openstax_content 0.0.2 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 578cf5b445c79d9e7dd1a904da90cb3daf527d2d8b7de128d07640e15d8bd066
-  data.tar.gz: b2bfe3524398a047b6722a047aa245de8ed52d02e4c9c37ca28f0b36840ba081
+  metadata.gz: ed22f20143aaa95ef57af3f123e218992d12320a2f3fc11181312aae26fc7372
+  data.tar.gz: 717b1f2bdce0faa6282fb3cbfef02845a590f2dab6be6cca515e4cb3e113c0dc
 SHA512:
-  metadata.gz: c2f7708cce0fb71727d985dcef91f6b0844f1b3ba2b134a55cc6f4eb3bddf0132c633cce04ed8221d16f8aac785a08f1ef16fe17e6f184c7f481c479054e496b
-  data.tar.gz: 07bd028f00e935ef240585da56b2fd88e82959da61213787f03421e1f07c0afecb8f3257b9b269ff1affea5483a227f70d299e1782a528e7ebb1c4d96dd2dd6b
+  metadata.gz: 31c7272f68a28f24aeb69b1a6f04bb8323976e484392e835d828b68c58838fe9a164ce54b2be073400c15485525d40359d7ea279fe3b8c64caeeb9d59476437b
+  data.tar.gz: 1eae72c6d0e78d02ae0152a741e90be6363ddac96b6670fe44080fe207b082e9fe54615ec9b316d1de8ce3394f7c0ed6aea79767c26b951486f5d793cb995c4f

data/lib/openstax/content/archive.rb

@@ -48,6 +48,10 @@ class OpenStax::Content::Archive
     if uri.path.start_with?('../')
       uri.path = uri.path.sub('..', '')
       "#{base_url}#{uri.to_s}"
+    elsif uri.path.start_with?(OpenStax::Content.archive_path) ||
+          uri.path.start_with?("/#{OpenStax::Content.archive_path}")
+      uri.path.start_with?('/') ? "https://#{OpenStax::Content.domain}#{uri.to_s}" :
+                                  "https://#{OpenStax::Content.domain}/#{uri.to_s}"
     else
       uri.path = "#{uri.path.chomp('.json').chomp('.xhtml')}.json"
 

data/lib/openstax/content/s3.rb

@@ -2,7 +2,7 @@ require 'aws-sdk-s3'
 
 class OpenStax::Content::S3
   def initialize
-    @ls = {}
+    @ls = Hash.new { |hash, key| hash[key] = Hash.new { |hash, key| hash[key] = {} } }
   end
 
   def bucket_name
@@ -21,24 +21,65 @@ class OpenStax::Content::S3
     )
   end
 
-  def ls(archive_version = nil)
-    return @ls[archive_version] unless @ls[archive_version].nil?
-    return unless bucket_configured?
-
+  # Returns the archive path for the given archive_version, book_id, page_uuid and extension
+  # If not all arguments are given, returns the prefix instead
+  def path_for(archive_version = nil, book_id = nil, page_uuid = nil, extension = nil)
     archive_path = OpenStax::Content.archive_path.chomp('/')
 
     if archive_version.nil?
-      prefix = "#{archive_path}/"
-      delimiter = '/'
+      "#{archive_path}/"
+    elsif book_id.nil?
+      "#{archive_path}/#{archive_version}/contents/"
+    elsif page_uuid.nil?
+      "#{archive_path}/#{archive_version}/contents/#{book_id}:"
+    elsif extension.nil?
+      "#{archive_path}/#{archive_version}/contents/#{book_id}:#{page_uuid}."
+    else
+      "#{archive_path}/#{archive_version}/contents/#{book_id}:#{page_uuid}.#{extension}"
+    end
+  end
+
+  # Without an archive version, returns a list of archive versions
+  # With an archive version, returns a list of book ids (uuid@version)
+  # With an archive version and a book, returns a list of page uuids
+  # With an archive version, book id and page uuid, returns the available extensions, if any
+  def ls(archive_version = nil, book_id = nil, page_uuid = nil)
+    return @ls[archive_version][book_id][page_uuid] \
+      unless @ls[archive_version][book_id][page_uuid].nil?
+    return unless bucket_configured?
+
+    prefix = path_for archive_version, book_id, page_uuid
+
+    delimiter = if archive_version.nil?
+      '/'
+    elsif book_id.nil?
+      ':'
+    elsif page_uuid.nil?
+      '.'
+    else
+      nil
+    end
+
+    responses = client.list_objects_v2 bucket: bucket_name, prefix: prefix, delimiter: delimiter
+
+    @ls[archive_version][book_id][page_uuid] = if page_uuid.nil?
+      responses.flat_map(&:common_prefixes).map do |common_prefix|
+        common_prefix.prefix.sub(prefix, '').chomp(delimiter)
+      end
     else
-      prefix = "#{archive_path}/#{archive_version}/contents/"
-      delimiter = ':'
+      responses.flat_map(&:contents).map { |content| content.key.sub(prefix, '') }
     end
+  end
 
-    @ls[archive_version] = client.list_objects_v2(
-      bucket: bucket_name, prefix: prefix, delimiter: delimiter
-    ).flat_map(&:common_prefixes).map do |common_prefix|
-      common_prefix.prefix.sub(prefix, '').chomp(delimiter)
+  # Checks all books for the given page uuid and returns the path to the first one found
+  def find_page(page_uuid, archive_version: nil, extension: 'json')
+    archive_version ||= ls.last
+
+    ls(archive_version).each do |book_id|
+      return path_for(archive_version, book_id, page_uuid, extension) \
+        if ls(archive_version, book_id, page_uuid).include?(extension)
     end
+
+    nil
   end
 end

data/lib/openstax/content/version.rb

@@ -1,5 +1,5 @@
 module OpenStax
   module Content
-    VERSION = '0.0.2'
+    VERSION = '0.1.0'
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: openstax_content
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.1.0
 platform: ruby
 authors:
 - Dante Soares
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-06-11 00:00:00.000000000 Z
+date: 2021-06-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aws-sdk-s3