word_count_analyzer 0.0.10 → 0.0.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 9100a7be1a916cb3872aa74e9e76ddfb83744d62
-  data.tar.gz: 99b5f6ddb28101de71f97d3b44e179d485f7ab84
+  metadata.gz: f1aca10a5014f22128759d78ac94e1344dde8056
+  data.tar.gz: 83ef54c2fb432cab8246aecda21cefdce01f0c82
 SHA512:
-  metadata.gz: e0727cd58ddf83c23ac0dde9f9c713d7cad4f5c21fd572bc0bd21064266f57e5ff3d2489326d951ec5b7e3a7d38370a36f98596bd725d08092ad9fdabea43480
-  data.tar.gz: decc5f7b4256f0884b61bf23191ae638509a68f948a2515333ddbef2a9e6d3d6c9004ebb1e5db1352a11a1212399497cd798bb559d1f91bb14b778342ec123c0
+  metadata.gz: da0391470b4bffa94c4830e0aa52066ee9ab8b1792b06c5e14049e08527a7c7321ee7f658e4ee6c90b6b9ce1aa63b080d76b2161b244c6870b3814cc803b062c
+  data.tar.gz: cb77e8ceda65956f77c9d3aabd33e8e39ec4d682e6cc215ac0ac48cb4ada8e9e55afc0bbed8d6f830a9d6e8e37635148f7fd3d4ce08d735dc9050ed9b80f9630

data/lib/word_count_analyzer/counter.rb

@@ -1,6 +1,6 @@
 module WordCountAnalyzer
   class Counter
-    attr_reader :text, :ellipsis, :hyperlink, :contraction, :hyphenated_word, :date, :number, :numbered_list, :xhtml, :forward_slash, :backslash, :dotted_line, :dashed_line, :underscore, :stray_punctuation, :tgr
+    attr_reader :text, :ellipsis, :hyperlink, :contraction, :hyphenated_word, :date, :number, :numbered_list, :xhtml, :forward_slash, :backslash, :dotted_line, :dashed_line, :underscore, :stray_punctuation, :equal_sign, :tgr
     def initialize(text:, **args)
       @text = text
       @ellipsis = args[:ellipsis] || 'ignore'
@@ -17,6 +17,7 @@ module WordCountAnalyzer
       @dashed_line = args[:dashed_line] || 'ignore'
       @underscore = args[:underscore] || 'ignore'
       @stray_punctuation = args[:stray_punctuation] || 'ignore'
+      @equal_sign = 'ignore'
       @tgr = EngTagger.new
     end
 
@@ -39,6 +40,7 @@ module WordCountAnalyzer
       @dashed_line = 'ignore'
       @underscore = 'ignore'
       @stray_punctuation = 'ignore'
+      @equal_sign = 'break'
       word_count
     end
 
@@ -77,6 +79,7 @@ module WordCountAnalyzer
       processed_text = process_dashed_line(processed_text)
       processed_text = process_underscore(processed_text)
       processed_text = process_stray_punctuation(processed_text)
+      processed_text = process_equal_sign(processed_text) if @equal_sign.eql?('break')
       processed_text.split(/\s+/).reject(&:empty?).size
     end
 
@@ -226,5 +229,9 @@ module WordCountAnalyzer
         raise 'The value you specified for stray_punctuation is not a valid option. Please use either `ignore` or `count`. The default option is `ignore`'
       end
     end
+
+    def process_equal_sign(txt)
+      txt.split('=').join(' ').split(/>(?=[a-zA-z]+)/).join(' ')
+    end
   end
 end

data/lib/word_count_analyzer/version.rb

@@ -1,3 +1,3 @@
 module WordCountAnalyzer
-  VERSION = "0.0.10"
+  VERSION = "0.0.11"
 end

data/spec/word_count_analyzer/counter_spec.rb

@@ -599,6 +599,12 @@ RSpec.describe WordCountAnalyzer::Counter do
       ws = WordCountAnalyzer::Counter.new(text: text)
       expect(ws.pages_count).to eq(9)
     end
+
+    it 'reverse engineers Pages word count #005' do
+      text = "<span class=\"large-text\">Hello world</span> <new-tag>Hello</new-tag>"
+      ws = WordCountAnalyzer::Counter.new(text: text)
+      expect(ws.pages_count).to eq(12)
+    end
   end
 
   context 'Microsoft Word Count' do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: word_count_analyzer
 version: !ruby/object:Gem::Version
-  version: 0.0.10
+  version: 0.0.11
 platform: ruby
 authors:
 - Kevin S. Dias