linkheaders-processor 0.1.13 → 0.1.17

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e88dd164547a9a21ce0f1a3ffa85f2af4190ea6a588da445387cdfa2dca7e25d
-  data.tar.gz: f800677c8d4cb18e274defb5dbda5d2f58431cab40899ae15cfb2f866fcf8644
+  metadata.gz: b10f24c0498058f393a2142465f0ff2e559dc2a8721ea8cf6a90178c8ff21789
+  data.tar.gz: 5cd760d37f6e82f63cb8271375bec72364264173010fda86e1ae32f1a424642d
 SHA512:
-  metadata.gz: ad4b8814c9ace9def1edd94e53e890c8d534a40d7cad4d55f7dbd0426e1310d29277a1849fd63ebada06fa3c9812a9d189c86ef2635ffe846fe52ff5f4864e2e
-  data.tar.gz: 99573e84fa6eb0412a5223cb188f333c37a8dbb3c877fcf5ea403a71bfc18d57d477920c6000b280fbb8cc920548ddd11ea95dcdb0df4f0a896e30a68fe84d9b
+  metadata.gz: e696571f2c9da932ff461af46740824de0728a6accd327ab6e109ab84530d313868d120cff5a62f670655efe6531069d7b6a358f1dcfcb94560cdd647614ff35
+  data.tar.gz: cf056d618d352bbcfaa43e59317fbc6f739186ac3e7ae26534e8f12721dec0d7ab2e8edfa07a2434e881a37d1cd18bf30e84f6cfac402fcc74ecd8c86170fa73

data/Gemfile.lock

@@ -1,10 +1,11 @@
 PATH
   remote: .
   specs:
-    linkheaders-processor (0.1.13)
+    linkheaders-processor (0.1.17)
       json (~> 2.0)
       json-ld (~> 3.2)
       json-ld-preloaded (~> 3.2)
+      link_header (~> 0.0.8)
       metainspector (~> 5.11.2)
       rest-client (~> 2.1)
       securerandom (~> 0.1.0)
@@ -18,7 +19,7 @@ GEM
     diff-lcs (1.5.0)
     domain_name (0.5.20190701)
       unf (>= 0.0.5, < 1.0.0)
-    faraday (1.10.0)
+    faraday (1.10.1)
       faraday-em_http (~> 1.0)
       faraday-em_synchrony (~> 1.0)
       faraday-excon (~> 1.1)
@@ -57,13 +58,13 @@ GEM
       domain_name (~> 0.5)
     json (2.6.2)
     json-canonicalization (0.3.0)
-    json-ld (3.2.1)
+    json-ld (3.2.3)
       htmlentities (~> 4.3)
       json-canonicalization (~> 0.3)
       link_header (~> 0.0, >= 0.0.8)
       multi_json (~> 1.15)
       rack (~> 2.2)
-      rdf (~> 3.2)
+      rdf (~> 3.2, >= 3.2.9)
     json-ld-preloaded (3.2.0)
       json-ld (~> 3.2)
       rdf (~> 3.2)
@@ -95,7 +96,7 @@ GEM
     rack (2.2.4)
     rainbow (3.1.1)
     rake (13.0.6)
-    rdf (3.2.8)
+    rdf (3.2.9)
       link_header (~> 0.0, >= 0.0.8)
     regexp_parser (2.5.0)
     rest-client (2.1.0)
@@ -117,7 +118,7 @@ GEM
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.11.0)
     rspec-support (3.11.0)
-    rubocop (1.32.0)
+    rubocop (1.33.0)
       json (~> 2.3)
       parallel (~> 1.10)
       parser (>= 3.1.0.0)

data/lib/linkheaders/link.rb

@@ -1,3 +1,4 @@
+
 module LinkHeaders
   class LinkFactory
 
@@ -191,5 +192,24 @@ module LinkHeaders
         self.send("#{k}=", v)
       end
     end
+
+    #
+    # Create an HTML version of the link
+    # @return [String]  HTML version of the Link object
+    #
+    def to_html
+      methods = self.linkmethods
+      href = self.href
+      rel = self.relation
+      anchor = self.anchor
+      properties = []
+      methods.each do |method|
+        value = self.send(method)
+        properties << [method, value]
+      end
+      properties << ["rel", rel]
+      properties << ["anchor", anchor]
+      LinkHeader::Link.new(href, properties).to_html
+    end
   end
 end

