ttl2html 3.0.0 → 3.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8120a06f3f763ca13d8b9f35c920f9129d57876a5dde434b264c052bd1aa9151
-  data.tar.gz: 9d1cd9d09f5a14a5d56856c00099b38abe059a9c27f775df2074b42db99e9a7a
+  metadata.gz: d7254a4cc200e719bbb09c33d88dea4a22a8360e1c6453085f5331d541320b82
+  data.tar.gz: 43385ec41709d6b8a552dfe7bb2c2d7186b1142b0655bd8480b2fc0b91b336d8
 SHA512:
-  metadata.gz: 4935a3824f19a696ea4657d652c4afd482af6a8a2e33aa98725d4739375480654b86dd4a9781dae357d816d422dc7807a974bb7dfb38fbe1d6cb5d84a84a3792
-  data.tar.gz: 9e561cb7abbfc195e8a13f09af9161a0b88f301e73b040ab149707e5f47905728a9ed603210b9f2818771350fd11670072941df4b58d92e4f5c268c608845274
+  metadata.gz: 0be4d1d1550a1eb410ffb8747eeffbd8b3154a297966cb99cde5ecb6c249973a8c345e88b01ab8cd7c1c791064ea78b6708ad7bc1579231b2f5424b8643af709
+  data.tar.gz: a5ec01247b296e2be84d110e4742412152465f9925223950211ca99a5c9f6112ebe81820fa6a422589f8cf45f5d41ecfb3b619c33e833694b9811cb80fb86541

data/lib/ttl2html/version.rb

@@ -1 +1 @@
-TTL2HTML::VERSION = "3.0.0"
+TTL2HTML::VERSION = "3.1.0"

data/lib/ttl2html.rb

@@ -72,9 +72,11 @@ module TTL2HTML
       $stderr.puts "#{count} triples. #{subjects.size} subjects."
       @data
     end
-    def format_turtle(subject, depth = 1)
+    def format_turtle(subject, depth = 1, force = false)
       turtle = RDF::Turtle::Writer.new
       result = ""
+      #p [:format_turtle, subject, depth, force]
+      return result if !force && @cache[:output_turtle_files].include?(subject)
       if subject =~ /^_:/
         result << "[\n#{"  "*depth}"
       else
@@ -92,14 +94,14 @@ module TTL2HTML
             [ @data[object.to_s][schema_position] ? @data[object.to_s][schema_position].first.to_i : Float::INFINITY,
               @data[object.to_s][qb_order] ? @data[object.to_s][qb_order].first.to_i : Float::INFINITY,
               @data[object.to_s][shacl_order] ? @data[object.to_s][shacl_order].first.to_i : Float::INFINITY,
-              format_turtle(object, depth + 1)
+              format_turtle(object, depth + 1, true)
             ]
           else
             [Float::INFINITY, Float::INFINITY, Float::INFINITY, object.to_s]
           end
         end.map do |object|
           if /^_:/ =~ object.to_s # blank node:
-            format_turtle(object, depth + 1)
+            format_turtle(object, depth + 1, force)
           elsif RDF::URI::IRI =~ object.to_s
             turtle.format_uri(RDF::URI.new object)
           elsif object.respond_to?(:first) and object.first.kind_of?(Symbol)
@@ -110,22 +112,117 @@ module TTL2HTML
         end.join(", ")
         str
       end.join(";\n#{"  "*depth}")
-      result << " ." if not subject =~ /^_:/
+      result << "." if not subject =~ /^_:/
       result << "\n"
       result << "#{"  "*(depth-1)}]" if subject =~ /^_:/
+      @cache[:output_turtle_files] << subject unless force
       result
     end
     def format_turtle_inverse(object)
-      result = ""
-      return result if not object.start_with? @config[:base_uri]
-      return result if not @data_inverse.has_key? object
-      @data_inverse[object].keys.sort.each do |predicate|
-        @data_inverse[object.to_s][predicate].sort.each do |subject|
-          next if subject =~ /^_:/
-          result << "<#{subject}> <#{predicate}> <#{object}>.\n"
+      triples = collect_inverse_triples(object)
+      return "" if triples.empty?
+      by_subject = build_subject_index(triples)
+      ref_count  = build_object_ref_count(triples)
+      roots      = find_inverse_roots(by_subject)
+      roots.map do |root|
+        "#{format_inverse_subject(root, by_subject, ref_count, Set.new, 1)}.\n"
+      end.join
+    end
+    def collect_inverse_triples(object, triples = Set.new, visited = Set.new)
+      return triples if object.to_s.start_with?("_:")
+      return triples unless object.to_s.start_with?(@config[:base_uri].to_s)
+      return triples unless @data_inverse.key?(object.to_s)
+      return triples if visited.include?(object.to_s)
+      visited << object.to_s
+      @data_inverse[object.to_s].each do |predicate, subjects|
+        subjects.each do |subject|
+          triples << [subject.to_s, predicate.to_s, object.to_s]
+          collect_inverse_triples_for_bnode(subject.to_s, triples, visited) if subject.to_s.start_with?("_:")
         end
       end
-      result
+      triples
+    end
+    def collect_inverse_triples_for_bnode(node, triples, visited)
+      return triples unless @data_inverse.key?(node)
+      return triples if visited.include?(node)
+      visited << node
+      @data_inverse[node].each do |predicate, subjects|
+        subjects.each do |subject|
+            triples << [subject.to_s, predicate.to_s, node]
+            collect_inverse_triples_for_bnode(subject.to_s, triples, visited) if subject.to_s.start_with?("_:")
+        end
+      end
+      triples
+    end
+    def build_subject_index(triples)
+      by_subject = Hash.new { |h, k| h[k] = Hash.new { |hh, kk| hh[kk] = [] } }
+      triples.each do |subject, predicate, object|
+        by_subject[subject][predicate] << object
+      end
+      by_subject.each_value do |predicates|
+        predicates.each_value(&:uniq!)
+      end
+      by_subject
+    end
+    def build_object_ref_count(triples)
+      count = Hash.new(0)
+      triples.each do |_subject, _predicate, object|
+        count[object] += 1 if object.to_s.start_with?("_:")
+      end
+      count
+    end
+    def find_inverse_roots(by_subject)
+      all_subjects = by_subject.keys
+      all_objects  = by_subject.values.flat_map { |preds| preds.values.flatten }.uniq
+      all_subjects.reject do |subject|
+        subject.start_with?("_:") || all_objects.include?(subject)
+      end.sort
+    end
+    def format_inverse_subject(subject, by_subject, ref_count, visited, depth = 1)
+      props = by_subject[subject]
+      return format_node(subject) if props.nil? || props.empty?
+      indent = "  " * (depth - 1)
+      inner  = "  " * depth
+      if subject.start_with?("_:")
+        return "[]" if visited.include?(subject)
+        visited = visited.dup
+        visited << subject
+        head = "[\n#{inner}"
+        tail = "\n#{indent}]"
+      else
+        head = "<#{subject}> "
+        tail = ""
+      end
+      body = props.keys.sort.map do |predicate|
+        objects = props[predicate].sort.map do |object|
+          format_inverse_object(object, by_subject, ref_count, visited, depth + 1)
+        end.join(", ")
+        "<#{predicate}> #{objects}"
+      end.join(";\n#{inner}")
+      head + body + tail
+    end
+    def format_inverse_object(object, by_subject, ref_count, visited, depth = 1)
+      if object.to_s.start_with?("_:") && by_subject.key?(object.to_s)
+        if ref_count[object.to_s] <= 1
+          format_inverse_subject(object.to_s, by_subject, ref_count, visited, depth)
+        else
+          object.to_s
+        end
+      else
+        format_node(object)
+      end
+    end
+    def format_node(value)
+      turtle = RDF::Turtle::Writer.new
+      if value.to_s.start_with?("_:")
+        value.to_s
+      elsif RDF::URI::IRI =~ value.to_s
+        "<#{value}>"
+      elsif value.respond_to?(:first) && value.first.kind_of?(Symbol)
+        turtle.format_literal(RDF::Literal.new(value[1], language: value[0]))
+      else
+        turtle.format_literal(value)
+      end
     end
 
     def each_data(label = :each_data)
@@ -521,6 +618,8 @@ module TTL2HTML
         end
         dir = File.dirname(file)
         FileUtils.mkdir_p(dir) if not File.exist?(dir)
+        @cache ||= {}
+        @cache[:output_turtle_files] = Set.new
         str = format_turtle(uri)
         str << format_turtle_inverse(uri)
         open(file, "w") do |io|

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: ttl2html
 version: !ruby/object:Gem::Version
-  version: 3.0.0
+  version: 3.1.0
 platform: ruby
 authors:
 - Masao Takaku
 bindir: bin
 cert_chain: []
-date: 2026-04-08 00:00:00.000000000 Z
+date: 2026-04-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri