openstax_content 0.2.0 → 1.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b17dd5822cd85c27b9d83850a0324d3633ef05bfd11793b9ac05bb734a0b8de3
-  data.tar.gz: 36be3946382c3f472ca4cc200ad2e7899505d4abc4101f0d94caebe287fc8d19
+  metadata.gz: 5ce0c0361f6d9e8205eb18893bb7d4f4bf726a2adf6fab3fe35649c56ae55876
+  data.tar.gz: ccec97c10020878437487261de88a742ccf16480e178227a323420ed2f9fe4fd
 SHA512:
-  metadata.gz: d32296abfb1ff02ea5781b62caf6616516459bb66f7a6615fdee3dc66481749f09c42e89df7ff28846b0a5fb97677b8d683150c0e60fa3c8eb4f0c7e870659b9
-  data.tar.gz: fab8643c3358e3cf02be92f59da7a3fc73338dbf679b1e38eeb4fc1364dc8d117c87913de960643e74c649806509694ebf56e68c38c1863e523191fa3d31f8e5
+  metadata.gz: 419d040e582db3b9233e2a6ebfcc3a6ad5bf908ed656a74aa67ecab1f111ad9278947e9bae3dab41d0479f36c7ad8750cabdac760dd6ca8ba73acacc8b94841d
+  data.tar.gz: aafb9064479bdaa157c557018b46c442f9c9abc4f0ce8fb48e12b7f81c8f0f09f2bba2829db6227bc7d5c36831b9909bb581839cea80da65cea87a5b807d33a9

data/lib/openstax/content/abl.rb

@@ -1,13 +1,81 @@
+require_relative 'archive'
+require_relative 'book'
+
 class OpenStax::Content::Abl
-  def body
-    @body ||= JSON.parse(Faraday.get(OpenStax::Content.abl_url).body).deep_symbolize_keys
+  def initialize(url: nil)
+    @url = url
+  end
+
+  def url
+    @url ||= OpenStax::Content.abl_url
+  end
+
+  def body_string
+    @body_string ||= Faraday.get(url).body
   end
 
-  def approved_books
-    body[:approved_books]
+  def body_hash
+    @body_hash ||= JSON.parse(body_string, symbolize_names: true)
   end
 
-  def approved_versions
-    body[:approved_versions]
+  def digest
+    Digest::SHA256.hexdigest body_string
+  end
+
+  def latest_approved_version_by_collection_id(archive: OpenStax::Content::Archive.new)
+    {}.tap do |hash|
+      body_hash[:approved_versions].each do |version|
+        next if version[:min_code_version] > archive.version
+
+        existing_version = hash[version[:collection_id]]
+
+        next if !existing_version.nil? &&
+                (existing_version[:content_version].split('.').map(&:to_i) <=>
+                 version[:content_version].split('.').map(&:to_i)) >= 0
+
+        hash[version[:collection_id]] = version
+      end
+    end
+  end
+
+  def approved_books(archive: OpenStax::Content::Archive.new)
+    # Can be removed once we have no more CNX books
+    version_by_collection_id = latest_approved_version_by_collection_id(archive: archive)
+
+    body_hash[:approved_books].flat_map do |approved_book|
+      if approved_book[:versions].nil?
+        # CNX-hosted book
+        version = version_by_collection_id[approved_book[:collection_id]]
+
+        next [] if version.nil?
+
+        approved_book[:books].map do |book|
+          OpenStax::Content::Book.new(
+            archive: archive,
+            uuid: book[:uuid],
+            version: version[:content_version].sub('1.', ''),
+            slug: book[:slug],
+            style: approved_book[:style]
+          )
+        end
+      else
+        # Git-hosted book
+        approved_book[:versions].flat_map do |version|
+          next [] if version[:min_code_version] > archive.version
+
+          commit_metadata = version[:commit_metadata]
+
+          commit_metadata[:books].map do |book|
+            OpenStax::Content::Book.new(
+              archive: archive,
+              uuid: book[:uuid],
+              version: version[:commit_sha][0..6],
+              slug: book[:slug],
+              style: book[:style]
+            )
+          end
+        end
+      end
+    end
   end
 end

data/lib/openstax/content/archive.rb

@@ -2,14 +2,22 @@ require 'addressable/uri'
 require 'faraday'
 
 class OpenStax::Content::Archive
-  def initialize(version)
+  def initialize(version: nil)
     @version = version
     @slugs = {}
   end
 
