picky 2.7.0 → 3.0.0.pre1

data/lib/picky/backend/files.rb

@@ -1,22 +1,26 @@
-module Backend
+module Picky
 
-  class Files < Base
+  module Backend
 
-    def initialize bundle_name, category
-      super bundle_name, category
+    class Files < Base
 
-      # Note: We marshal the similarity, as the
-      #       Yajl json lib cannot load symbolized
-      #       values, just keys.
-      #
-      @inverted      = File::JSON.new    category.index_path(bundle_name, :inverted)
-      @weights       = File::JSON.new    category.index_path(bundle_name, :weights)
-      @similarity    = File::Marshal.new category.index_path(bundle_name, :similarity)
-      @configuration = File::JSON.new    category.index_path(bundle_name, :configuration)
-    end
+      def initialize bundle
+        super bundle
+
+        # Note: We marshal the similarity, as the
+        #       Yajl json lib cannot load symbolized
+        #       values, just keys.
+        #
+        @inverted      = File::JSON.new    bundle.index_path(:inverted)
+        @weights       = File::JSON.new    bundle.index_path(:weights)
+        @similarity    = File::Marshal.new bundle.index_path(:similarity)
+        @configuration = File::JSON.new    bundle.index_path(:configuration)
+      end
+
+      def to_s
+        "#{self.class}(#{[@inverted, @weights, @similarity, @configuration].join(', ')})"
+      end
 
-    def to_s
-      "#{self.class}(#{[@prepared, @inverted, @weights, @similarity, @configuration].join(', ')})"
     end
 
   end

data/lib/picky/backend/redis/basic.rb

@@ -1,89 +1,93 @@
-module Backend
+module Picky
 
-  class Redis
+  module Backend
 
-    # Redis Backend Accessor.
-    #
-    # Provides necessary helper methods for its
-    # subclasses.
-    # Not directly useable, as it does not provide
-    # dump/load methods.
-    #
-    class Basic
+    class Redis
 
-      attr_reader :namespace, :backend
-
-      # An index cache takes a path, without file extension,
-      # which will be provided by the subclasses.
+      # Redis Backend Accessor.
+      #
+      # Provides necessary helper methods for its
+      # subclasses.
+      # Not directly useable, as it does not provide
+      # dump/load methods.
       #
-      def initialize namespace
-        @namespace = namespace
+      class Basic
+
+        attr_reader :namespace, :backend
 
-        # TODO Turn this inside out such that people can pass in
-        # their own Redis instance.
+        # An index cache takes a path, without file extension,
+        # which will be provided by the subclasses.
         #
-        # TODO Make the :db a real option.
+        def initialize namespace
+          @namespace = namespace
+
+          # TODO Turn this inside out such that people can pass in
+          # their own Redis instance.
+          #
+          # TODO Make the :db a real option.
+          #
+          @backend = ::Redis.new :db => 15
+        end
+
+        # Does nothing.
         #
-        @backend = ::Redis.new :db => 15
-      end
-
-      # Does nothing.
-      #
-      def load
-        # Nothing.
-      end
-      # We do not use Redis to retrieve data.
-      #
-      def retrieve
-        # Nothing.
-      end
+        def load
+          # Nothing.
+        end
+        # We do not use Redis to retrieve data.
+        #
+        def retrieve
+          # Nothing.
+        end
 
-      # Redis does not backup.
-      #
-      def backup
-        # Nothing.
-      end
+        # Redis does not backup.
+        #
+        def backup
+          # Nothing.
+        end
 
-      # Deletes the Redis index namespace.
-      #
-      def delete
-        # Not implemented here.
-        # Note: backend.flushdb might be the way to go,
-        #       but since we cannot delete by key pattern,
-        #       we don't do anything.
-      end
+        # Deletes the Redis index namespace.
+        #
+        def delete
+          # Not implemented here.
+          # Note: backend.flushdb might be the way to go,
+          #       but since we cannot delete by key pattern,
+          #       we don't do anything.
+        end
+
+        # Checks.
+        #
 
-      # Checks.
-      #
+        # Is this cache suspiciously small?
+        #
+        def cache_small?
+          size < 1
+        end
 
-      # Is this cache suspiciously small?
-      #
-      def cache_small?
-        size < 1
-      end
+        # Is the cache ok?
+        #
+        # A small cache is still ok.
+        #
+        def cache_ok?
+          size > 0
+        end
 
-      # Is the cache ok?
-      #
-      # A small cache is still ok.
-      #
-      def cache_ok?
-        size > 0
-      end
+        # Extracts the size of the file in Bytes.
+        #
+        # Note: This is a very forgiving implementation.
+        #       But as long as Redis does not implement
+        #       DBSIZE KEYPATTERN, we are stuck with this.
+        #
+        def size
+          backend.dbsize
+        end
 
-      # Extracts the size of the file in Bytes.
-      #
-      # Note: This is a very forgiving implementation.
-      #       But as long as Redis does not implement
-      #       DBSIZE KEYPATTERN, we are stuck with this.
-      #
-      def size
-        backend.dbsize
-      end
+        #
+        #
+        def to_s
+          "#{self.class}(#{namespace}:*)"
+        end
 
-      #
-      #
-      def to_s
-        "#{self.class}(#{namespace}:*)"
       end
 
     end

data/lib/picky/backend/redis/list_hash.rb

@@ -1,45 +1,49 @@
-module Backend
-
-  class Redis
-
-    class ListHash < Basic
-
-      # Writes the hash into Redis.
-      #
-      def dump hash
-        clear
-        hash.each_pair do |key, values|
-          redis_key = "#{namespace}:#{key}"
-          i = 0
-          values.each do |value|
-            i += 1
-            backend.zadd redis_key, i, value
+module Picky
+
+  module Backend
+
+    class Redis
+
+      class ListHash < Basic
+
+        # Writes the hash into Redis.
+        #
+        def dump hash
+          clear
+          hash.each_pair do |key, values|
+            redis_key = "#{namespace}:#{key}"
+            i = 0
+            values.each do |value|
+              i += 1
+              backend.zadd redis_key, i, value
+            end
           end
         end
-      end
 
-      # Clear the index for this list.
-      #
-      # Note: Perhaps we can use a server only command.
-      #       This is not the optimal way to do it.
-      #
-      def clear
-        redis_key = "#{namespace}:*"
-        backend.keys(redis_key).each do |key|
-          backend.del key
+        # Clear the index for this list.
+        #
+        # Note: Perhaps we can use a server only command.
+        #       This is not the optimal way to do it.
+        #
+        def clear
+          redis_key = "#{namespace}:*"
+          backend.keys(redis_key).each do |key|
+            backend.del key
+          end
         end
-      end
 
-      # Get a collection.
-      #
-      def collection key
-        backend.zrange "#{namespace}:#{key}", 0, -1
-      end
+        # Get a collection.
+        #
+        def collection key
+          backend.zrange "#{namespace}:#{key}", 0, -1
+        end
+
+        # Get a single value.
+        #
+        def member key
+          raise "Can't retrieve single value :#{key} from a Redis ListHash. Use Indexes::Redis::StringHash."
+        end
 
-      # Get a single value.
-      #
-      def member key
-        raise "Can't retrieve single value :#{key} from a Redis ListHash. Use Index::Redis::StringHash."
       end
 
     end

data/lib/picky/backend/redis/string_hash.rb

@@ -1,36 +1,40 @@
-module Backend
+module Picky
 
-  class Redis
+  module Backend
 
-    class StringHash < Basic
+    class Redis
 
-      # Writes the hash into Redis.
-      #
-      # Note: We could use multi, but it did not help.
-      #
-      def dump hash
-        clear
-        hash.each_pair do |key, value|
-          backend.hset namespace, key, value
+      class StringHash < Basic
+
+        # Writes the hash into Redis.
+        #
+        # Note: We could use multi, but it did not help.
+        #
+        def dump hash
+          clear
+          hash.each_pair do |key, value|
+            backend.hset namespace, key, value
+          end
         end
-      end
 
-      # Clears the hash.
-      #
-      def clear
-        backend.del namespace
-      end
+        # Clears the hash.
+        #
+        def clear
+          backend.del namespace
+        end
 
-      # Get a collection.
-      #
-      def collection key
-        raise "Can't retrieve collection for :#{key} from a StringHash. Use Index::Redis::ListHash."
-      end
+        # Get a collection.
+        #
+        def collection key
+          raise "Can't retrieve collection for :#{key} from a StringHash. Use Indexes::Redis::ListHash."
+        end
+
+        # Get a single value.
+        #
+        def member key
+          backend.hget namespace, key
+        end
 
-      # Get a single value.
-      #
-      def member key
-        backend.hget namespace, key
       end
 
     end

data/lib/picky/backend/redis.rb

@@ -1,38 +1,42 @@
-module Backend
+module Picky
 
-  #
-  #
-  class Redis < Base
+  module Backend
 
-    def initialize bundle_name, category
-      super bundle_name, category
+    #
+    #
+    class Redis < Base
+
+      def initialize bundle
+        super bundle
+
+        # Refine a few Redis "types".
+        #
+        @inverted      = Redis::ListHash.new   "#{bundle.identifier}:inverted"
+        @weights       = Redis::StringHash.new "#{bundle.identifier}:weights"
+        @similarity    = Redis::ListHash.new   "#{bundle.identifier}:similarity"
+        @configuration = Redis::StringHash.new "#{bundle.identifier}:configuration"
+      end
 
-      # Refine a few Redis "types".
+      # Delegate to the right collection.
       #
-      @inverted      = Redis::ListHash.new   "#{category.identifier}:#{bundle_name}:inverted"
-      @weights       = Redis::StringHash.new "#{category.identifier}:#{bundle_name}:weights"
-      @similarity    = Redis::ListHash.new   "#{category.identifier}:#{bundle_name}:similarity"
-      @configuration = Redis::StringHash.new "#{category.identifier}:#{bundle_name}:configuration"
-    end
+      def ids sym
+        inverted.collection sym
+      end
 
-    # Delegate to the right collection.
-    #
-    def ids sym
-      inverted.collection sym
-    end
+      # Delegate to the right member value.
+      #
+      # Note: Converts to float.
+      #
+      def weight sym
+        weights.member(sym).to_f
+      end
 
-    # Delegate to the right member value.
-    #
-    # Note: Converts to float.
-    #
-    def weight sym
-      weights.member(sym).to_f
-    end
+      # Delegate to a member value.
+      #
+      def setting sym
+        configuration.member sym
+      end
 
-    # Delegate to a member value.
-    #
-    def setting sym
-      configuration.member sym
     end
 
   end

data/lib/picky/bundle.rb

@@ -1,62 +1,87 @@
-# A Bundle is a number of indexes
-# per [index, category] combination.
-#
-# At most, there are three indexes:
-# * *core* index (always used)
-# * *weights* index (always used)
-# * *similarity* index (used with similarity)
-#
-# In Picky, indexing is separated from the index
-# handling itself through a parallel structure.
-#
-# Both use methods provided by this base class, but
-# have very different goals:
-#
-# * *Indexing*::*Bundle*::*Base* is just concerned with creating index
-#   files / redis entries and providing helper functions to e.g. check
-#   the indexes.
-#
-# * *Index*::*Bundle*::*Base* is concerned with loading these index files into
-#   memory / redis and looking up search data as fast as possible.
-#
-class Bundle
-
-  attr_reader   :identifier,
-                :files
-  attr_accessor :inverted,
-                :weights,
-                :similarity,
-                :configuration,
-                :similarity_strategy
-
-  delegate :clear,    :to => :inverted
-  delegate :[], :[]=, :to => :configuration
-
-  def initialize name, category, similarity_strategy
-    @identifier    = "#{category.identifier}:#{name}"
-    @files         = Backend::Files.new name, category
-
-    @inverted      = {}
-    @weights       = {}
-    @similarity    = {}
-    @configuration = {} # A hash with config options.
-
-    @similarity_strategy = similarity_strategy
-  end
-
-  # Get a list of similar texts.
+module Picky
+  # A Bundle is a number of indexes
+  # per [index, category] combination.
   #
-  # Note: Does not return itself.
+  # At most, there are three indexes:
+  # * *core* index (always used)
+  # * *weights* index (always used)
+  # * *similarity* index (used with similarity)
   #
-  def similar text
-    code = similarity_strategy.encoded text
-    similar_codes = code && @similarity[code]
-    similar_codes.delete text if similar_codes
-    similar_codes || []
-  end
+  # In Picky, indexing is separated from the index
+  # handling itself through a parallel structure.
+  #
+  # Both use methods provided by this base class, but
+  # have very different goals:
+  #
+  # * *Indexing*::*Bundle*::*Base* is just concerned with creating index
+  #   files / redis entries and providing helper functions to e.g. check
+  #   the indexes.
+  #
+  # * *Index*::*Bundle*::*Base* is concerned with loading these index files into
+  #   memory / redis and looking up search data as fast as possible.
+  #
+  class Bundle
 
-  def to_s
-    "#{self.class}(#{identifier}, #{files})"
-  end
+    attr_reader :name,
+                :category
+
+    attr_accessor :inverted,
+                  :weights,
+                  :similarity,
+                  :configuration,
+                  :similarity_strategy
 
+    delegate :clear,           :to => :inverted
+    delegate :[], :[]=,        :to => :configuration
+    delegate :index_directory, :to => :category
+
+    def initialize name, category, similarity_strategy, options = {}
+      @name          = name
+      @category      = category
+
+      @inverted      = {}
+      @weights       = {}
+      @similarity    = {}
+      @configuration = {} # A hash with config options.
+
+      @similarity_strategy = similarity_strategy
+    end
+    def identifier
+      "#{category.identifier}:#{name}"
+    end
+
+    # Get a list of similar texts.
+    #
+    # Note: Does not return itself.
+    #
+    def similar text
+      code = similarity_strategy.encoded text
+      similar_codes = code && @similarity[code]
+      similar_codes.delete text if similar_codes
+      similar_codes || []
+    end
+
+    # If a key format is set, use it, else delegate to the category.
+    #
+    def key_format
+      @key_format || @category.key_format
+    end
+
+    # Path and partial filename of a specific subindex.
+    #
+    # Subindexes are:
+    #  * inverted index
+    #  * weights index
+    #  * partial index
+    #  * similarity index
+    #
+    def index_path type
+      ::File.join index_directory, "#{category.name}_#{name}_#{type}"
+    end
+
+    def to_s
+      "#{self.class}(#{identifier})"
+    end
+
+  end
 end

data/lib/picky/calculations/location.rb

@@ -1,55 +1,59 @@
-module Calculations # :nodoc:all
-
-  # A location calculation recalculates a 1-d location
-  # to the Picky internal 1-d "grid".
-  #
-  # For example, if you have a location x == 12.3456,
-  # it will be recalculated into 3, if the minimum is 9
-  # and the gridlength is 1.
-  #
-  class Location
-
-    attr_reader :minimum, :precision, :grid
-
-    def initialize user_grid, precision = nil
-      @user_grid = user_grid
-      @precision = precision || 1
-      @grid      = @user_grid / (@precision + 0.5)
-    end
-
-    def minimum= minimum
-      # Add a margin of 1 user grid.
-      #
-      minimum -= @user_grid
-
-      # Add plus 1 grid so that the index key never falls on 0.
-      # Why? to_i maps by default to 0.
-      #
-      minimum -= @grid
+module Picky
 
-      @minimum = minimum
-    end
+  module Calculations # :nodoc:all
 
+    # A location calculation recalculates a 1-d location
+    # to the Picky internal 1-d "grid".
     #
+    # For example, if you have a location x == 12.3456,
+    # it will be recalculated into 3, if the minimum is 9
+    # and the gridlength is 1.
     #
-    def add_margin length
-      @minimum -= length
-    end
+    class Location
+
+      attr_reader :minimum, :precision, :grid
+
+      def initialize user_grid, precision = nil
+        @user_grid = user_grid
+        @precision = precision || 1
+        @grid      = @user_grid / (@precision + 0.5)
+      end
+
+      def minimum= minimum
+        # Add a margin of 1 user grid.
+        #
+        minimum -= @user_grid
+
+        # Add plus 1 grid so that the index key never falls on 0.
+        # Why? to_i maps by default to 0.
+        #
+        minimum -= @grid
+
+        @minimum = minimum
+      end
+
+      #
+      #
+      def add_margin length
+        @minimum -= length
+      end
+
+      #
+      #
+      def recalculated_range location
+        range recalculate(location)
+      end
+      #
+      #
+      def range around_location
+        (around_location - @precision)..(around_location + @precision)
+      end
+      #
+      #
+      def recalculate location
+        ((location - @minimum) / @grid).floor
+      end
 
-    #
-    #
-    def recalculated_range location
-      range recalculate(location)
-    end
-    #
-    #
-    def range around_location
-      (around_location - @precision)..(around_location + @precision)
-    end
-    #
-    #
-    def recalculate location
-      ((location - @minimum) / @grid).floor
     end
 
   end