! analyzer, + phonetics, + category doc

server/lib/picky/analyzer.rb

@@ -40,6 +40,7 @@ def analyze bundle
   end
   def cardinality identifier, index
     return if index.size.zero?
+    return unless index.respond_to?(:each_pair)
 
     key_length_average = 0
     ids_length_average = 0
@@ -49,7 +50,7 @@ def cardinality identifier, index
     min_key_length = 1.0/0 # Infinity
     max_key_length =     0
 
-    key_size, ids_size =
+    key_size, ids_size = 0, 0
     index.each_pair do |key, ids|
       key_size = key.size
       if key_size < min_key_length
@@ -91,7 +92,8 @@ def index_analysis
     end
   end
   def weights index
-    return if index.size.zero?
+    return if !index.respond_to?(:size) || index.size.zero?
+    return unless index.respond_to?(:each_pair)
 
     min_weight = 1.0/0 # Infinity
     max_weight =   0.0
@@ -131,11 +133,11 @@ def to_s
   end
   def index_to_s
     return if analysis[:__keys].zero?
-    [
-      "index key cardinality:                #{"%10d" % analysis[:__keys]}",
-      "index key length range (avg):         #{"%10s" % analysis[:index][:key_length]} (#{analysis[:index][:key_length_average].round(2)})",
-      "index ids per key length range (avg): #{"%10s" % analysis[:index][:ids_length]} (#{analysis[:index][:ids_length_average].round(2)})"
-    ].join("\n")
+    ary = ["index key cardinality:                #{"%10d" % analysis[:__keys]}"]
+    return ary.join "\n" unless analysis[:index]
+    ary << "index key length range (avg):         #{"%10s" % analysis[:index][:key_length]} (#{analysis[:index][:key_length_average].round(2)})"
+    ary << "index ids per key length range (avg): #{"%10s" % analysis[:index][:ids_length]} (#{analysis[:index][:ids_length_average].round(2)})"
+    ary.join "\n"
   end
   def weights_to_s
     return unless analysis[:weights]

server/lib/picky/backends/file/json.rb

@@ -40,6 +40,12 @@ def clear
           self.mapping.clear
         end
 
+        # Size of the index is equal to the mapping size.
+        #
+        def size
+          self.mapping.size
+        end
+
         # Loads the mapping hash from json format.
         #
         def load

server/lib/picky/backends/redis/list.rb

@@ -18,6 +18,15 @@ def clear
           end
         end
 
+        # Size of the list(s).
+        #
+        def size
+          redis_key = "#{namespace}:*"
+          client.keys(redis_key).inject(0) do |total, key|
+            total + client.zcard(key)
+          end
+        end
+
         # Deletes the list for the key.
         #
         def delete key

server/lib/picky/backends/redis/string.rb

@@ -12,6 +12,12 @@ def clear
           client.del namespace
         end
 
+        # Returns the size of the hash.
+        #
+        def size
+          client.hlen namespace
+        end
+
         # Deletes the single value.
         #
         def delete key

server/lib/picky/backends/sqlite/array.rb

@@ -10,6 +10,11 @@ def create_table
           db.execute 'create table key_value (key varchar(255), value text);'
         end
 
+        def size
+          result = db.execute 'SELECT COUNT(*) FROM key_value'
+          result.first.first.to_i
+        end
+
         def []= key, array
           unless array.empty?
             db.execute 'INSERT OR REPLACE INTO key_value (key,value) VALUES (?,?)',

server/lib/picky/bundle.rb

@@ -91,11 +91,8 @@ def create_backends
     # the strategy itself pretends to be an index.
     #
     def initialize_backends
-      # @inverted, @weights, @similarity, @configuration, @realtime = backend.initial weight_strategy
       @inverted      = @backend_inverted.initial
-      # TODO @weights = @weight_strategy.initial || @backend_weights.initial
-      #
-      @weights       = @weight_strategy.saved?? @backend_weights.initial : @weight_strategy
+      @weights       = @weight_strategy.saved? ? @backend_weights.initial : @weight_strategy
       @similarity    = @backend_similarity.initial
       @configuration = @backend_configuration.initial
       @realtime      = @backend_realtime.initial
@@ -106,8 +103,6 @@ def initialize_backends
     #
     def empty
       @inverted = @backend_inverted.empty
-      # THINK about this. Perhaps the strategies should implement the backend methods?
-      #
       @weights = @weight_strategy.saved? ? @backend_weights.empty : @weight_strategy
       @similarity = @backend_similarity.empty
       @configuration = @backend_configuration.empty

server/lib/picky/bundle_indexed.rb

@@ -25,7 +25,7 @@ class Bundle
     #
     def ids sym_or_string
       @inverted[sym_or_string] || []
-      # TODO ?
+      # THINK Place the key_format conversion here – or move into the backend?
       #
       # if @key_format
       #   class << self
@@ -79,8 +79,6 @@ def load_inverted
     # Loads the weights index.
     #
     def load_weights
-      # TODO @weights = @weight_strategy.load || @backend_weights.load
-      #
       self.weights = @backend_weights.load if @weight_strategy.saved?
     end
     # Loads the similarity index.

server/lib/picky/bundle_realtime.rb

@@ -74,6 +74,8 @@ def add_similarity str_or_sym, where = :unshift
         #
         similars.delete str_or_sym if similars.include? str_or_sym
         similars << str_or_sym
+
+        self.similarity_strategy.prioritize similars, str_or_sym
       end
     end
 

server/lib/picky/category.rb

@@ -16,13 +16,16 @@ class Category
     #  * partial: Partial::None.new, Partial::Substring.new(from:start_char, to:up_to_char) (defaults from:-3, to:-1)
     #  * similarity: Similarity::None.new (default), Similarity::DoubleMetaphone.new(amount_of_similarly_linked_words)
     #  * from: The source category identifier to take the data from.
+    #  * key_format: What this category's keys are formatted with (default is :to_i)
+    #  * backend: The backend to use. Default is Backends::Memory.new.
+    #  Other options are: Backends::Redis.new, Backends::SQLite.new, Backends::File.new.
+    #  * qualifiers: Which qualifiers can be used to predefine the category. E.g. "title:bla".
     #
     # Advanced Options:
-    #  * source: Use if the category should use a different source.
-    #  * weights: Query::Weights.new( [:category1, :category2] => +2, ... )
+    #  * weight: Query::Weights.new( [:category1, :category2] => +2, ... )
     #  * tokenizer: Use a subclass of Tokenizers::Base that implements #tokens_for and #empty_tokens.
-    #  * key_format: What this category's keys are formatted with (default is :to_i)
     #  * use_symbols: Whether to use symbols internally instead of strings.
+    #  * source: Use if the category should use a different source.
     #
     def initialize name, index, options = {}
       @name  = name

server/lib/picky/generators/similarity/phonetic.rb

@@ -36,11 +36,9 @@ def check_gem # :nodoc:
 
         # Sorts the index values in place.
         #
-        # Not used currently.
-        #
-        def prioritize! ary, code
+        def prioritize ary, code
           ary.sort_by_levenshtein! code
-          ary.slice! amount, ary.size # THINK size is not perfectly correct, but anyway
+          ary.slice! amount, ary.size # Note: The ary.size is not perfectly correct.
         end
 
       end

server/lib/picky/interfaces/live_parameters/master_child.rb

@@ -130,7 +130,8 @@ def extract_configuration
           }
         end
 
-        # TODO Move to Interface object.
+        # THINK What to do about this? Standardize the tokenizer interface,
+        # then access each individual tokenizer.
         #
         def querying_removes_characters
           regexp = Tokenizer.searching.instance_variable_get :@removes_characters_regexp

server/spec/lib/bundle_indexing_spec.rb

@@ -25,7 +25,7 @@
       bundle.similar(:dragon).should == [:dargon]
     end
     it 'returns the right similars' do
-      bundle.similar(:trkn).should == [:dragon, :dargon]
+      bundle.similar(:trkn).should == [:dargon, :dragon]
     end
     it 'performs' do
       performance_of { bundle.similar(:dragon) }.should < 0.000075

server/spec/lib/generators/similarity/phonetic_spec.rb

@@ -21,53 +21,61 @@
     instance.send :initialize
   end
 
-  describe 'prioritize!' do
+  describe 'prioritize' do
     let(:phonetic) { described_class.allocate }
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 2
 
       ary = [:a, :b, :c]
-      phonetic.prioritize! ary, :b
+      phonetic.prioritize ary, :b
 
       ary.should == [:b, :a]
     end
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 2
 
       ary = [:aaa, :aa, :aaaa]
-      phonetic.prioritize! ary, :aaa
+      phonetic.prioritize ary, :aaa
 
       ary.should == [:aaa, :aa]
     end
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 3
 
       ary = [:aaa, :aa, :aaaa]
-      phonetic.prioritize! ary, :aaa
+      phonetic.prioritize ary, :aaa
 
       ary.should == [:aaa, :aa, :aaaa]
     end
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 3
 
       ary = [:aaaaa, :aa, :aaaa]
-      phonetic.prioritize! ary, :aaa
+      phonetic.prioritize ary, :aaa
 
       ary.should == [:aaaa, :aa, :aaaaa]
     end
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 3
 
       ary = [:aaaaa, :aa]
-      phonetic.prioritize! ary, :aaa
+      phonetic.prioritize ary, :aaa
 
       ary.should == [:aa, :aaaaa]
     end
     it 'sorts correctly' do
       phonetic.instance_variable_set :@amount, 3
 
       ary = [:aaa]
-      phonetic.prioritize! ary, :aaa
+      phonetic.prioritize ary, :aaa
+
+      ary.should == [:aaa]
+    end
+    it 'sorts correctly' do
+      phonetic.instance_variable_set :@amount, 1
+
+      ary = [:a, :aa, :aaa]
+      phonetic.prioritize ary, :aaa
 
       ary.should == [:aaa]
     end