tao_rdfizer 0.10 → 0.11.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a2f897b03c9e6856aaf752fd645808c47e6d52422ea1b6934b44b694d8d0d9a1
-  data.tar.gz: 8d9fb426258bf3fceabb6471a81fcaeb99cd18dcace251575a0f2f715fa4d787
+  metadata.gz: ef68cf0cfd37026eda3abfb19f294d866c343294c1a9cc8f4ed2aa9af1f1443c
+  data.tar.gz: fdde442909c60d0dbfe81ea445b049a52829752d01b06f5f599786071b09b62c
 SHA512:
-  metadata.gz: f0d230f25945bf45aed05f8f0cb3d46113d5eb4213b17ce4812cbfbcfa5636d540343b8ee8fe94cf1bbbc1a1eb4db35ac076bf1d7cd3cb7abdd92db446c94f41
-  data.tar.gz: 6f00c1aeb735b8e76da2e9185ad141c0c29c1d45d94faad1b3d89ec903a1c931e22957f4f356dce24ff6beadd53a77ac7fcd5aad2c17cb281a5d0128cd062b89
+  metadata.gz: a0219d82d900259fd8e7dab5ea5382bc06bb504ea21207302b21b59047f33ba899491aa53043a9456af1c9e1693d9a957223a9cad31d8a2e21f4df60aebdbf4e
+  data.tar.gz: 5250d9a3daf7716e5c8f2e01fe2560589862fb3b7cdf7a03c067f07e47930f91b6912b7b9316d8de652fcf983f677e1ab0d9de16954f2fdf3880478e7f50344a

data/bin/tao_rdfizer

@@ -3,6 +3,7 @@ require 'tao_rdfizer'
 require 'json'
 
 mode = nil
+options = {}
 
 ## command line option processing
 require 'optparse'
@@ -17,6 +18,14 @@ optparse = OptionParser.new do |opts|
     mode = :spans
   end
 
+  opts.on('-x', '--x-prefixes', 'without prefixes.') do
+    options[:with_prefixes] = false
+  end
+
+  opts.on('-o', '--only-prefixes', 'only prefixes.') do
+    options[:only_prefixes] = true
+  end
+
   opts.on('-h', '--help', 'displays this screen.') do
     puts opts
     exit
@@ -34,7 +43,7 @@ begin
 	annotations = JSON.parse File.read(ARGV[0]), :symbolize_names => true
 	annotations = [annotations] unless annotations.class == Array
 	rdfizer = TAO::RDFizer.new(mode)
-	puts rdfizer.rdfize(annotations)
+	puts rdfizer.rdfize(annotations, options)
 rescue ArgumentError, IOError => e
 	puts e.message
 end

data/lib/tao_rdfizer/tao_rdfizer.rb

@@ -14,18 +14,45 @@ class TAO::RDFizer
 		else
 			ERB_ANNOTATIONS_TTL
 		end
-
 		@tao_ttl_erb = ERB.new(template, nil, '-')
 		@prefix_ttl_erb = ERB.new(ERB_PREFIXES_TTL, nil, '-')
 	end
 
-	def rdfize(annotations_col)
+	def rdfize(annotations_col, options = nil)
+		options ||= {}
+		only_prefixes = options.has_key?(:only_prefixes) ? options[:only_prefixes] == true : false
+		with_prefixes = options.has_key?(:with_prefixes) ? options[:with_prefixes] == true : true
+
+		# check the format
+		annotations_col.each do |annotations|
+			raise "'target' is missing" unless annotations.has_key? :target
+		end
+
 		# namespaces
 		namespaces = {}
 
 		anns = annotations_col.first
 		anns[:namespaces].each {|n| namespaces[n[:prefix]] = n[:uri]} unless anns[:namespaces].nil?
 
+		prefixes_ttl = @prefix_ttl_erb.result_with_hash(namespaces:namespaces) if only_prefixes || with_prefixes
+
+		if only_prefixes
+			prefixes_ttl
+		else
+			annotations_ttl = get_annotations_ttl(annotations_col, namespaces)
+			if with_prefixes
+				prefixes_ttl + annotations_ttl
+			else
+				annotations_ttl
+			end
+		end
+	end
+
+	private
+
+	def get_annotations_ttl(annotations_col, namespaces)
+		anns = annotations_col.first
+
 		unless @mode ==:spans
 			raise ArgumentError, "A project name has to be specified." unless anns.has_key?(:project)
 			prefix_for_this = anns[:project].downcase.gsub(/ /, '_')
@@ -53,14 +80,14 @@ class TAO::RDFizer
 			_relations = annotations[:relations] || []
 			if @mode == :spans && annotations.has_key?(:tracks)
 				annotations[:tracks].each do |track|
-					_denotations += track[:denotations]
-					_attributes += track[:attributes]
-					_relations += track[:relations]
+					_denotations += track[:denotations] if track.has_key? :denotations
+					_attributes += track[:attributes] if track.has_key? :attributes
+					_relations += track[:relations] if track.has_key? :relations
 				end
 			end
 