+  def s3
+    @s3 ||= OpenStax::Content::S3.new
+  end
+
+  def version
+    @version ||= s3.ls.last
+  end
+
   def base_url
     @base_url ||= "https://#{OpenStax::Content.domain}/#{
-                  OpenStax::Content.archive_path}/#{@version}"
+                  OpenStax::Content.archive_path}/#{version}"
   end
 
   def url_for(object)
@@ -28,10 +36,6 @@ class OpenStax::Content::Archive
     end
 
     if uri.absolute?
-      OpenStax::Content.logger.warn do
-        "#{self.class.name} received an unexpected absolute URL in url_for: \"#{object}\""
-      end
-
       # Force absolute URLs to be https
       uri.scheme = 'https'
       return uri.to_s
@@ -76,20 +80,16 @@ class OpenStax::Content::Archive
     end
   end
 
-  def s3
-    @s3 ||= OpenStax::Content::S3.new
-  end
-
   def add_latest_book_version_if_missing(object)
     book_id, page_id = object.split(':', 2)
     book_uuid, book_version = book_id.split('@', 2)
     return object unless book_version.nil? && s3.bucket_configured?
 
-    s3.ls(@version).each do |book|
-      uuid, version = book.split('@')
-      next unless uuid == book_uuid
+    s3.ls(version).each do |book|
+      s3_uuid, s3_version = book.split('@')
+      next unless s3_uuid == book_uuid
 
-      book_version = version
+      book_version = s3_version
       break
     end
 

data/lib/openstax/content/book.rb

@@ -1,33 +1,28 @@
-require_relative 'archive'
 require_relative 'book_part'
 
 class OpenStax::Content::Book
-  def initialize(
-    archive_version:, uuid: nil, version: nil, hash: nil, title: nil, tree: nil, root_book_part: nil
-  )
-    @uuid            = uuid || (hash || {})['id']
-    raise ArgumentError, 'Either uuid or hash with id key is required' if @uuid.nil?
+  extend Forwardable
 
-    @version         = version || (hash || {})['version']
-    raise ArgumentError, 'Either version or hash with version key is required' if @version.nil?
+  attr_reader :archive, :uuid, :version, :slug, :style
 
-    @archive_version = archive_version
-    @hash            = hash
-    @title           = title
-    @tree            = tree
-    @root_book_part  = root_book_part
-  end
-
-  attr_reader :archive_version, :uuid, :version
-
-  def archive
-    @archive ||= OpenStax::Content::Archive.new archive_version
+  def initialize(archive:, uuid:, version:, url: nil, hash: nil, slug: nil, style: nil)
+    @archive = archive
+    @uuid = uuid
+    @version = version
+    @url = url
+    @hash = hash
+    @slug = slug
+    @style = style
   end
 
   def url
     @url ||= archive.url_for "#{uuid}@#{version}"
   end
 
+  def hash
+    @hash ||= archive.json url
+  end
+
   def url_fragment
     @url_fragment ||= url.chomp('.json')
   end
@@ -40,22 +35,10 @@ class OpenStax::Content::Book
     @collated ||= hash.fetch('collated', false)
   end
 
-  def hash
-    @hash ||= archive.json url
-  end
-
-  def uuid
-    @uuid ||= hash.fetch('id')
-  end
-
   def short_id
     @short_id ||= hash['shortId']
   end
 
