Split sentences on emojis

lib/keyphrase.rb

@@ -9,7 +9,7 @@ class Keyphrase
   CLEAN_REGEX = /([^\p{L}a-zA-Z0-9\'\- \.]|(?<!\w)\.)/ # don't remove ' because it might be part of a stop word
   BLACKLIST_REGEX = /(?:^|\s)[^a-zA-Z\p{L}0-9]+\b|\'|\-/ # remove words with no letters, ie 123.23.12. And last chance to remove ' and -
   CLEAN_SPACES_REGEX = /^[0-9\s\.]+$|\s+/ # last phase. Remove extra whitespace and lone numbers
-  SENTENCES_REGEX = /[+!?,;:&\[\]\{\}\<\>\=\/\n\t\\"\\(\\)\u2019\u2013\|]|-(?!\w)|'(?=s)|(?<!\s)\.(?![a-zA-Z0-9])|(?<!\w)\#(?=\w)/u
+  SENTENCES_REGEX = /[+!?,;:&\[\]\{\}\<\>\=\/\n\t\\"\\(\\)\u2019\u2013\|]|-(?!\w)|'(?=s)|(?<!\s)\.(?![a-zA-Z0-9])|(?<!\w)\#(?=\w)|\p{Extended_Pictographic}+/u
 
   def self.analyse text, options={}
     @@keyphrase ||= Keyphrase.new

lib/keyphrase/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class Keyphrase
-  VERSION = "0.2.2"
+  VERSION = "0.2.3"
 end

spec/keyphrase_spec.rb

@@ -109,6 +109,12 @@
         expect(result.keys).to eq ["你好 我一方面什麼 這麼好"]
       end
 
+      it 'should split on emoji' do
+        result = Keyphrase.analyse "Study Ghibli Music For 2 Hours 🍀 Heal Your Body With Ghibli, Relax, Work, Sleep Deeply"
+
+        expect(result.keys).to eq ["Study Ghibli Music", "Sleep Deeply", "Ghibli", "2 Hours", "Heal", "Body", "Relax"]
+      end
+
     end
 
   end