tao_rdfizer 0.9.10 → 0.10.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 839f30915cc8896477342a8cfdb87d8d4f23d8ffbbc133b9b0b087939050befe
-  data.tar.gz: bfa3efcbae03859f8a778afbf81f37daffec3467b537671cf5282c0905ed1d5f
+  metadata.gz: 4baf53459b357e899ab4e569f15ba00e4240653641912488c2478ef0a65ac2bf
+  data.tar.gz: cd7d1a7121abbd4f725b1cbf5dff8cf28db27efdcab025e1c43305d0350157ab
 SHA512:
-  metadata.gz: 2a9744b372d7a9774774262a17771c43533b44c3a95ad19cd7e60add25d6135afb3c85ecad8d6933ffdff61b39632be0386ecf9d85340f1f0bea1efc9176e8d5
-  data.tar.gz: 77f1d13bf567b6e5ed28668034da90e0810520358e2aa2d27dc747ce78f090316b1519be8f1b2a30050fce5a20408b46a932887c955979a217304c4a53cd709e
+  metadata.gz: ba755e5d41a06b36d592dc52261ed65b9018c47ed4da5b29fa0bf2a49196e0ae29c046c1cf463be9bbaad510ddca2bed464ce17898c58480e3258668374a7675
+  data.tar.gz: cb71ac538c0573812317caa83a54b3069dca0221d85388dd2ede2e1251b4d820379e43ac45744dbc450faf4a3dcd0db542832075a4febdc03389d956a7cbd9ff

data/lib/tao_rdfizer/tao_rdfizer.rb

@@ -27,6 +27,7 @@ class TAO::RDFizer
 		anns[:namespaces].each {|n| namespaces[n[:prefix]] = n[:uri]} unless anns[:namespaces].nil?
 
 		unless @mode ==:spans
+			raise ArgumentError, "A project name has to be specified." unless anns.has_key?(:project)
 			prefix_for_this = anns[:project].downcase.gsub(/ /, '_')
 			raise ArgumentError, "'#{prefix_for_this}' is a reserved prefix for this project." if namespaces.has_key?(prefix_for_this)
 			project_uri = 'http://pubannotation.org/projects/' + anns[:project]
@@ -34,6 +35,7 @@ class TAO::RDFizer
 		end
 
 		denotations = []
+		attributes = []
 		relations = []
 		spans = []
 
@@ -46,34 +48,57 @@ class TAO::RDFizer
 			end
 
 			# denotations and relations
-			_denotations = annotations[:denotations]
-			_relations = annotations[:relations]
-			_denotations = [] if _denotations.nil?
-			_relations = [] if _relations.nil?
+			_denotations = annotations[:denotations] || []
+			_attributes = annotations[:attributes] || []
+			_relations = annotations[:relations] || []
 			if @mode == :spans && annotations.has_key?(:tracks)
 				annotations[:tracks].each do |track|
 					_denotations += track[:denotations]
+					_attributes += track[:attributes]
 					_relations += track[:relations]
 				end
 			end
 
 			begin
-				# denotations preprocessing
-				_denotations.each do |d|
-					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]}"
-					d[:cls_uri] = find_uri(d[:obj], namespaces, prefix_for_this)
-				end
+				if @mode == :annotations
+					# index attributes
+					attributesh = _attributes.inject({}) do |h, a|
+						if a[:pred].end_with?('_id')
+							subj = a[:subj]
+							h[subj] = [] unless h.has_key? subj
+							h[subj] << a[:obj]
+						end
+						h
+					end
 
-				# relations preprocessing
-				_relations.each do |r|
-					r[:subj_uri] = "#{prefix_for_this}:#{text_id}-#{r[:subj]}"
-					r[:obj_uri] = "#{prefix_for_this}:#{text_id}-#{r[:obj]}"
-					r[:pred_uri] = find_uri(r[:pred], namespaces, prefix_for_this)
+					# denotations preprocessing
+					_denotations.each do |d|
+						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])
+						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
+					end
+
+					# relations preprocessing
+					_relations.each do |r|
+						r[:subj_uri] = "#{prefix_for_this}:#{text_id}-#{r[:subj]}"
+						r[:obj_uri] = "#{prefix_for_this}:#{text_id}-#{r[:obj]}"
+						r[:pred_uri] = find_uri(r[:pred], namespaces, prefix_for_this)
+					rescue ArgumentError => e
+						raise ArgumentError, "[#{sourcedb}-#{sourceid}-#{r[:id]}] " + e.message
+					end
+				else
+					# denotations preprocessing
+					_denotations.each do |d|
+						span_uri = "<#{text_uri}/spans/#{d[:span][:begin]}-#{d[:span][:end]}>"
+						d[:span_uri] = span_uri
+					rescue ArgumentError => e
+						raise ArgumentError, "[#{sourcedb}-#{sourceid}-#{d[:id]}] " + e.message
+					end
 				end
-			rescue ArgumentError => e
-				raise ArgumentError, "[#{sourcedb}-#{sourceid}] " + e
 			end
 
 			unless @mode == :annotations
@@ -88,7 +113,7 @@ class TAO::RDFizer
 					s[:text] = text[s[:begin] ... s[:end]]
 				end
 
-				# index
+				# index spans
 				spanh = _spans.inject({}){|r, s| r[s[:span_uri]] = s; r}
 
 				# add denotation information
@@ -165,23 +190,23 @@ class TAO::RDFizer
 		raise ArgumentError, "A label including a whitespace character found: #{label}." if label.match(/\s/)
 		delimiter_position = label.index(':')
 		if !delimiter_position.nil? && namespaces.keys.include?(label[0...delimiter_position])
-			label
+			label.gsub('(', '\(').gsub(')', '\)') # brackets have to be escaped
 		elsif label =~ %r[^https?://]
 			"<#{label}>"
 		else
 			clabel = if label.match(/^\W+$/)
 				'SYM'
 			else
-				label.sub(/^\W+/, '').sub(/\W+$/, '').gsub(/ +/, '_')
+				label.sub(/^\W+/, '').sub(/[^a-zA-Z0-9_)]+$/, '').gsub(/ +/, '_')
 			end
-			namespaces.has_key?('_base') ? "<#{clabel}>" : "#{prefix_for_this}:#{clabel}"
+			namespaces.has_key?('_base') ? "<#{clabel}>" : "#{prefix_for_this}:#{clabel.gsub('(', '\(').gsub(')', '\)')}"
 		end
 	end
 
 	ERB_ANNOTATIONS_TTL = <<~HEREDOC
 		<% denotations.each do |d| -%>
 		<%= d[:obj_uri] %> tao:denoted_by <%= d[:span_uri] %> ;
-			rdf:type <%= d[:cls_uri] %> .
+		<%= d[:class_uris].map{|c| "\trdf:type " + c}.join(" ;\n") + " ." %>
 		<% end -%>
 		<%# relations -%>
 		<% relations.each do |r| -%>

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tao_rdfizer
 version: !ruby/object:Gem::Version
-  version: 0.9.10
+  version: 0.10.1
 platform: ruby
 authors:
 - Jin-Dong Kim
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-07-11 00:00:00.000000000 Z
+date: 2020-12-24 00:00:00.000000000 Z
 dependencies: []
 description: It uses TAO (text annotation ontology) for representation of annotations
   to text.
@@ -40,8 +40,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.9
+rubygems_version: 3.0.8
 signing_key: 
 specification_version: 4
 summary: A RDF statement generator for annotations in the PubAnnotation JSON format.