Merge pull request #12

Migrate PDF generation from `wicked_pdf` to `grover`, refactor related components

Author: paranoicsan

.env.docker

@@ -56,8 +56,9 @@ ELASTICSEARCH_ADDRESS=0.0.0.0
 # PDF GENERATION
 # Tools and settings for PDF document generation
 # ==============================================================================
-#WKHTMLTOPDF_PATH=           # Path to wkhtmltopdf binary
-#PUPPETEER_TIMEOUT=          # Puppeteer timeout in milliseconds
+GROVER_EXECUTABLE_PATH=/usr/bin/chromium
+GROVER_NO_SANDBOX=true
+PUPPETEER_TIMEOUT=3000
 
 # ==============================================================================
 # ERROR TRACKING

.env.example

@@ -4,6 +4,7 @@
 # ==============================================================================
 APP_NAME=                   # Application name
 APPLICATION_DOMAIN=         # Domain where the application is hosted
+FORCE_PREVIEW_GENERATION=   # Force generation of document and materials previews
 
 # ==============================================================================
 # DATABASE CONFIGURATION
@@ -56,7 +57,7 @@ ELASTICSEARCH_ADDRESS=      # Elasticsearch server address
 # PDF GENERATION
 # Tools and settings for PDF document generation
 # ==============================================================================
-WKHTMLTOPDF_PATH=           # Path to wkhtmltopdf binary
+GROVER_EXECUTABLE_PATH=     # Path to the Google Chrome or Chromium executable
 PUPPETEER_TIMEOUT=          # Puppeteer timeout in milliseconds
 
 # ==============================================================================

Dockerfile.dev

@@ -1,4 +1,34 @@
-FROM ruby:3.4.7-slim
+FROM ruby:3.4.7-slim AS base
+
+# Install base packages + Chromium for PDF generation
+RUN apt-get update -qq && \
+    apt-get install --no-install-recommends -y \
+    curl \
+    libjemalloc2 \
+    libvips \
+    postgresql-client \
+    # Chromium for Grover PDF generation
+    chromium \
+    chromium-driver \
+    # Required Chromium dependencies
+    fonts-liberation \
+    libappindicator3-1 \
+    libasound2 \
+    libatk-bridge2.0-0 \
+    libatk1.0-0 \
+    libcups2 \
+    libdbus-1-3 \
+    libnspr4 \
+    libnss3 \
+    libx11-xcb1 \
+    libxcomposite1 \
+    libxdamage1 \
+    libxrandr2 \
+    xdg-utils && \
+    ln -s /usr/lib/$(uname -m)-linux-gnu/libjemalloc.so.2 /usr/local/lib/libjemalloc.so && \
+    rm -rf /var/lib/apt/lists /var/cache/apt/archives
+
+FROM base AS dev
 
 # Install development packages
 RUN apt-get update -qq && \

Gemfile

@@ -66,6 +66,11 @@ group :development do
   gem "web-console"
 end
 
+# ============================================================================
+# New dependencies
+# ============================================================================
+gem "grover"
+
 # ============================================================================
 # Dependencies migrated from lcms-engine gem
 # ============================================================================
@@ -123,8 +128,6 @@ gem "combine_pdf", "~> 1.0"
 gem "rubyzip", "~> 2.3"
 gem "nokogiri", "~> 1.16"
 gem "sanitize", "~> 6.1"
-gem "wicked_pdf", "~> 2.1"
-gem "wkhtmltopdf-binary", "~> 0.12.6"
 
 # HTTP & External APIs
 gem "httparty", "~> 0.22"

Gemfile.lock

@@ -342,6 +342,8 @@ GEM
       multi_json (~> 1.11)
       os (>= 0.9, < 2.0)
       signet (>= 0.16, < 2.a)
+    grover (1.2.4)
+      nokogiri (~> 1)
     hashie (5.0.0)
     hiredis-client (0.26.1)
       redis-client (= 0.26.1)
@@ -790,16 +792,12 @@ GEM
       base64
       websocket-extensions (>= 0.1.0)
     websocket-extensions (0.1.5)
-    wicked_pdf (2.8.2)
-      activesupport
-      ostruct
     will_paginate (4.0.1)
     will_paginate-bootstrap-style (0.3.0)
       will_paginate (~> 4.0, >= 4.0.0)
     with_advisory_lock (5.3.0)
       activerecord (>= 6.1)
       zeitwerk (>= 2.6)
-    wkhtmltopdf-binary (0.12.6.10)
     xpath (3.2.0)
       nokogiri (~> 1.8)
     zeitwerk (2.7.3)
@@ -847,6 +845,7 @@ DEPENDENCIES
   fog-aws (~> 3.5, >= 3.5.2)
   google-apis-drive_v3 (~> 0.66)
   google-apis-script_v1 (~> 0.28)
+  grover
   hiredis-client
   httparty (~> 0.22)
   image_processing (~> 1.2)
@@ -900,11 +899,9 @@ DEPENDENCIES
   virtus (~> 1.0, >= 1.0.5)
   web-console
   webdrivers
-  wicked_pdf (~> 2.1)
   will_paginate (~> 4.0)
   will_paginate-bootstrap-style (~> 0.3)
   with_advisory_lock (~> 5.0)
-  wkhtmltopdf-binary (~> 0.12.6)
 
 BUNDLED WITH
    2.6.9

app/controllers/materials_controller.rb

@@ -4,15 +4,15 @@ class MaterialsController < Admin::AdminController
   before_action :set_material
 
   def preview_pdf
-    if !ENV.fetch('FORCE_PREVIEW_GENERATION', false) && (url = @material.preview_links["pdf"]).present?
+    if !ENV.fetch("FORCE_PREVIEW_GENERATION", false) && (url = @material.preview_links["pdf"]).present?
       return redirect_to url
     end
 
     preview_for :pdf
   end
 
   def preview_gdoc
-    if !ENV.fetch('FORCE_PREVIEW_GENERATION', false) &&  (url = @material.preview_links["gdoc"]).present?
+    if !ENV.fetch("FORCE_PREVIEW_GENERATION", false) &&  (url = @material.preview_links["gdoc"]).present?
       return redirect_to url
     end
 
@@ -28,7 +28,7 @@ def preview_for(preview_type, options = {})
     if service.perform
       links = @material.preview_links
       @material.update preview_links: links.merge(preview_type => service.url)
-      redirect_to service.url
+      redirect_to service.url, allow_other_host: true
     else
       redirect_to material_path(@material), alert: service.error
     end

app/helpers/asset_helper.rb

@@ -0,0 +1,51 @@
+# frozen_string_literal: true
+
+module AssetHelper
+  REDIS_PREFIX = "ub-b64-asset"
+
+  class << self
+    def base64_encoded(path, cache: false)
+      key = "#{REDIS_PREFIX}#{path}"
+
+      if cache
+        b64_asset = redis.get(key)
+        return b64_asset if b64_asset.present?
+      end
+
+      b64_asset = encode path
+      redis.set key, b64_asset, ex: 1.day.to_i if cache
+      b64_asset
+    end
+
+    def inlined(path)
+      if Rails.env.development? || Rails.env.test?
+        asset = Rails.application.assets.find_asset(path)
+      else
+        filesystem_path = Rails.application.assets_manifest.assets[path]
+        asset = File.read(Rails.root.join("public", "assets", filesystem_path))
+      end
+      asset
+    end
+
+    private
+
+    def encode(path)
+      if Rails.env.development? || Rails.env.test?
+        asset = Rails.application.assets.find_asset(path)
+        content_type = asset&.content_type
+      elsif (filesystem_path = Rails.application.assets_manifest.assets[path])
+        asset = File.read(Rails.root.join("public", "assets", filesystem_path))
+        content_type = Mime::Type.lookup_by_extension(File.extname(path).split(".").last)
+      end
+      raise "Could not find asset '#{path}'" if asset.nil?
+      raise "Unknown MimeType for asset '#{path}'" if content_type.nil?
+
+      encoded = Base64.encode64(asset.to_s).gsub(/\s+/, "")
+      "data:#{content_type};base64,#{Rack::Utils.escape(encoded)}"
+    end
+
+    def redis
+      Rails.application.config.redis
+    end
+  end
+end

app/presenters/document_presenter.rb

@@ -9,7 +9,6 @@ def content_for(context_type, options = {})
     with_excludes = (options[:excludes] || []).any?
     content = render_content(context_type, options)
     content = update_activity_timing(content) if with_excludes
-    content = remove_optional_break(content) if ela? && with_excludes
     content
   end
 
@@ -21,20 +20,10 @@ def doc_type
     "lesson"
   end
 
-  def remove_optional_break(content)
-    html = Nokogiri::HTML.fragment content
-    html.at_css(".o-ld-optbreak-wrapper")&.remove
-    html.to_html
-  end
-
   def base_metadata
     @base_metadata ||= DocTemplate::Objects::Document.build_from(metadata)
   end
 
-  def module_value
-    ela? ? send(:module) : unit
-  end
-
   def pdf_filename
     name = short_breadcrumb(join_with: "_", with_short_lesson: true)
     name += PDF_SUBTITLES[content_type.to_sym]
@@ -63,6 +52,7 @@ def render_content(context_type, options = {})
   def short_breadcrumb(join_with: " / ", with_short_lesson: false, with_subject: true, unit_level: false)
     lesson_abbr = with_short_lesson ? "L#{lesson}" : "Lesson #{lesson}" \
       unless unit_level
+    module_value = ela? ? send(:module) : unit
     [
       with_subject ? SUBJECT_FULL[subject] || subject : nil,
       grade.to_i.zero? ? grade : "G#{grade}",

app/presenters/material_presenter.rb

@@ -1,9 +1,7 @@
 # frozen_string_literal: true
 
 class MaterialPresenter < ContentPresenter
-  attr_accessor :document
-
-  delegate :name_date, :show_title, :subject, to: :base_metadata
+  delegate :grade, :name_date, :show_title, :subject, to: :base_metadata
 
   DEFAULT_TITLE = "Material"
   MATERIAL_TYPES = {
@@ -14,22 +12,17 @@ class MaterialPresenter < ContentPresenter
 
   def base_filename(with_version: true)
     name = base_metadata.identifier
-    name = "#{document.short_breadcrumb(join_with: '_', with_short_lesson: true)}_#{name}"
     with_version ? "#{name}_v#{version.presence || 1}" : name
   end
 
   def cc_attribution
-    base_metadata.cc_attribution.presence || document&.cc_attribution
+    base_metadata.cc_attribution.to_s
   end
 
   def content_for(context_type, options = {})
     render_content(context_type, options)
   end
 
-  def gdoc_folder
-    "#{document.id}_v#{document.version}"
-  end
-
   def gdoc_preview_title
     preview_links["gdoc"].present? ? "Preview Google Document" : "Generate Google Document"
   end
@@ -51,7 +44,7 @@ def orientation
   end
 
   def pdf_filename
-    "#{document.id}/#{base_filename}"
+    "#{id}/#{base_filename}"
   end
 
   def pdf_url
@@ -93,11 +86,7 @@ def base_metadata
     @base_metadata ||= DocTemplate::Objects::Material.build_from(metadata)
   end
 
-  def material_links
-    @material_links ||= (document || @lesson).links["materials"]&.dig(id.to_s)
-  end
-
   def material_url(key)
-    material_links&.dig(key).to_s
+    "TBD"
   end
 end

app/services/material_preview_generator.rb

@@ -15,8 +15,6 @@ def initialize(material, options = {})
   end
 
   def perform
-    return false unless assign_document
-
     options[:type] == :pdf ? generate_pdf : generate_gdoc
   rescue StandardError => e
     @error = e.message
@@ -27,18 +25,6 @@ def perform
 
   attr_reader :material, :options
 
-  # TODO: Check why we need the connection between material and document
-  def assign_document # rubocop:disable Naming/PredicateMethod
-    document = material.documents.last || Document.last
-    unless document.present?
-      @error = "Can't generate PDF for preview: no documents exist"
-      return false
-    end
-    material.document = DocumentGenerator.document_presenter.new document
-
-    true
-  end
-
   def generate_gdoc
     folder_id = options[:folder_id]
     file_id = material.preview_links["gdoc"].to_s.match(GDOC_RE)&.[](1)
@@ -47,7 +33,7 @@ def generate_gdoc
   end
 
   def generate_pdf # rubocop:disable Naming/PredicateMethod
-    pdf_filename = "#{PDF_S3_FOLDER}/#{material.id}/#{material.base_filename}#{ContentPresenter::PDF_EXT}"
+    pdf_filename = "#{PDF_S3_FOLDER}/#{material.base_filename}#{ContentPresenter::PDF_EXT}"
     pdf = DocumentExporter::Pdf::Material.new(material).export
     @url = S3Service.upload pdf_filename, pdf
     true