data/lib/linkheaders/processor/version.rb

@@ -3,6 +3,6 @@
 
 module LinkHeaders
   class Processor
-    VERSION = "0.1.13"
+    VERSION = "0.1.17"
   end
 end

data/lib/linkheaders/processor.rb

@@ -1,10 +1,9 @@
 # frozen_string_literal: true
 
 require_relative 'processor/version'
-require_relative 'constants'
 require_relative 'link'
 require_relative 'web_utils'
-
+require 'link_header'
 require 'json'
 require 'rest-client'
 require 'securerandom'
@@ -63,10 +62,10 @@ module LinkHeaders
       newlinks = parse_http_link_headers(head) # pass guid to check against anchors in linksets
       warn "HTTPlinks #{newlinks.inspect}"
 
-      HTML_FORMATS['html'].each do |format|
+      ['text/html','text/xhtml+xml', 'application/xhtml+xml'].each do |format|
         if head[:content_type] and head[:content_type].match(format)
           warn "found #{format} content - parsing"
-          htmllinks = parse_html_link_headers(body) # pass html body to find HTML link headers
+          htmllinks = parse_html_link_headers(body: body, anchor: default_anchor) # pass html body to find HTML link headers
           warn "htmllinks #{htmllinks.inspect}"
         end
       end
@@ -124,7 +123,7 @@ module LinkHeaders
         relation = sections['rel']
         sections.delete('rel')
         relations = relation.split(/\s+/)  # handle the multiple relation case
-        $stderr.puts "RELATIONS #{relations}"
+        warn "RELATIONS #{relations}"
 
         relations.each do |rel|
           next unless rel.match?(/\w/)
@@ -139,8 +138,8 @@ module LinkHeaders
     #
     # @param [String] body The HTML of the page containing HTML Link headers
     #
-    def parse_html_link_headers(body)
-      m = MetaInspector.new('http://example.org', document: body)
+    def parse_html_link_headers(body:, anchor: '')
+      m = MetaInspector.new(anchor, document: body)
       # an array of elements that look like this: [{:rel=>"alternate", :type=>"application/ld+json", :href=>"http://scidata.vitk.lv/dataset/303.jsonld"}]
       newlinks = Array.new
       m.head_links.each do |l|
@@ -155,7 +154,7 @@ module LinkHeaders
         l.delete(:href)        
         
         relations = relation.split(/\s+/)  # handle the multiple relation case
-        $stderr.puts "RELATIONS #{relations}"
+        warn "RELATIONS #{relations}"
 
         relations.each do |rel|
           next unless rel.match?(/\w/)
@@ -189,7 +188,7 @@ module LinkHeaders
     end
 
     def processJSONLinkset(href:)
-      _headers, linkset = fetch(href, { 'Accept' => 'application/linkset+json' })
+      _headers, linkset = lhfetch(href, { 'Accept' => 'application/linkset+json' })
       # warn "Linkset body #{linkset.inspect}"
       newlinks = Array.new
       return nil unless linkset
@@ -242,7 +241,7 @@ module LinkHeaders
 
     def processTextLinkset(href:)
       newlinks = Array.new
-      headers, linkset = fetch(href, { 'Accept' => 'application/linkset' })
+      headers, linkset = lhfetch(href, { 'Accept' => 'application/linkset' })
       # warn "linkset body #{linkset.inspect}"
       return {} unless linkset
 

data/lib/linkheaders/web_utils.rb

@@ -1,4 +1,4 @@
-def fetch(url, headers = ACCEPT_ALL_HEADER)  # we will try to retrieve turtle whenever possible
+def lhfetch(url, headers = {accept: "*/*"})
   # warn "In fetch routine now.  "
 
   # warn "executing call over the Web to #{url.to_s}"

data/spec/linkheader/parser_spec.rb

@@ -1,7 +1,37 @@
 # frozen_string_literal: true
+require_relative '../../lib/linkheaders/processor'
+require 'rest-client'
+
+
+url1 = "https://w3id.org/a2a-fair-metrics/22-http-html-citeas-describedby-mixed/"
+p = LinkHeaders::Processor.new(default_anchor: url1)
+r = RestClient.get(url1)
+p.extract_and_parse(response: r)
+factory = p.factory  # LinkHeaders::LinkFactory
+
+
+RSpec.describe LinkHeaders::Processor do
 
-RSpec.describe LinkHeader::Parser do
   it 'has a version number' do
-    expect(LinkHeader::Parser::VERSION).not_to be nil
+    expect(LinkHeaders::Processor::VERSION).not_to be nil
+  end
+
+  it "should find PURL citeas which has described-by and cite-as in mixed HTTP and HTML headers" do
+    expect(factory.all_links.length).to eq 5
+  end
+  it "should find find href on all links" do
+    expect(factory.all_links.select{|l| l.href}.length).to eq 5
+  end
+  it "should find find href on all links" do
+    expect(factory.all_links.select{|l| l.anchor}.length).to eq 5
+  end
+  it "should find 5 links in mixed HTTP and HTML headers" do
+    expect(factory.all_links.select{|l| l.relation}.length).to eq 5
+  end
+  it "should find one citeas in mixed HTTP and HTML headers" do
+    expect(factory.all_links.select{|l| l.relation == 'cite-as'}.length).to eq 1
+  end
+  it "should find described-by in mixed HTTP and HTML headers" do
+    expect(factory.all_links.select{|l| l.relation == 'describedby'}.length).to eq 1
   end
 end

data/spec/spec_helper.rb

@@ -7,7 +7,7 @@ RSpec.configure do |config|
   config.example_status_persistence_file_path = ".rspec_status"
 
   # Disable RSpec exposing methods globally on `Module` and `main`
-  config.disable_monkey_patching!
+  # config.disable_monkey_patching!
 
   config.expect_with :rspec do |c|
     c.syntax = :expect

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: linkheaders-processor
 version: !ruby/object:Gem::Version
-  version: 0.1.13
+  version: 0.1.17
 platform: ruby
 authors:
 - Mark Wilkinson
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-08-02 00:00:00.000000000 Z
+date: 2022-08-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -108,6 +108,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: 5.11.2
+- !ruby/object:Gem::Dependency
+  name: link_header
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.8
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.8
 description: A parser/processor for Link Headers and Linksets in both JSON and Text
   formats.
 email:
@@ -123,7 +137,6 @@ files:
 - README.md
 - Rakefile
 - launch.json
-- lib/linkheaders/constants.rb
 - lib/linkheaders/link.rb
 - lib/linkheaders/processor.rb
 - lib/linkheaders/processor/version.rb

data/lib/linkheaders/constants.rb

@@ -1,28 +0,0 @@
-ACCEPT_ALL_HEADER = {'Accept' => 'text/turtle, application/ld+json, application/rdf+xml, text/xhtml+xml, application/n3, application/rdf+n3, application/turtle, application/x-turtle, text/n3, text/turtle, text/rdf+n3, text/rdf+turtle, application/n-triples' }
-
-TEXT_FORMATS = {
-    'text' => ['text/plain',],
-}
-
-RDF_FORMATS = {
-  'jsonld'  => ['application/ld+json', 'application/vnd.schemaorg.ld+json'],  # NEW FOR DATACITE
-  'turtle'  => ['text/turtle','application/n3','application/rdf+n3',
-               'application/turtle', 'application/x-turtle','text/n3','text/turtle',
-               'text/rdf+n3', 'text/rdf+turtle'],
-  #'rdfa'    => ['text/xhtml+xml', 'application/xhtml+xml'],
-  'rdfxml'  => ['application/rdf+xml'],
-  'triples' => ['application/n-triples','application/n-quads', 'application/trig']
-}
-
-XML_FORMATS = {
-  'xml' => ['text/xhtml','text/xml',]
-}
-
-HTML_FORMATS = {
-  'html' => ['text/html','text/xhtml+xml', 'application/xhtml+xml']
-}
-
-JSON_FORMATS = {
-            'json' => ['application/json',]
-}
-