -599,7 +605,7 @@ DEPENDENCIES
   omniauth
   omniauth-linkedin-oauth2
   overcommit
-  paperclip (~> 5.3)
+  paperclip (~> 6.1)
   pg (~> 0.18)
   pg_search
   phantomjs

app/models/asset.rb

@@ -10,16 +10,34 @@ class Asset < ApplicationRecord
 
   # Paperclip validation
   has_attached_file :file,
-                    styles: {
-                      large: [Constants::LARGE_PIC_FORMAT, :jpg],
-                      medium: [Constants::MEDIUM_PIC_FORMAT, :jpg],
-                      original: { processors: [:image_quality_calculate] }
+                    styles: lambda { |a|
+                      if a.previewable_document?
+                        {
+                          large: { processors: [:custom_file_preview],
+                                   geometry: Constants::LARGE_PIC_FORMAT,
+                                   format: :jpg },
+                          medium: { processors: [:custom_file_preview],
+                                    geometry: Constants::MEDIUM_PIC_FORMAT,
+                                    format: :jpg }
+                        }
+                      else
+                        {
+                          large: [Constants::LARGE_PIC_FORMAT, :jpg],
+                          medium: [Constants::MEDIUM_PIC_FORMAT, :jpg],
+                          original: { processors: [:image_quality_calculate] }
+                        }
+                      end
                     },
                     convert_options: {
                       medium: '-quality 70 -strip',
                       all: '-background "#d2d2d2" -flatten +matte'
                     }
 
+  before_post_process :previewable?
+
+  # adds image processing in background job
+  process_in_background :file, processing_image_url: '/images/:style/processing.gif'
+
   validates_attachment :file,
                        presence: true,
                        size: {
@@ -31,20 +49,6 @@ class Asset < ApplicationRecord
   # Should be checked for any security leaks
   do_not_validate_attachment_file_type :file
 
-  # adds image processing in background job
-  process_in_background :file,
-                        only_process: lambda { |a|
-                                        if a.content_type ==
-                                           %r{^image/#{ Regexp.union(
-                                             Constants::WHITELISTED_IMAGE_TYPES
-                                           ) }}
-                                          %i(large medium original)
-                                        else
-                                          {}
-                                        end
-                                      },
-                        processing_image_url: '/images/:style/processing.gif'
-
   # Asset validation
   # This could cause some problems if you create empty asset and want to
   # assign it to result
@@ -196,6 +200,10 @@ class Asset < ApplicationRecord
     end
   end
 
+  def previewable?
+    file.previewable_image? || file.previewable_document?
+  end
+
   def is_image?
     %r{^image/#{Regexp.union(Constants::WHITELISTED_IMAGE_TYPES)}} ===
       file.content_type

app/views/shared/_file_preview_icon.html.erb

@@ -1,4 +1,8 @@
 <div>
-  <i class="fas fa-10x <%= file_fa_icon_class(asset) %>"></i>
+  <% if asset.file.previewable_document? %>
+    <%= image_tag(asset.url(:large)) %>
+  <% else %>
+    <i class="fas fa-10x <%= file_fa_icon_class(asset) %>"></i>
+  <% end %>
   <h3 class="file-name"></h3>
 </div>

app/views/steps/attachments/_item.html.erb

@@ -1,4 +1,4 @@
-<% if asset.file.processing? && asset.is_image? %>
+<% if asset.file.processing? && (asset.is_image? || asset.file.previewable_document?) %>
   <% asset_status = 'asset-loading' %>
   <% present_url = step_file_present_asset_path(asset.id) %>
 <% else %>

app/views/steps/attachments/_placeholder.html.erb

@@ -1,4 +1,4 @@
-<% image_preview = asset.is_image? ? asset.url(:medium) : nil %>
+<% image_preview = (asset.is_image? || asset.file.previewable_document?) ? asset.url(:medium) : nil %>
 
 <div class="attachment-placeholder pull-left">
   <div class="attachment-thumbnail <%= image_preview ? '' : 'no-shadow' %>">

config/initializers/constants.rb

@@ -218,6 +218,8 @@ class Constants
     'text/plain'
   ].freeze
 
+  PREVIEWABLE_FILE_TYPES = TEXT_EXTRACT_FILE_TYPES
+
   WHITELISTED_IMAGE_TYPES = [
     'gif', 'jpeg', 'pjpeg', 'png', 'x-png', 'svg+xml', 'bmp', 'tiff'
   ].freeze

config/initializers/paperclip.rb

@@ -61,6 +61,27 @@ Paperclip::Attachment.class_eval do
   def fetch
     Paperclip.io_adapters.for self
   end
+
+  def previewable_document?
+    previewable = Constants::PREVIEWABLE_FILE_TYPES.include?(content_type)
+
+    extensions = %w(.xlsx .docx .pptx .xls .doc .ppt)
+    # Mimetype sometimes recognizes Office files as zip files
+    # In this case we also check the extension of the given file
+    # Otherwise the conversion should fail if the file is being something else
+    previewable ||= (content_type == 'application/zip' && extensions.include?(File.extname(original_filename)))
+
+    # Mimetype also sometimes recognizes '.xls' and '.ppt' files as
+    # application/x-ole-storage (https://github.com/minad/mimemagic/issues/50)
+    previewable ||=
+      (content_type == 'application/x-ole-storage' && %w(.xls .ppt).include?(File.extname(original_filename)))
+
+    previewable
+  end
+
+  def previewable_image?
+    content_type == %r{^image/#{Regexp.union(Constants::WHITELISTED_IMAGE_TYPES)}}
+  end
 end
 
 module Paperclip

lib/paperclip_processors/custom_file_preview.rb

@@ -0,0 +1,32 @@
+# frozen_string_literal: true
+
+module Paperclip
+  class CustomFilePreview < Processor
+    def make
+      libreoffice_path = ENV['LIBREOFFICE_PATH'] || 'libreoffice'
+      directory = File.dirname(@file.path)
+      basename  = File.basename(@file.path, '.*')
+      original_preview_file = File.join(directory, "#{basename}.png")
+      dst = TempfileFactory.new.generate("#{basename}.#{options[:format]}")
+
+      begin
+        Paperclip.run(
+          libreoffice_path,
+          "--headless --invisible --convert-to png --outdir #{directory} #{@file.path}"
+        )
+
+        convert(
+          ":source -resize '#{options[:geometry]}' -format #{options[:format]} #{options[:convert_options]} :dest",
+          source: File.expand_path(original_preview_file),
+          dest: File.expand_path(dst.path)
+        )
+      ensure
+        File.delete(original_preview_file) if File.file?(original_preview_file)
+      end
+
+      dst
+    rescue StandardError => e
+      raise Paperclip::Error, "There was an error generating document preview - #{e}"
+    end
+  end
+end