wordfinder 0.0.1 → 0.0.2

data/lib/word_finder.rb

@@ -5,9 +5,9 @@
 # See LICENSE file for details.
 ##
 
-require "ffi"
-require "ffi/aspell"
 require "word_finder/version"
+require "word_finder/checker"
+require "word_finder/node"
 
 module WordFinder
   class << self
@@ -24,37 +24,9 @@ module WordFinder
     # @return [Array<String>]
     #   an Array containing all words found, in order
     def words_in(phrase, options = {})
-      speller = FFI::Aspell::Speller.new(options[:lang] || "en_US")
-
-      matched   = []
-      unmatched = ""
-
-      phrase.each_char do |c|
-        unmatched << c
-
-        tmp, matched = matched, []
-
-        if tmp.empty?
-          if speller.correct?(unmatched)
-            matched.push(unmatched)
-            unmatched = ""
-          end
-        else
-          (0..tmp.length).each do |i|
-            word = tmp[i..-1].join + unmatched
-
-            if speller.correct?(word)
-              matched.push(word)
-              unmatched = ""
-              break
-            elsif tmp[i]
-              matched << tmp[i]
-            end
-          end
-        end
-      end
-
-      matched.reject{|m| m !~ /\w/} # remove punctuation marks
+      Node.new(options).tap{|node|
+        phrase.each_char(&node.method(:insert))
+      }.words.reject{|m| m !~ /\w/}
     end
   end
 end

data/lib/word_finder/checker.rb

@@ -0,0 +1,24 @@
+##
+# WordFinder.
+# Copyright © 2012 Chris Corbyn.
+#
+# See LICENSE file for details.
+##
+
+require "ffi"
+require "ffi/aspell"
+
+module WordFinder
+  # Caching subclass of Aspell.
+  class Checker < FFI::Aspell::Speller
+    def initialize(*)
+      super
+      @cache = {}
+    end
+
+    def word?(str)
+      @cache[str] = correct?(str) unless @cache.key?(str)
+      @cache[str]
+    end
+  end
+end

data/lib/word_finder/node.rb

@@ -0,0 +1,65 @@
+##
+# WordFinder.
+# Copyright © 2012 Chris Corbyn.
+#
+# See LICENSE file for details.
+##
+
+module WordFinder
+  # A node in the search tree.
+  #
+  # A search is represented by a tree of nodes.
+  # Each node represents a branch of valid words.
+  class Node
+    attr_reader :offset
+
+    def initialize(options = {})
+      @words    = []
+      @buf      = ""
+      @children = []
+      @offset   = 0
+      @checker  = Checker.new(options[:lang] || "en_US")
+    end
+
+    def insert(char)
+      @children.each{|c| c.insert(char)}
+
+      @buf << char
+
+      if @checker.word?(@buf)
+        @words << @buf
+        @buf = ""
+        condense
+      end
+    end
+
+    def condense
+      (@offset...(@words.length - 1)).each do |i|
+        (@words[i..-1].join).tap do |word|
+          if @checker.word?(word)
+            @children << dup
+            @words[i..-1] = word
+            @buf = ""
+            @offset = @words.length
+            return
+          end
+        end
+      end
+    end
+
+    def words
+      (best = @words)
+      @children.each do |c|
+        other = c.words
+        best = other if other.length < best.length
+      end
+      best
+    end
+
+    def initialize_copy(parent)
+      @words    = parent.instance_variable_get(:@words).dup
+      @buf      = parent.instance_variable_get(:@buf).dup
+      @children = []
+    end
+  end
+end

data/lib/word_finder/version.rb

@@ -6,5 +6,5 @@
 ##
 
 module WordFinder
-  VERSION = "0.0.1"
+  VERSION = "0.0.2"
 end

data/spec/wordfinder_spec.rb

@@ -67,5 +67,13 @@ describe WordFinder do
         words.should == %w[I need it now]
       end
     end
+
+    context "with ambiguous constructions" do
+      let(:phrase) { "threelittlepigswenttomarket" }
+
+      it "uses the fewest words possible" do
+        words.should == %w[three little pigs went to market]
+      end
+    end
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: wordfinder
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-12-08 00:00:00.000000000 Z
+date: 2012-12-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -73,6 +73,8 @@ files:
 - README.md
 - Rakefile
 - lib/word_finder.rb
+- lib/word_finder/checker.rb
+- lib/word_finder/node.rb
 - lib/word_finder/version.rb
 - lib/wordfinder.rb
 - spec/spec_helper.rb