thomas 0.1.4 → 0.1.5

data/lib/thomas.rb

@@ -26,11 +26,16 @@ module Adocca
         @word = word
         @amount = amount
       end
+      def adjust(n)
+        @amount += n
+      end
       def execute
 	model = Module.const_get(@model_name)
-        instance = model.find_by_corpus_and_word(@corpus, @word)
-        instance = model.create(:corpus => @corpus, :word => @word) unless instance
-        instance.update_attribute(:amount, (instance.amount || 0) + @amount)
+        if model.connection.update(model.sanitizeSQL(["UPDATE #{model.table_name} SET amount = amount + ? WHERE corpus = ? AND word = ?",
+                                                      @amount, @corpus, @word])) == 0
+          model.create(:corpus => @corpus, :word => @word, :amount => @amount)
+        end
+        model.expire_cached_namespace('CachedSupermodel:' + model.name + ':find_by_corpus_and_word:' + [@corpus, @word].inspect)
       end
     end
 
@@ -38,7 +43,7 @@ module Adocca
 
       attr_accessor :model
 
-      @@updates = []
+      @@updates = {}
       
       def initialize(model)
         @model = model
@@ -67,14 +72,24 @@ module Adocca
 
       def train(corpus, text, options = {})
         tokenize(text, options).each do |word|
-          @@updates << Update.new(@model, corpus, word, 1)
+          @@updates[corpus] ||= {}
+          if update = @@updates[corpus][word]
+            update.adjust(1)
+          else
+            @@updates[corpus][word] = Update.new(@model, corpus, word, 1)
+          end
           flush_updates
         end
       end
 
       def untrain(corpus, text, options = {})
         tokenize(text, options).each do |word|
-          @@updates << Update.new(@model, corpus, word, -1)
+          @@updates[corpus] ||= {}
+          if update = @@updates[corpus][word]
+            update.adjust(-1)
+          else
+            @@updates[corpus][word] = Update.new(@model, corpus, word, -1)
+          end
           flush_updates
         end
       end
@@ -104,14 +119,23 @@ module Adocca
 
       protected
 
+      def updates_needed
+        rval = 0
+        @@updates.each do |corpus, words|
+          rval += words.size
+        end
+        return rval
+      end
+
       def flush_updates
-        if @@updates.size > 200 || (defined?($FLUSH_THOMAS) && $FLUSH_THOMAS)
+        if (defined?($FLUSH_THOMAS) && $FLUSH_THOMAS) || updates_needed > 200
           retries = 0
           begin
             @model.transaction do
-              my_updates = @@updates.clone
-              until my_updates.empty?
-                my_updates.pop.execute
+              @@updates.clone.each do |corpus, words|
+                words.each do |s, update|
+                  update.execute
+                end
               end
             end
           rescue ActiveRecord::StatementInvalid => e

metadata

@@ -1,10 +1,10 @@
 --- !ruby/object:Gem::Specification 
-rubygems_version: 0.8.11
+rubygems_version: 0.9.2
 specification_version: 1
 name: thomas
 version: !ruby/object:Gem::Version 
-  version: 0.1.4
-date: 2007-01-30 00:00:00 +01:00
+  version: 0.1.5
+date: 2007-04-11 00:00:00 +02:00
 summary: Bayes classifier using memcached
 require_paths: 
 - lib
@@ -25,6 +25,7 @@ required_ruby_version: !ruby/object:Gem::Version::Requirement
 platform: ruby
 signing_key: 
 cert_chain: 
+post_install_message: 
 authors: 
 - Adocca Entertainment AB
 files: