vector_embed 0.5.0 → 0.5.1

data/CHANGELOG

@@ -1,3 +1,9 @@
+0.5.1 / 2013-07-09
+
+* Bug fixes
+
+  * Don't drop 0 labels like you would for features (sorry!)
+
 0.5.0 / 2013-07-08
 
 * Breaking changes

data/lib/vector_embed.rb

@@ -109,7 +109,7 @@ class VectorEmbed
 
   def label_maker(label)
     @label_maker || @mutex.synchronize do
-      @label_maker ||= Maker.pick(LABEL_MAKERS, 'label', label, self)
+      @label_maker ||= Maker.pick(LABEL_MAKERS, 'label', label, self, keep_zero: true)
     end
   end
 

data/lib/vector_embed/maker.rb

@@ -7,13 +7,13 @@ require 'vector_embed/maker/date'
 class VectorEmbed
   class Maker
     class << self
-      def pick(choices, k, first_v, parent)
+      def pick(choices, k, first_v, parent, options = nil)
         if (feature_types = parent.options[:features]) and (type = feature_types.detect { |kk, v| kk.to_s == k.to_s })
           klass = const_get type[1].to_sym
-          klass.new k, parent
+          klass.new k, parent, options
         elsif klass = choices.detect { |klass| klass.want?(first_v, parent) }
           parent.logger.debug { "Interpreting #{k.inspect} as #{klass.name.split('::').last} given first value #{first_v.inspect}" }
-          klass.new k, parent
+          klass.new k, parent, options
         else
           raise "Can't use #{first_v.class} for #{k.inspect} given #{first_v.inspect} and choices #{choices.inspect}"
         end
@@ -23,11 +23,13 @@ class VectorEmbed
     attr_accessor :cardinality
     attr_reader :parent
     attr_reader :k
+    attr_reader :options
 
-    def initialize(k, parent)
+    def initialize(k, parent, options = nil)
       @k = k
       @parent = parent
       @cardinality = 0
+      @options = options
     end
 
     def pairs(v)

data/lib/vector_embed/maker/ngram.rb

@@ -15,7 +15,7 @@ class VectorEmbed
       attr_reader :len
       attr_reader :delim
 
-      def initialize(k, parent)
+      def initialize(k, parent, options = nil)
         super
         @len = parent.options[:ngram_len].to_i
         raise ArgumentError, ":ngram_len must be > 0" unless @len > 0

data/lib/vector_embed/maker/number.rb

@@ -23,22 +23,29 @@ class VectorEmbed
           else
             v
           end
-          if num.nonzero?
-            '%.16g' % num
-          end
         end
       end
+      
+      FORMAT = '%.16g'
 
       def value(v)
         case v
         when Numeric, JUST_A_NUMBER, UGLY_FLOAT
-          Number.numify v
+          num = Number.numify v
+          if num.nonzero? or keep_zero?
+            FORMAT % num
+          end
         when NilClass, NULL, SLASH_N
           nil
         else
           raise ArgumentError, "Can't embed #{v.inspect} in number feature #{k.inspect}"
         end
       end
+
+      def keep_zero?
+        return @keep_zero_query if defined?(@keep_zero_query)
+        @keep_zero = options && !!options[:keep_zero]
+      end
     end
   end
 end

data/lib/vector_embed/version.rb

@@ -1,3 +1,3 @@
 class VectorEmbed
-  VERSION = '0.5.0'
+  VERSION = '0.5.1'
 end

data/spec/vector_embed_spec.rb

@@ -12,6 +12,7 @@ describe VectorEmbed do
 
     it "stores numbers as numbers" do
       v = VectorEmbed.new
+      v.line(0).should == '0'
       v.line(5.4).should == '5.4'
       v.line(-3.9).should == '-3.9'
     end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: vector_embed
 version: !ruby/object:Gem::Version
-  version: 0.5.0
+  version: 0.5.1
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-07-08 00:00:00.000000000 Z
+date: 2013-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: murmurhash3