 			begin
-				if @mode == :annotations
+				unless @mode == :span
 					# index attributes
 					attributesh = _attributes.inject({}) do |h, a|
 						if a[:pred].end_with?('_id')
@@ -76,7 +103,7 @@ class TAO::RDFizer
 						span_uri = "<#{text_uri}/spans/#{d[:span][:begin]}-#{d[:span][:end]}>"
 						d[:span_uri] = span_uri
 						d[:obj_uri] = "#{prefix_for_this}:#{text_id}-#{d[:id]}"
-						class_uris = attributesh[d[:id]].push(d[:obj])
+						class_uris = (attributesh[d[:id]] || []).push(d[:obj])
 						d[:class_uris] = class_uris.map{|uri| find_uri(uri, namespaces, prefix_for_this)}
 					rescue ArgumentError => e
 						raise ArgumentError, "[#{sourcedb}-#{sourceid}-#{d[:id]}] " + e.message
@@ -167,11 +194,9 @@ class TAO::RDFizer
 			spans += _spans unless @mode == :annotations
 		end
 
-		ttl = @prefix_ttl_erb.result(binding) + @tao_ttl_erb.result(binding)
+		@tao_ttl_erb.result(binding)
 	end
 
-	private
-
 	def include_parent?(spans, span)
 		# spans.each{|s| return true if (s[:begin] <= span[:begin] && s[:end] > span[:end]) || (s[:begin] < span[:begin] && s[:end] >= span[:end])}
 		spans.each{|s| return true if s[:begin] <= span[:begin] && s[:end] >= span[:end]}
@@ -187,7 +212,10 @@ class TAO::RDFizer
 	end
 
 	def find_uri (label, namespaces, prefix_for_this)
-		raise ArgumentError, "A label including a whitespace character found: #{label}." if label.match(/\s/)
+		if label.match(/\s/)
+			# raise ArgumentError, "A label including a whitespace character found: #{label}."
+			label.gsub(/\s/, '_')
+		end
 		delimiter_position = label.index(':')
 		if !delimiter_position.nil? && namespaces.keys.include?(label[0...delimiter_position])
 			label.gsub('(', '\(').gsub(')', '\)') # brackets have to be escaped
@@ -203,6 +231,18 @@ class TAO::RDFizer
 		end
 	end
 
+	def rdf_literal_escape(string)
+		string.gsub('\\', '\\\\').
+					 gsub("\t", '\\t').
+					 gsub("\b", '\\b').
+					 gsub("\n", '\\n').
+					 gsub("\r", '\\r').
+					 gsub("\f", '\\f').
+					 gsub('"', '\\"').
+					 freeze
+	end
+
+	# variable: denotations, relations
 	ERB_ANNOTATIONS_TTL = <<~HEREDOC
 		<% denotations.each do |d| -%>
 		<%= d[:obj_uri] %> tao:denoted_by <%= d[:span_uri] %> ;
@@ -214,6 +254,7 @@ class TAO::RDFizer
 		<% end -%>
 	HEREDOC
 
+	# variable: spans
 	ERB_SPANS_TTL = <<~HEREDOC
 		<% spans.each do |s| -%>
 		<%= s[:span_uri] %> rdf:type tao:Text_span ;
@@ -223,13 +264,14 @@ class TAO::RDFizer
 		<% s[:children].each do |s| -%>
 			tao:contains <%= s[:span_uri] %> ;
 		<% end -%>
-			tao:has_text <%= s[:text].inspect %> ;
+			tao:has_text "<%= rdf_literal_escape(s[:text]) %>" ;
 			tao:belongs_to <<%= s[:source_uri] %>> ;
 			tao:begins_at <%= s[:begin] %> ;
 			tao:ends_at <%= s[:end] %> .
 		<% end -%>
 	HEREDOC
 
+	# variable: namespaces
 	ERB_PREFIXES_TTL = <<~HEREDOC
 		@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
 		@prefix tao: <http://pubannotation.org/ontology/tao.owl#> .

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tao_rdfizer
 version: !ruby/object:Gem::Version
-  version: '0.10'
+  version: 0.11.2
 platform: ruby
 authors:
 - Jin-Dong Kim
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-12-20 00:00:00.000000000 Z
+date: 2021-05-23 00:00:00.000000000 Z
 dependencies: []
 description: It uses TAO (text annotation ontology) for representation of annotations
   to text.
@@ -40,7 +40,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.8
+rubygems_version: 3.0.9
 signing_key: 
 specification_version: 4
 summary: A RDF statement generator for annotations in the PubAnnotation JSON format.