opener-property-tagger 3.3.3 → 3.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 200701c1c3f256b0fbfa6fb284e321a16d10f399ced1fa750f7f9d4b66ed5a6f
-  data.tar.gz: 6da856c2c5beb56d778f12e93db99c6e5109728c1c89a6a8f8cceec488a556d2
+  metadata.gz: cd341e6aded0a9e7690a95a6d56ad68fd530df0b55458a54e3831b0b3a743393
+  data.tar.gz: 1ecc43a166b60c8ee56dcd5345efbbb792acfac285a4c2b1d1448c0b5aa91c79
 SHA512:
-  metadata.gz: 5d7056c547b3c7b845b4d3ad7d9f43019629c57e447b7ba5d10ada342809e708f4b62a6ac468daef61617a62e60b35391b541e9f393722d26590b531f6292559
-  data.tar.gz: 7a79912c1317fb629c95244c8f0bb9cf73690bf3b194984c8a7c818a6cc3123f8eaa5957d6bbe32db01f000457472a9db91ba7f1c3b8b556fdfcd440ff033c1a
+  metadata.gz: efb8fb35c77c6886d01929d04e29d590e25ff0ecb70e98110c85c221d4e47557bca951793ea278a994f48660e1016c4c1df08e13c552f7b601b8fb65734f3293
+  data.tar.gz: 6f51528b4be7f5eb5985720914ac555066f7105832e8459db68eb729dfa4d45b8023751f491ce06776ca9af784f27951e6225246c1305324ce2c539452940110

data/lib/opener/property_tagger.rb

@@ -1,6 +1,6 @@
 require 'open3'
 require 'slop'
-require 'oga'
+require 'nokogiri'
 require 'monitor'
 require 'httpclient'
 require 'hashie'

data/lib/opener/property_tagger/file_aspects_cache.rb

@@ -33,12 +33,17 @@ module Opener
       # @param [String] path
       #
       def load_aspects(path)
-        mapping = Hash.new { |hash, key| hash[key] = [] }
+        mapping = Hash.new{ |hash, key| hash[key] = [] }
 
-        File.foreach(path) do |line|
+        File.foreach path do |line|
           lemma, pos, aspect = line.chomp.split("\t")
+          l = Hashie::Mash.new(
+            lemma:  lemma,
+            pos:    pos,
+            aspect: aspect,
+          )
 
-          mapping[lemma.to_sym] << aspect
+          mapping[l.lemma.to_sym] << l
         end
 
         return mapping

data/lib/opener/property_tagger/processor.rb

@@ -6,7 +6,8 @@ module Opener
     class Processor
 
       attr_accessor :document
-      attr_accessor :aspects, :aspects_path, :aspects_url
+      attr_accessor :aspects_path, :aspects_url
+      attr_accessor :aspects, :lexicons
       attr_accessor :timestamp, :pretty
 
       ##
@@ -25,18 +26,19 @@ module Opener
       #  by default due to the performance overhead.
       #
       def initialize file, params: {}, url: nil, path: nil, timestamp: true, pretty: false
-        @document     = Oga.parse_xml file
+        @document     = Nokogiri.XML file
         raise 'Error parsing input. Input is required to be KAF' unless is_kaf?
         @timestamp    = timestamp
         @pretty       = pretty
 
         @params       = params
-        @cache_keys   = params[:cache_keys] || {lang: language}
         @remote       = !url.nil?
         @aspects_path = path
         @aspects_url  = url
+        @cache_keys   = params[:cache_keys] || {}
+        @cache_keys.merge! lang: @document.root.attr('xml:lang')
 
-        @aspects = if @remote then REMOTE_ASPECTS_CACHE[**@cache_keys].aspects else FILE_ASPECTS_CACHE[aspects_file] end
+        @lexicons = if @remote then REMOTE_ASPECTS_CACHE[**@cache_keys].aspects else FILE_ASPECTS_CACHE[aspects_file] end
       end
 
       ##
@@ -44,83 +46,79 @@ module Opener
       # @return [String]
       #
       def process
-        existing_aspects = extract_aspects
-
         add_features_layer
         add_properties_layer
 
-        existing_aspects.each_with_index do |(key, value), index|
+        extract_aspects.each.with_index do |(lemma, values), index|
           index += 1
 
-          add_property(key, value, index)
+          add_property lemma, values, index
         end
 
         add_linguistic_processor
 
-        return pretty ? pretty_print(document) : document.to_xml
+        pretty ? pretty_print(document) : document.to_xml
       end
 
-      ##
-      # Get the language of the input file.
-      #
-      # @return [String]
-      #
       def language
-        return @language ||= document.at_xpath('KAF').get('xml:lang')
+        @language ||= document.at_xpath('KAF').attr('xml:lang')
       end
 
-      ##
-      # Get the terms from the input file
-      # @return [Hash]
-      #
       def terms
         unless @terms
           @terms = {}
 
           document.xpath('KAF/terms/term').each do |term|
-            @terms[term.get('tid').to_sym] = term.get('lemma')
+            @terms[term.attr('tid').to_sym] = { lemma: term.attr('lemma'), text: term.attr('text')}
           end
         end
 
-        return @terms
+        @terms
       end
 
+      # Use of n-grams to determine if a unigram (1 lemma) or bigram (2
+      # lemmas) belong to a property.
+      MAX_NGRAM = 2
+
       ##
       # Check which terms belong to an aspect (property)
+      # Text have priority over Lemmas, overriding if there is a conflict
       # @return [Hash]
       #
       def extract_aspects
-        term_ids  = terms.keys
-        lemmas    = terms.values
-
-        current_token = 0
-        # Use of n-grams to determine if a unigram (1 lemma) or bigram (2
-        # lemmas) belong to a property.
-        max_ngram = 2
-
-        uniq_aspects = Hash.new { |hash, key| hash[key] = [] }
-
-        while current_token < terms.count
-          (0..max_ngram).each do |tam_ngram|
-            if current_token + tam_ngram <= terms.count
-              ngram = lemmas[current_token..current_token+tam_ngram].join(" ").downcase
-
-              if aspects[ngram.to_sym]
-                properties = aspects[ngram.to_sym]
-                ids        = term_ids[current_token..current_token+tam_ngram]
-
-                properties.uniq.each do |property|
-                  next if !property or property.strip.empty?
-
-                  uniq_aspects[property.to_sym] << [ids,ngram]
+        all_term_ids = terms.keys
+        lemmas       = terms.values
+        uniq_aspects = Hash.new{ |hash, lemma| hash[lemma] = [] }
+
+        [:lemma, :text].each do |k|
+          current_token = 0
+
+          while current_token < terms.count
+            (0..MAX_NGRAM).each do |tam_ngram|
+              next unless current_token + tam_ngram <= terms.count
+
+              ngram = lemmas[current_token..current_token+tam_ngram].map{ |a| a[k] }.join(" ").downcase
+
+              @lexicons[ngram.to_sym]&.each do |l|
+                properties = if l.aspects.present? then l.aspects else [l.aspect] end
+                properties.each do |p|
+                  next if p.blank?
+                  term_ids = all_term_ids[current_token..current_token+tam_ngram]
+                  next if uniq_aspects[p.to_sym].find{ |v| v.term_ids == term_ids }
+
+                  uniq_aspects[p.to_sym] << Hashie::Mash.new(
+                    term_ids: term_ids,
+                    ngram:    ngram,
+                    lexicon:  l,
+                  )
                 end
               end
             end
+            current_token += 1
           end
-          current_token += 1
         end
 
-        return Hash[uniq_aspects.sort]
+        Hash[uniq_aspects.sort]
       end
 
       ##
@@ -140,25 +138,25 @@ module Opener
         new_node("properties", "KAF/features")
       end
 
-      def add_property(key, value, index)
+      def add_property lemma, values, index
         property_node = new_node("property", "KAF/features/properties")
 
-        property_node.set('lemma', key.to_s)
-        property_node.set('pid', "p#{index.to_s}")
+        property_node['lemma'] = lemma.to_s
+        property_node['pid']   = "p#{index.to_s}"
 
         references_node = new_node("references", property_node)
 
-        value.uniq.each do |v|
-          comment = Oga::XML::Comment.new(:text => " #{v.last} ")
-
-          references_node.children << comment
+        values.each do |v|
+          comm_node = Nokogiri::XML::Comment.new(references_node, " #{v.ngram} ")
+          references_node.add_child comm_node
 
-          span_node = new_node("span", references_node)
+          span_node = new_node 'span', references_node
 
-          v.first.each do |val|
-            target_node = new_node("target", span_node)
+          v.term_ids.each do |id|
+            target_node       = new_node 'target', span_node
 
-            target_node.set('id', val.to_s)
+            target_node['id'] = id.to_s
+            target_node['lexicon-id'] = v.lexicon.id if v.lexicon.id
           end
         end
       end
@@ -169,19 +167,19 @@ module Opener
         version     = '2.0'
 
         node = new_node('linguisticProcessors', 'KAF/kafHeader')
-        node.set('layer', 'features')
+        node['layer'] = 'features'
 
         lp_node = new_node('lp', node)
 
-        lp_node.set('version', "#{last_edited}-#{version}")
-        lp_node.set('name', description)
+        lp_node['version'] = "#{last_edited}-#{version}"
+        lp_node['name']    = description
 
         if timestamp
           format = '%Y-%m-%dT%H:%M:%S%Z'
 
-          lp_node.set('timestamp', Time.now.strftime(format))
+          lp_node['timestamp'] = Time.now.strftime(format)
         else
-          lp_node.set('timestamp', '*')
+          lp_node['timestamp'] = '*'
         end
       end
 
@@ -200,7 +198,7 @@ module Opener
         formatter.compact = true
         formatter.write(doc, out)
 
-        return out.strip
+        out.strip
       end
 
       protected
@@ -212,11 +210,11 @@ module Opener
           parent_node = parent
         end
 
-        node = Oga::XML::Element.new(:name => tag)
+        node = Nokogiri::XML::Element.new(tag, document)
 
-        parent_node.children << node
+        parent_node.add_child node
 
-        return node
+        node
       end
 
       ##
@@ -224,7 +222,7 @@ module Opener
       # @return [Boolean]
       #
       def is_kaf?
-        return !!document.at_xpath('KAF')
+        !!document.at_xpath('KAF')
       end
 
       ##

data/lib/opener/property_tagger/remote_aspects_cache.rb

@@ -17,10 +17,11 @@ module Opener
       end
 
       def [] **params
+        existing = @cache[params]
+        return existing if existing and existing.from > UPDATE_INTERVAL.ago
+
         synchronize do
-          existing = @cache[params]
-          break existing if existing and existing.from > UPDATE_INTERVAL.ago
-          @cache[params] = cache_update existing, **params
+          @cache[params] = cache_update @cache[params], **params
         end
       end
       alias_method :get, :[]
@@ -49,7 +50,10 @@ module Opener
         lexicons = lexicons['data'].map{ |l| Hashie::Mash.new l }
         mapping  = Hash.new{ |hash, key| hash[key] = [] }
         lexicons.each do |l|
-          mapping[l.lemma.to_sym] << l.aspect
+          mapping[l.lemma.to_sym] << l
+          l.variants&.each do |v|
+            mapping[v.lemma.to_sym] << l
+          end
         end
 
         mapping

data/lib/opener/property_tagger/version.rb

@@ -1,7 +1,7 @@
 module Opener
   class PropertyTagger
 
-    VERSION = '3.3.3'
+    VERSION = '3.4.1'
 
   end
 end

data/opener-property-tagger.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |gem|
   gem.add_dependency 'opener-webservice', '~> 2.1'
   gem.add_dependency 'opener-core', '~> 2.2'
 
-  gem.add_dependency 'oga', ['~> 1.0', '>= 1.3.1']
+  gem.add_dependency 'nokogiri'
   gem.add_dependency 'httpclient'
   gem.add_dependency 'hashie'
   gem.add_dependency 'activesupport'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: opener-property-tagger
 version: !ruby/object:Gem::Version
-  version: 3.3.3
+  version: 3.4.1
 platform: ruby
 authors:
 - development@olery.com
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-11-02 00:00:00.000000000 Z
+date: 2021-02-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: opener-daemons
@@ -53,25 +53,19 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '2.2'
 - !ruby/object:Gem::Dependency
-  name: oga
+  name: nokogiri
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.0'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.3.1
+        version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.0'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.3.1
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: httpclient
   requirement: !ruby/object:Gem::Requirement
@@ -173,9 +167,9 @@ dependencies:
 description: Property tagger for hotels in Dutch and English.
 email:
 executables:
-- property-tagger-server
 - property-tagger
 - property-tagger-daemon
+- property-tagger-server
 extensions: []
 extra_rdoc_files: []
 files: