iev 0.4.0 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 382780f439ab289a11d44053962bb7dbf043688fd68cb2328b8746429c19c6d3
-  data.tar.gz: 72151b973c479bfc82e61b3d0da4648f11ac54e9344b8c71e3bbf5d91a3a3c3b
+  metadata.gz: 3d7d47000d44c3dcf02934083f9d05b0a195491d6ef130cdc9ca198feeb0d8ad
+  data.tar.gz: 84b2da3efd17c94b8278bdbe963f1d356433fb0d20ccc6e06e1c2bffd05e8851
 SHA512:
-  metadata.gz: d36fad5c1d853d39ffa5d8d67e98ecaedd3c40206237706ccb177ff14957ebea9616420bd41878b84ac34cdab1e9c5c7890143024b0a3070fb6c8a994d769518
-  data.tar.gz: 8c5e7f0a9510fd562fc92db4ba225e5f9b7580e0e703525ec062b3d404ab57c804da824f94af0e3047738033577aaae8394970d47c349b1ac6b1979f7fbfa02d
+  metadata.gz: ba7fbe8bbdcbdec627b43aa62288d2b282a5476a2616ae74a56afcc81c7b983c0190dd43e92378024ad11155c0c9f9b5cf2f2ae4042c4c2536c68cecb4380262
+  data.tar.gz: ee080f35c1d17e1309217def4db29acc019e58c5ad3d104efbfc351155addc43c38a1e0602033024d213e91ce13692084f868cc1f1b64521145d90a763ae2f85

data/.github/workflows/rake.yml

@@ -1,7 +1,7 @@
 name: rake
 
 permissions:
-  contents: read
+  contents: write
 
 on:
   push:

data/lib/iev/cli/command.rb

@@ -34,6 +34,8 @@ module Iev
                              desc: "Enables debug messages about authoritative sources recognition"
       option :debug_relaton, type: :boolean, default: false,
                              desc: "Enables debug messages about Relaton integration"
+      option :relaton, type: :boolean, default: false,
+                       desc: "Fetch source URLs via Relaton (slow, makes network requests)"
       def export(file)
         handle_generic_options(options)
 
@@ -42,6 +44,7 @@ module Iev
           output_dir: options[:output],
           only_concepts: options[:only_concepts],
           only_languages: options[:only_languages],
+          fetch_relaton_links: options[:relaton],
         ).export
 
         info "Done!"

data/lib/iev/converter/mathml_to_asciimath.rb

@@ -48,8 +48,17 @@ module Iev
           mathml_to_asciimath(input)
         end
 
+        # Clear the Plurimath expression cache. Call between export runs.
+        def clear_cache
+          @math_cache = nil
+        end
+
         private
 
+        def math_cache
+          @math_cache ||= {}
+        end
+
         def mathml_to_asciimath(input)
           return input unless input&.match?(/<|&/)
 
@@ -58,17 +67,18 @@ module Iev
           to_asciimath = Nokogiri::HTML.fragment(input, "UTF-8")
 
           to_asciimath.css("math").each do |math_element|
-            asciimath = Plurimath::Math.parse(
-              math_element.to_xml, :mathml
-            ).to_asciimath.strip
+            math_xml = math_element.to_xml
+            asciimath = math_cache[math_xml] ||= begin
+              Plurimath::Math.parse(math_xml, :mathml).to_asciimath.strip
+            rescue Plurimath::Math::ParseError
+              ""
+            end
 
             if asciimath.empty?
               math_element.remove
             else
               math_element.replace "stem:[#{asciimath}]"
             end
-          rescue Plurimath::Math::ParseError
-            math_element.remove
           end
 
           html_to_asciimath(
@@ -79,6 +89,13 @@ module Iev
         def html_to_asciimath(input)
           return input if input.nil? || input.empty?
 
+          # Fast path: if no HTML elements remain that need Nokogiri processing
+          # (after parse_anchor_tag handles <i>/<sub>/<sup>/<ol>/<ul>/<font>),
+          # just do the Greek entity replacement.
+          unless input.match?(/<([iI]|sub|sup|ol|ul|font)\b/)
+            return html_entities_to_stem(input)
+          end
+
           to_asciimath = Nokogiri::HTML.fragment(input, "UTF-8")
 
           to_asciimath.css("i").each do |math_element|

data/lib/iev/exporter.rb

@@ -27,12 +27,15 @@ module Iev
     # @param output_dir [String, Pathname] destination for YAML files
     # @param only_concepts [String, nil] SQL LIKE pattern for IEVREF filtering
     # @param only_languages [String, nil] comma-separated language codes
+    # @param fetch_relaton_links [Boolean] whether to fetch source URLs via Relaton
     def initialize(input_path, output_dir: Dir.pwd,
-                   only_concepts: nil, only_languages: nil)
+                   only_concepts: nil, only_languages: nil,
+                   fetch_relaton_links: false)
       @input_path = Pathname.new(input_path)
       validate_input!
 
       @output_dir = Pathname.new(output_dir)
+      @fetch_relaton_links = fetch_relaton_links
       @filters = {
         only_concepts: only_concepts,
         only_languages: only_languages,
@@ -102,15 +105,28 @@ module Iev
     end
 
     def build_collection(dataset)
-      Glossarist::ManagedConceptCollection.new.tap do |collection|
-        dataset.each do |row|
-          term = TermBuilder.build_from(row)
-          next unless term
+      SourceParser.relaton_enabled = @fetch_relaton_links
 
-          concept = collection.fetch_or_initialize(term.id)
-          concept.add_l10n(term)
+      # Use a hash index for O(1) concept lookup instead of
+      # Glossarist's O(n) fetch_or_initialize which does linear scan.
+      concept_index = {}
+      collection = Glossarist::ManagedConceptCollection.new
+
+      dataset.each do |row|
+        term = TermBuilder.build_from(row)
+        next unless term
+
+        concept = concept_index[term.id] ||= begin
+          c = Glossarist::ManagedConcept.new(data: { "id" => term.id })
+          collection.store(c)
+          c
         end
+        concept.add_l10n(term)
       end
+
+      collection
+    ensure
+      SourceParser.relaton_enabled = true
     end
 
     def save_collection(collection)

data/lib/iev/source_parser.rb

@@ -14,6 +14,12 @@ module Iev
     include Utilities
     using DataConversions
 
+    # When false, obtain_source_link skips Relaton network calls.
+    @relaton_enabled = true
+    class << self
+      attr_accessor :relaton_enabled
+    end
+
     attr_reader :src_split, :parsed_sources, :raw_str, :src_str
 
     def initialize(source_str, term_domain)
@@ -89,7 +95,7 @@ module Iev
         origin: origin,
         modification: relationship[:modification],
       )
-    rescue ::RelatonBib::RequestError => e
+    rescue ::RelatonBib::RequestError, Socket::ResolutionError, SocketError => e
       warn e.message
     end
 
@@ -347,10 +353,11 @@ module Iev
 
     # Uses Relaton to obtain link for given source ref.
     def obtain_source_link(ref)
+      return nil unless self.class.relaton_enabled
       return nil unless defined?(RelatonDb)
 
       RelatonDb.instance.fetch(ref)&.url
-    rescue ::RelatonBib::RequestError => e
+    rescue ::RelatonBib::RequestError, Socket::ResolutionError, SocketError => e
       warn e.message
       nil
     end

data/lib/iev/term_builder.rb

@@ -298,6 +298,8 @@ module Iev
     end
 
     def strip_html_comments(str)
+      return str unless str.include?("<!--")
+
       doc = Nokogiri::HTML::DocumentFragment.parse(str)
       comments = doc.children.select(&:comment?)
       return str if comments.empty?

data/lib/iev/utilities.rb

@@ -13,10 +13,14 @@ module Iev
 
     def parse_anchor_tag(text, term_domain)
       return nil if text.nil?
+      return text unless text.include?("<")
 
       text = process_simg_figures(text, term_domain)
       text = fix_unquoted_href(text)
 
+      # Second check: regex substitutions may have consumed all tags
+      return text unless text.include?("<")
+
       doc = Nokogiri::HTML::DocumentFragment.parse(text)
       nodes_to_adoc(doc.children, term_domain)
     end
@@ -86,11 +90,45 @@ module Iev
         "#{IMAGE_PATH_PREFIX}/#{term_domain}/#{src}[]"
       when "p", "div", "span"
         inner
+      when "i"
+        convert_italic(inner)
+      when "sub"
+        inner.empty? ? "" : "~#{inner}~"
+      when "sup"
+        inner.empty? ? "" : "^#{inner}^"
+      when "ol"
+        convert_list(node, ". ")
+      when "ul"
+        convert_list(node, "* ")
+      when "li"
+        inner
+      when "font"
+        convert_font(node, inner)
       else
         node.to_s
       end
     end
 
+    def convert_italic(text)
+      case text.length
+      when 0
+        ""
+      when 1..12
+        "stem:[#{text}]"
+      else
+        "_#{text}_"
+      end
+    end
+
+    def convert_list(node, prefix)
+      node.css("li").map { |li| "#{prefix}#{li.text}" }.join
+    end
+
+    def convert_font(node, inner)
+      style = node["style"].to_s
+      style.include?("sans-serif") ? "`#{inner}`" : inner
+    end
+
     def convert_link(node, inner)
       href = (node["href"] || "").to_s.strip
 

data/lib/iev/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Iev
-  VERSION = "0.4.0"
+  VERSION = "0.4.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: iev
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.4.1
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-04-27 00:00:00.000000000 Z
+date: 2026-04-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: creek