json-inference 0.0.1 → 0.0.2

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    json-inference (0.0.1)
+    json-inference (0.0.2)
 
 GEM
   remote: https://rubygems.org/

data/lib/json-inference.rb

@@ -3,147 +3,84 @@ module JsonInference
     Report.new
   end
 
-  class BaseNode
-    def initialize
-      @value_classes = Hash.new 0
-      @sub_nodes = Hash.new { |h,k| 
-        if k == :nth_child
-          sub_node = NthChildNode.new(self)
-        else
-          sub_node = Node.new(k, self)
-        end
-        h[k] = sub_node
-      }
-    end
-
-    def <<(value)
-      if value.is_a?(Hash)
-        value.each do |key, sub_value|
-          @sub_nodes[key] << sub_value
-        end
-      elsif value.is_a?(Array)
-        @sub_nodes[:nth_child]
-        value.each do |sub_value|
-          @sub_nodes[:nth_child] << sub_value
-        end
-      end
-      if value.class == String && value =~ /^(\d){4}-(\d){2}-(\d){2}T(\d){2}:(\d){2}:(\d){2}\.(\d){3}Z$/
-        @value_classes[Date] += 1
-      elsif [true, false].include?(value)
-        @value_classes['Boolean'] += 1
-      else
-        @value_classes[value.class] += 1
-      end
-    end
-
-    def each_sub_node
-      @sub_nodes.keys.sort.each do |key|
-        sub_node = @sub_nodes[key]
-        yield sub_node
-      end
-    end
-
-    def indent_level
-      @parent.indent_level + 1
-    end
-
-    def indent
-      '  ' * indent_level
-    end
-
-    def total_count
-      @value_classes.values.inject { |sum, i| sum + i } || 0
-    end
+  def self.percent_string(numerator, denominator)
+    "#{(numerator / denominator.to_f * 100).round}%"
   end
 
-  class Node < BaseNode
-    def initialize(name = nil, parent = nil)
-      super()
-      @name, @parent = name, parent
-    end
-
-    def selector
-      "#{@parent.selector} > .#{@name}"
+  class Report
+    def initialize
+      @documents = []
+      @root = RootNode.new
     end
 
-    def selector_line(documents_count)
-      "#{indent}#{selector}: #{total_count}/#{documents_count} (#{(total_count.to_f / documents_count * 100).round}%)\n"
+    def <<(document)
+      @documents << document
+      @root << document
     end
 
-    def to_s(documents_count)
-      str = ""
-      str << selector_line(documents_count)
-      @value_classes.each do |klass, count|
-        str << "  #{indent}#{klass}: #{(count / total_count.to_f * 100).round}%\n"
-      end
-      each_sub_node do |sub_node|
-        str << sub_node.to_s(documents_count)
-      end
+    def to_s
+      str = "JsonInference report: #{@documents.size} documents\n"
+      str << @root.to_s(@documents.size)
       str
     end
   end
 
-  class NthChildNode < BaseNode
-    def initialize(parent)
-      super()
-      @parent = parent
+  class NodeValuesCollection
+    def initialize
+      @value_counters = Hash.new { |h,k| h[k] = ValueCounter.new(k) }
     end
 
-    def selector
-      "#{@parent.selector}:nth-child()"
+    def <<(value)
+      if value.class == String && value =~ /^(\d){4}-(\d){2}-(\d){2}T(\d){2}:(\d){2}:(\d){2}\.(\d){3}Z$/
+        @value_counters[Date] << value
+      elsif [true, false].include?(value)
+        @value_counters['Boolean'] << value
+      else
+        @value_counters[value.class] << value
+      end
     end
 
-    def selector_line(documents_count)
-      "#{indent}#{selector}: #{total_count} child#{'ren' unless total_count == 1}\n"
+    def size
+      @value_counters.values.inject(0) { |sum, counter| sum + counter.size } || 0
     end
 
-    def to_s(documents_count)
+    def to_s(indent)
       str = ""
-      str << selector_line(documents_count)
-      @value_classes.each do |klass, count|
-        str << "  #{indent}#{klass}: #{(count / total_count.to_f * 100).round}%\n"
-      end
-      each_sub_node do |sub_node|
-        str << sub_node.to_s(total_count)
+      @value_counters.values.each do |value_counter|
+        str << "  #{indent}#{value_counter.to_s(size)}\n"
       end
       str
     end
-  end
 
-  class RootNode < BaseNode
-    def indent_level
-      -1
-    end
+    class ValueCounter
+      attr_reader :size
 
-    def selector
-      ':root'
-    end
-
-    def to_s(documents_count)
-      str = ""
-      each_sub_node do |sub_node|
-        str << sub_node.to_s(documents_count)
-        str << "\n"
+      def initialize(reported_class)
+        @reported_class = reported_class
+        @size = 0
+        @empties = 0
       end
-      str
-    end
-  end
 
-  class Report
-    def initialize
-      @documents = []
-      @root = RootNode.new
-    end
-
-    def <<(document)
-      @documents << document
-      @root << document
-    end
+      def <<(value)
+        @size += 1
+        if [Array, String].include?(@reported_class)
+          @empties += 1 if value.empty?
+        end
+      end
 
-    def to_s
-      str = "JsonInference report: #{@documents.size} documents\n"
-      str << @root.to_s(@documents.size)
-      str
+      def to_s(all_values_count)
+        str = "#{@reported_class}: #{JsonInference.percent_string(size, all_values_count)}"
+        if [Array, String].include?(@reported_class)
+          str << ", #{JsonInference.percent_string(@empties, size)} empty"
+        end
+        str
+      end
     end
   end
 end
+ 
+dir = File.dirname(__FILE__) + "/json-inference"
+require "#{dir}/base_node"
+require "#{dir}/node"
+require "#{dir}/nth_child_node"
+require "#{dir}/root_node"

data/lib/json-inference/base_node.rb

@@ -0,0 +1,44 @@
+module JsonInference
+  class BaseNode
+    def initialize
+      @values = NodeValuesCollection.new
+      @sub_nodes = Hash.new { |h,k| 
+        if k == :nth_child
+          sub_node = JsonInference::NthChildNode.new(self)
+        else
+          sub_node = JsonInference::Node.new(k, self)
+        end
+        h[k] = sub_node
+      }
+    end
+
+    def <<(value)
+      if value.is_a?(Hash)
+        value.each do |key, sub_value|
+          @sub_nodes[key] << sub_value
+        end
+      elsif value.is_a?(Array)
+        @sub_nodes[:nth_child]
+        value.each do |sub_value|
+          @sub_nodes[:nth_child] << sub_value
+        end
+      end
+      @values << value
+    end
+
+    def each_sub_node
+      @sub_nodes.keys.sort.each do |key|
+        sub_node = @sub_nodes[key]
+        yield sub_node
+      end
+    end
+
+    def indent_level
+      @parent.indent_level + 1
+    end
+
+    def indent
+      '  ' * indent_level
+    end
+  end
+end

data/lib/json-inference/node.rb

@@ -0,0 +1,26 @@
+module JsonInference
+  class Node < BaseNode
+    def initialize(name = nil, parent = nil)
+      super()
+      @name, @parent = name, parent
+    end
+
+    def selector
+      "#{@parent.selector} > .#{@name}"
+    end
+
+    def selector_line(documents_count)
+      "#{indent}#{selector}: #{@values.size}/#{documents_count} (#{JsonInference.percent_string(@values.size, documents_count)})\n"
+    end
+
+    def to_s(documents_count)
+      str = ""
+      str << selector_line(documents_count)
+      str << @values.to_s(indent)
+      each_sub_node do |sub_node|
+        str << sub_node.to_s(documents_count)
+      end
+      str
+    end
+  end
+end

data/lib/json-inference/nth_child_node.rb

@@ -0,0 +1,26 @@
+module JsonInference
+  class NthChildNode < BaseNode
+    def initialize(parent)
+      super()
+      @parent = parent
+    end
+
+    def selector
+      "#{@parent.selector}:nth-child()"
+    end
+
+    def selector_line(documents_count)
+      "#{indent}#{selector}: #{@values.size} child#{'ren' unless @values.size == 1}\n"
+    end
+
+    def to_s(documents_count)
+      str = ""
+      str << selector_line(documents_count)
+      str << @values.to_s(indent)
+      each_sub_node do |sub_node|
+        str << sub_node.to_s(@values.size)
+      end
+      str
+    end
+  end
+end

data/lib/json-inference/root_node.rb

@@ -0,0 +1,20 @@
+module JsonInference
+  class RootNode < BaseNode
+    def indent_level
+      -1
+    end
+
+    def selector
+      ':root'
+    end
+
+    def to_s(documents_count)
+      str = ""
+      each_sub_node do |sub_node|
+        str << sub_node.to_s(documents_count)
+        str << "\n"
+      end
+      str
+    end
+  end
+end

data/lib/json-inference/version.rb

@@ -1,5 +1,5 @@
 module Json
   module Inference
-    VERSION = "0.0.1"
+    VERSION = "0.0.2"
   end
 end

data/test/json_inference_test.rb

@@ -158,4 +158,18 @@ class JsonInferenceTestCase < Test::Unit::TestCase
       assert_match(/:root > .items:nth-child\(\): 0 children$/, @string)
     end
   end
+
+  context "field with empty strings" do
+    setup do
+      report = JsonInference.new_report
+      report << {foo: 'one'}
+      report << {foo: '', bar: 'ONE'}
+      report << {foo: '', baz: 'won'}
+      @string = report.to_s
+    end
+
+    should "note how likely it is to be empty" do
+      assert_match(/String: 100%, 67% empty/, @string)
+    end
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: json-inference
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-01-20 00:00:00.000000000 Z
+date: 2014-01-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -59,6 +59,10 @@ files:
 - Rakefile
 - json-inference.gemspec
 - lib/json-inference.rb
+- lib/json-inference/base_node.rb
+- lib/json-inference/node.rb
+- lib/json-inference/nth_child_node.rb
+- lib/json-inference/root_node.rb
 - lib/json-inference/version.rb
 - test/json_inference_test.rb
 homepage: ''
@@ -76,7 +80,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2986096997133027698
+      hash: -3344776264180278235
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -85,7 +89,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2986096997133027698
+      hash: -3344776264180278235
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.23