lite 0.0.5 → 0.0.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 883e359e761b59f56fa7fd759d9a83b37d61aa4a
-  data.tar.gz: 8189eebfd8e3aa1ca2b225d6bbcd46cca7c918f3
+  metadata.gz: 618acdff29331d957b2efd4c4774b7681938296d
+  data.tar.gz: c64f0d41f19ad4427974577d4c16a5baf75c7780
 SHA512:
-  metadata.gz: ba14c42f2cbc1fc769de835eb4a6e644a383db5b3edc7dd2bb06ad811851814c266ab359164d08bce90955c0118f245c3c3c2431591a64c2e06e4732923e4a7e
-  data.tar.gz: 2af2cf2adf7db9f0118cedcf72734e51bf68fe3f37d41d93a2614b7833e403232b77df85b3c066001ece7cba3d60d7f1d131cbb3eb1b881d8cd1ab43ab2703d5
+  metadata.gz: 36d9bdfa99e5eaeed2986141866e9824863f3a34a952395f00a143fc51722122e7c893c0979a70ba8a5a64cc5b126e9a9cbd053933fd9c1babc008bbc782642e
+  data.tar.gz: 86193ff687c02a9c1cb80e849a294b6766fc1d324a2207f36679206daab77853394f7ec6a2102ee8f0451436b47d5913ce3cfd1b2cf723f9f3396261066a92ba

data/lib/lite/vmm.rb

@@ -1,4 +1,4 @@
-require "json"
+require "msgpack"
 
 module VMM
 
@@ -6,50 +6,51 @@ class PPM
 
   def initialize ab,d=5
     @trie = Trie.new
-    @ab = ab
+    @ab = ABet.new ab
     @d = d
   end
 
   def learn str
-    (str.size - @d-1 ).times{|i| @trie.grow(str[i..i+@d-1],str[i+@d]) }
+    (str.size - @d-1 ).times{|i| @trie.grow(str[i..i+@d-1].chars.map{|sym| @ab.sym_to_i( sym )}, @ab.sym_to_i( str[i+@d] ) ) }
   end
 
   def log_eval str
     (str.size - @d).times.inject(0.0) do |agg,i|
-      agg += path_predict( str[@d], @trie.path( str[i..i+@d-1] ) )
+      agg += path_predict( @ab.sym_to_i(str[@d]) , @trie.path( str[i..i+@d-1].chars.map{|sym| @ab.sym_to_i(sym)} ) )
     end
-
   end
 
   def path_predict sym, path
     path.reverse.inject( 0.0 ) do |agg, context|
       agg += Math.log( single_pr( sym, context ), 2.0 )
-      break( agg ) if context===@trie.root or !context[:c].has_key?(sym)
+      break( agg ) if context===@trie.root or !context[0].has_key?( sym )
       agg
     end
   end
 
   def single_pr sym, context
-    ( context[:v].has_key?( sym ) ? context[:v][sym] : context[:v].size ) /  
-      (context[:v].values.inject(:+) + context[:v].size).to_f  
+    ( context[1].has_key?( sym ) ? context[1][sym] : context[1].size ) /  
+      (context[1].values.inject(:+) + context[1].size).to_f  
   end
 
-
-  def to_json 
-    { :trie => @trie.root, :ab => @ab, :d => @d }.to_json
+  def to_file file_path 
+    msg = { 'trie' => @trie.root, 'ab' => @ab.sym_arr, 'd' => @d }.to_msgpack
+    out = File.new( file_path, "w" )
+    out.print msg
+    out.close
   end
 
-  def self.load json
-    model = JSON.parse( json, :symbolize_names => true)
-    ppm = PPM.new( model[:ab],model[:d] )
-    ppm.instance_variable_set( :@trie, Trie.new( model[:trie] ) )
+  def self.from_file file_path
+    model = MessagePack.unpack( File.new( file_path,"r").readlines.join )
+    ppm = PPM.new( model['ab'],model['d'] )
+    ppm.instance_variable_set( :@trie, Trie.new( model['trie'] ) )
     ppm
   end
 end
 
 class Trie
   def initialize(root=nil)
-    @root = root.nil? ? ({ :c => {}, :v => { } }) : root
+    @root = root.nil? ? ( [{}, {} ]) : root
   end
 
   def root
@@ -58,29 +59,51 @@ class Trie
 
   def grow(context, symbol) 
     node = @root
-    @root[:v][symbol]||=0
-    @root[:v][symbol]+=1
-    context.each_char do |ch|
-      node[:c][ch] ||= new_node symbol
-      node[:v][symbol]||=0 
-      node[:v][symbol] += 1
-      node = node[:c][ch]
+    @root[1][symbol]||=0
+    @root[1][symbol]+=1
+    context.each do |ch|
+      node[0][ch] ||= new_node symbol
+      node[1][symbol]||=0 
+      node[1][symbol] += 1
+      node = node[0][ch]
     end
     true
   end
 
-  def path str
-    str.chars.inject([@root]) do |agg,ch|
-      next(agg) unless agg.last[:c][ch]
-      agg << agg.last[:c][ch]
+  def path sym_arr
+    sym_arr.inject([@root]) do |agg,ch|
+      next(agg) unless agg.last[0][ch]
+      agg << agg.last[0][ch]
       agg
     end
   end
 
   def new_node( v )
-    { :c => {}, :v => { v => 1} }
+    [{}, { v => 1}]
   end
 end
 
+class ABet
+  def initialize sym_arr
+    @ab = sym_arr
+    @sym_to_i = Hash[ sym_arr.zip( (0..sym_arr.size-1).to_a ) ]
+  end
+
+  def sym_to_i sym
+    @sym_to_i[ sym ] || @ab.size
+  end
+
+  def i_to_sym id
+    @ab[ id ] || @ab[ size ]
+  end
+
+  def sym_arr
+    @ab
+  end
+
+  def size
+    @ab.size + 1
+  end
+end
 
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: lite
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.6
 platform: ruby
 authors:
 - ronbee
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-08-24 00:00:00.000000000 Z
+date: 2013-08-28 00:00:00.000000000 Z
 dependencies: []
 description: 'lite machine learning tools: clasifier, annotator, and more'
 email: void@mailinator.com