-  def version
-    @version ||= hash.fetch('version')
-  end
-
   def title
     @title ||= hash.fetch('title')
   end
@@ -67,4 +50,6 @@ class OpenStax::Content::Book
   def root_book_part
     @root_book_part ||= OpenStax::Content::BookPart.new(hash: tree, is_root: true, book: self)
   end
+
+  def_delegator :root_book_part, :all_pages
 end

data/lib/openstax/content/book_part.rb

@@ -44,4 +44,10 @@ class OpenStax::Content::BookPart
       end
     end
   end
+
+  def all_pages
+    @all_pages ||= parts.flat_map do |part|
+      part.is_a?(OpenStax::Content::Page) ? [ part ] : part.all_pages
+    end
+  end
 end

data/lib/openstax/content/fragment/html.rb

@@ -12,17 +12,28 @@ class OpenStax::Content::Fragment::Html < OpenStax::Content::Fragment
     @to_html = @node.to_html
   end
 
-  def as_json(*args)
-    # Don't attempt to serialize @node (it would fail)
-    super.except('node')
+  # Serialization methods use #instance_variables to iterate through and dump all instance variables
+  # Nokogiri classes are not serializable, so we do not want to dump the @node variable
+  # Instead, we recreate it by parsing the HTML again if needed
+  def instance_variables
+    super - [ :@node ]
   end
 
-  def html?
-    !to_html.empty?
+  def blank?
+    return @blank unless @blank.nil?
+
+    @blank = if to_html.nil? || to_html.strip.empty?
+      true
+    else
+      node_without_title = node.dup
+      node_without_title.css('[data-type="document-title"]').remove
+      text = node_without_title.text
+      text.nil? || text.strip.empty?
+    end
   end
 
-  def blank?
-    !html?
+  def html?
+    !blank?
   end
 
   def node

data/lib/openstax/content/fragment_splitter.rb

@@ -40,9 +40,11 @@ class OpenStax::Content::FragmentSplitter
     # Flatten, remove empty nodes and transform remaining nodes into reading fragments
     result.map do |obj|
       next obj unless obj.is_a?(Nokogiri::XML::Node)
-      next if obj.content.nil? || obj.content.strip.empty?
 
-      OpenStax::Content::Fragment::Reading.new node: obj, reference_view_url: reference_view_url
+      fragment = OpenStax::Content::Fragment::Reading.new(
+        node: obj, reference_view_url: reference_view_url
+      )
+      fragment unless fragment.blank?
     end.compact.tap do |result|
       @media_nodes.each do |node|
         # Media processing instructions

data/lib/openstax/content/s3.rb

@@ -1,12 +1,13 @@
 require 'aws-sdk-s3'
 
 class OpenStax::Content::S3
-  def initialize
+  def initialize(bucket_name: nil)
+    @bucket_name = bucket_name
     @ls = Hash.new { |hash, key| hash[key] = Hash.new { |hash, key| hash[key] = {} } }
   end
 
   def bucket_name
-    OpenStax::Content.bucket_name
+    @bucket_name ||= OpenStax::Content.bucket_name
   end
 
   def bucket_configured?

data/lib/openstax/content/version.rb

@@ -1,5 +1,5 @@
 module OpenStax
   module Content
-    VERSION = '0.2.0'
+    VERSION = '1.0.0'
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: openstax_content
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 1.0.0
 platform: ruby
 authors:
 - Dante Soares
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-06-30 00:00:00.000000000 Z
+date: 2022-03-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aws-sdk-s3
@@ -168,7 +168,7 @@ homepage: https://github.com/openstax/content-ruby
 licenses:
 - AGPL-3.0
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -183,8 +183,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.19
-signing_key: 
+rubygems_version: 3.1.4
+signing_key:
 specification_version: 4
 summary: Ruby bindings to read and parse the OpenStax ABL and the content archive
 test_files: []