r_nlp 0.1.4 → 0.1.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 10e2b17b4d37d3b4ee59f74a7bead3795d8e0093
-  data.tar.gz: 0b696821a20ef70111c46f3322aacdc3573b1759
+  metadata.gz: 3d61657fc30af65e20df603b8ab79b6a85671d02
+  data.tar.gz: 97517fb59114c31571d9af656428c483583b113c
 SHA512:
-  metadata.gz: 38812887a46335e36b718b3f494eab97861c3100db68edb14f8b53d3df34d305b449fdadd9fff75668e3415a7c4e0a53ff0f8b76f57c57d25d9cbb853db2b9b2
-  data.tar.gz: bf4db7dd32390a4f5760282106913b50172c2c10fc0d7657e02bca58fc922b39ce9fcf5940368680bd8d61c33bb88a503e5f77c631b56fea1143f39845e6560c
+  metadata.gz: 7eca6f24c111164ca86573a1edc0ff34863e93f6c321b5139fa649408a70c926389ad817ed2a957fab5735297165cc17f0f8a5b142ef363199dd840330e4d7b1
+  data.tar.gz: 8d0a61c57d946406ee892fac514a1ccfaf74a0a73edc8196a21747fd28c674a495d0c4e2fdd8135c707ea161aea813ae8734756f2d698d6a60c30080c237a52b

data/lib/r_nlp/idf.rb

@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+
+module RNlp
+  class Idf
+    # compatible with ja or en
+    attr_reader :lang
+    def initialize(lang)
+      @lang = lang
+      unless lang == 'ja' || lang == 'en'
+        puts "#{@lang} is not compatible language\nlang should be 'ja' or 'en'"
+        exit
+      end
+    end
+    # documents should be array of string
+    def calc_idf(word, documents)
+      @word = word
+      @documents = documents
+      n = @documents.size
+      df = 0.0
+      @documents.each do |document|
+        df += 1 if document =~ /#{@word}/
+      end
+      idf = Math.log2(n/df) + 1
+      return idf
+    end
+  end
+end

data/lib/r_nlp/tf.rb

@@ -7,6 +7,10 @@ module RNlp
     attr_reader :lang
     def initialize(lang)
       @lang = lang
+      unless lang == 'ja' || lang == 'en'
+        puts "lang #{@lang} is not compatible."
+        exit
+      end
     end
     def count(text)
       tf = Hash.new
@@ -23,7 +27,7 @@ module RNlp
           end
         end
       elsif @lang == 'en'
-        text.split(" ").each do |line|
+        text.split("\n").each do |line|
           line.split(" ").each do |word|
             if tf[word] == nil
               tf[word] = 1
@@ -32,9 +36,6 @@ module RNlp
             end
           end
         end
-      else
-        puts "lang #{@lang} is not compatible."
-        exit
       end
       return tf
     end

data/lib/r_nlp/version.rb

@@ -1,3 +1,3 @@
 module RNlp
-  VERSION = "0.1.4"
+  VERSION = "0.1.5"
 end

data/lib/r_nlp.rb

@@ -1,5 +1,6 @@
 require "r_nlp/version"
 require "r_nlp/tf"
+require 'r_nlp/idf'
 
 module RNlp
   # Your code goes here...

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: r_nlp
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.1.5
 platform: ruby
 authors:
 - himkt
@@ -70,6 +70,7 @@ files:
 - bin/console
 - bin/setup
 - lib/r_nlp.rb
+- lib/r_nlp/idf.rb
 - lib/r_nlp/tf.rb
 - lib/r_nlp/version.rb
 - r_nlp.gemspec