charty 0.2.7 → 0.2.11

data/lib/charty/table_adapters.rb

@@ -23,3 +23,4 @@ require_relative 'table_adapters/daru_adapter'
 require_relative 'table_adapters/active_record_adapter'
 require_relative 'table_adapters/nmatrix_adapter'
 require_relative 'table_adapters/pandas_adapter'
+require_relative 'table_adapters/arrow_adapter'

data/lib/charty/vector.rb

@@ -49,7 +49,33 @@ module Charty
 
     alias completecases notnull
 
-    def_delegators :adapter, :mean, :stdev
+    def_delegators :adapter, :mean, :stdev, :percentile
+
+    def_delegators :adapter, :scale, :scale_inverse
+
+    def scale(method)
+      case method
+      when :linear
+        self
+      when :log
+        adapter.log_scale(method)
+      else
+        raise ArgumentError,
+              "Invalid scaling method: %p" % method
+      end
+    end
+
+    def scale_inverse(method)
+      case method
+      when :linear
+        self
+      when :log
+        adapter.inverse_log_scale(method)
+      else
+        raise ArgumentError,
+              "Invalid scaling method: %p" % method
+      end
+    end
 
     # TODO: write test
     def categorical_order(order=nil)
@@ -59,7 +85,9 @@ module Charty
           order = categories
         else
           order = unique_values.compact
-          order.sort! if numeric?
+          if numeric?
+            order.sort_by! {|x| Util.missing?(x) ? Float::INFINITY : x }
+          end
         end
         order.compact!
       end

data/lib/charty/vector_adapters/array_adapter.rb

@@ -30,7 +30,7 @@ module Charty
       include NameSupport
       include IndexSupport
 
-      def_delegators :data, :values_at
+      def_delegators :data, :values_at, :to_a
 
       def where(mask)
         masked_data, masked_index = where_in_array(mask)

data/lib/charty/vector_adapters/arrow_adapter.rb

@@ -0,0 +1,156 @@
+module Charty
+  module VectorAdapters
+    class ArrowAdapter < BaseAdapter
+      VectorAdapters.register(:arrow, self)
+
+      include Enumerable
+      include NameSupport
+      include IndexSupport
+
+      def self.supported?(data)
+        (defined?(Arrow::Array) && data.is_a?(Arrow::Array)) ||
+          (defined?(Arrow::ChunkedArray) && data.is_a?(Arrow::ChunkedArray))
+      end
+
+      def initialize(data)
+        @data = check_data(data)
+        self.index = index || RangeIndex.new(0 ... length)
+      end
+
+      def size
+        @data.length
+      end
+
+      def empty?
+        @data.length.zero?
+      end
+
+      def where(mask)
+        mask = check_mask_vector(mask)
+        mask_data = mask.data
+        unless mask_data.is_a?(Arrow::BooleanArray)
+          mask_data = mask.to_a
+          mask_data = mask_data.map(&:nonzero?) if mask_data[0].is_a?(Integer)
+          mask_data = Arrow::BooleanArray.new(mask_data)
+        end
+        masked_data = @data.filter(mask_data)
+        masked_index = []
+        mask_data.to_a.each_with_index do |boolean, i|
+          masked_index << index[i] if boolean
+        end
+        Vector.new(masked_data, index: masked_index, name: name)
+      end
+
+      def boolean?
+        case @data
+        when Arrow::BooleanArray
+          true
+        when Arrow::ChunkedArray
+          @data.value_data_type.is_a?(Arrow::BooleanDataType)
+        else
+          false
+        end
+      end
+
+      def numeric?
+        case @data
+        when Arrow::NumericArray
+          true
+        when Arrow::ChunkedArray
+          @data.value_data_type.is_a?(Arrow::NumericDataType)
+        else
+          false
+        end
+      end
+
+      def categorical?
+        case @data
+        when Arrow::StringArray, Arrow::DictionaryArray
+          true
+        when Arrow::ChunkedArray
+          case @data.value_data_type
+          when Arrow::StringArray, Arrow::DictionaryDataType
+            true
+          else
+            false
+          end
+        else
+          false
+        end
+      end
+
+      def categories
+        if @data.respond_to?(:dictionary)
+          dictionary = @data.dictionary
+        else
+          dictionary = @data.dictionary_encode.dictionary
+        end
+        dictionary.to_a
+      end
+
+      def unique_values
+        @data.unique.to_a
+      end
+
+      def group_by(grouper)
+        grouper = Vector.new(grouper) unless grouper.is_a?(Vector)
+        group_keys = grouper.unique_values
+        grouper_data = grouper.data
+        unless grouper_data.is_a?(Arrow::Array)
+          grouper_data = Arrow::Array.new(grouper.to_a)
+        end
+        equal = Arrow::Function.find("equal")
+        group_keys.map { |key|
+          if key.nil?
+            target_vector = Vector.new([nil] * @data.n_nulls)
+          else
+            mask = equal.execute([grouper_data, key]).value
+            target_vector = Vector.new(@data.filter(mask))
+          end
+          [key, target_vector]
+        }.to_h
+      end
+
+      def drop_na
+        if @data.n_nulls.zero?
+          Vector.new(@data, index: index, name: name)
+        else
+          data_without_null =
+            Arrow::Function.find("drop_null").execute([@data]).value
+          Vector.new(data_without_null)
+        end
+      end
+
+      def eq(val)
+        mask = Arrow::Function.find("equal").execute([@data, val]).value
+        Vector.new(mask, index: index, name: name)
+      end
+
+      def notnull
+        if @data.n_nulls.zero?
+          mask = Arrow::BooleanArray.new([true] * @data.length)
+        else
+          mask = Arrow::BooleanArray.new(@data.length,
+                                         @data.null_bitmap,
+                                         nil,
+                                         0)
+        end
+        Vector.new(mask, index: index, name: name)
+      end
+
+      def mean
+        @data.mean
+      end
+
+      def stdev(population: false)
+        options = Arrow::VarianceOptions.new
+        if population
+          options.ddof = 0
+        else
+          options.ddof = 1
+        end
+        Arrow::Function.find("stddev").execute([@data], options).value.value
+      end
+    end
+  end
+end

data/lib/charty/vector_adapters/daru_adapter.rb

@@ -34,12 +34,8 @@ module Charty
         case other
         when DaruVectorAdapter
           data == other.data
-        when ArrayAdapter
-          data.to_a == other.data
-        when NArrayAdapter, NMatrixAdapter, NumpyAdapter, PandasSeriesAdapter
-          other.compare_data_equality(self)
         else
-          data == other.data.to_a
+          to_a == other.to_a
         end
       end
 
@@ -156,7 +152,8 @@ module Charty
       end
 
       def percentile(q)
-        data.linear_percentile(q)
+        a = data.reject_values(*Daru::MISSING_VALUES).to_a
+        Statistics.percentile(a, q)
       end
     end
   end

data/lib/charty/vector_adapters/narray_adapter.rb

@@ -22,13 +22,22 @@ module Charty
         when NumpyAdapter, PandasSeriesAdapter
           other.compare_data_equality(self)
         else
-          data == other.data.to_a
+          data == other.to_a
         end
       end
 
       include NameSupport
       include IndexSupport
 
+      def to_a
+        case data
+        when Numo::Bit
+          map {|bit| bit == 1 }
+        else
+          super
+        end
+      end
+
       def where(mask)
         mask = check_mask_vector(mask)
         case mask.data

data/lib/charty/vector_adapters/nmatrix_adapter.rb

@@ -19,12 +19,8 @@ module Charty
         case other
         when NMatrixAdapter
           data == other.data
-        when ArrayAdapter, DaruVectorAdapter
-          data.to_a == other.data.to_a
-        when NArrayAdapter, NumpyAdapter, PandasSeriesAdapter
-          other.compare_data_equality(self)
         else
-          data == other.data.to_a
+          data.to_a == other.data.to_a
         end
       end
 

data/lib/charty/vector_adapters/numpy_adapter.rb

@@ -163,6 +163,10 @@ module Charty
       def stdev(population: false)
         Numpy.std(data, ddof: population ? 0 : 1)
       end
+
+      def percentile(q)
+        Numpy.nanpercentile(data, q)
+      end
     end
   end
 end

data/lib/charty/vector_adapters/pandas_adapter.rb

@@ -152,7 +152,8 @@ module Charty
           group_keys = grouper.unique.to_a
           groups = data.groupby(grouper)
           group_keys.map {|g|
-            [g, Charty::Vector.new(groups.get_group(g))]
+            g_vals = groups.get_group(g) rescue []
+            [g, Charty::Vector.new(g_vals)]
           }.to_h
         when Charty::Vector
           case grouper.adapter
@@ -194,6 +195,14 @@ module Charty
         q = q.map {|x| x / 100.0 }
         data.quantile(q)
       end
+
+      def log_scale(method)
+        Charty::Vector.new(Numpy.log10(data))
+      end
+
+      def inverse_log_scale(method)
+        Charty::Vector.new(Numpy.power(10, data))
+      end
     end
   end
 end

data/lib/charty/vector_adapters/vector_adapter.rb

@@ -0,0 +1,62 @@
+module Charty
+  module VectorAdapters
+    class VectorAdapter < BaseAdapter
+      VectorAdapters.register(:vector, self)
+
+      extend Forwardable
+      include Enumerable
+
+      def self.supported?(data)
+        data.is_a?(Vector)
+      end
+
+      def initialize(data, index: nil)
+        data = check_data(data)
+        @data = reduce_nested_vector(data)
+        self.index = index || RangeIndex.new(0 ... length)
+      end
+
+      include NameSupport
+      include IndexSupport
+
+      def_delegators :data,
+                     :boolean?,
+                     :categorical?,
+                     :categories,
+                     :drop_na,
+                     :each,
+                     :eq,
+                     :first_nonnil,
+                     :group_by,
+                     :notnull,
+                     :numeric?,
+                     :to_a,
+                     :uniq,
+                     :unique_values,
+                     :values_at,
+                     :where
+
+      def compare_data_equality(other)
+        if other.is_a?(self.class)
+          other = reduce_nested_vector(other.data).adapter
+        end
+        if other.is_a?(self.class)
+          @data.adapter.data == other.data
+        elsif @data.adapter.respond_to?(:compare_data_equality)
+          @data.adapter.compare_data_equality(other)
+        elsif other.respond_to?(:compare_data_equality)
+          other.compare_data_equality(@data.adapter)
+        else
+          @data.adapter.to_a == other.to_a
+        end
+      end
+
+      private def reduce_nested_vector(vector)
+        while vector.adapter.is_a?(self.class)
+          vector = vector.adapter.data
+        end
+        vector
+      end
+    end
+  end
+end

data/lib/charty/vector_adapters.rb

@@ -99,6 +99,26 @@ module Charty
       def stdev(population: false)
         Statistics.stdev(data, population: population)
       end
+
+      def percentile(q)
+        Statistics.percentile(data, q)
+      end
+
+      def log_scale(method)
+        Charty::Vector.new(
+          self.map {|x| Math.log10(x) },
+          index: index,
+          name: name
+        )
+      end
+
+      def inverse_log_scale(method)
+        Charty::Vector.new(
+          self.map {|x| 10.0 ** x },
+          index: index,
+          name: name
+        )
+      end
     end
 
     module NameSupport
@@ -180,8 +200,10 @@ module Charty
 end
 
 require_relative "vector_adapters/array_adapter"
+require_relative "vector_adapters/arrow_adapter"
 require_relative "vector_adapters/daru_adapter"
 require_relative "vector_adapters/narray_adapter"
 require_relative "vector_adapters/nmatrix_adapter"
 require_relative "vector_adapters/numpy_adapter"
 require_relative "vector_adapters/pandas_adapter"
+require_relative "vector_adapters/vector_adapter"

data/lib/charty/version.rb

@@ -1,5 +1,5 @@
 module Charty
-  VERSION = "0.2.7"
+  VERSION = "0.2.11"
 
   module Version
     numbers, TAG = VERSION.split("-")

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: charty
 version: !ruby/object:Gem::Version
-  version: 0.2.7
+  version: 0.2.11
 platform: ruby
 authors:
 - youchan
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2021-06-21 00:00:00.000000000 Z
+date: 2021-09-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: red-colors
@@ -208,6 +208,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 0.7.0
+- !ruby/object:Gem::Dependency
+  name: csv
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Visualizing your data in a simple way.
 email:
 - youchan01@gmail.com
@@ -232,6 +246,8 @@ files:
 - charty.gemspec
 - examples/Gemfile
 - examples/active_record.ipynb
+- examples/bar_plot.rb
+- examples/box_plot.rb
 - examples/daru.ipynb
 - examples/iris_dataset.ipynb
 - examples/nmatrix.ipynb
@@ -270,6 +286,7 @@ files:
 - examples/sample_images/subplot_pyplot.png
 - examples/sample_pyplot.ipynb
 - examples/sample_rubyplot.ipynb
+- examples/scatter_plot.rb
 - images/design_concept.png
 - images/penguins_body_mass_g_flipper_length_mm_scatter_plot.png
 - images/penguins_body_mass_g_flipper_length_mm_species_scatter_plot.png
@@ -319,6 +336,7 @@ files:
 - lib/charty/table.rb
 - lib/charty/table_adapters.rb
 - lib/charty/table_adapters/active_record_adapter.rb
+- lib/charty/table_adapters/arrow_adapter.rb
 - lib/charty/table_adapters/base_adapter.rb
 - lib/charty/table_adapters/daru_adapter.rb
 - lib/charty/table_adapters/datasets_adapter.rb
@@ -330,11 +348,13 @@ files:
 - lib/charty/vector.rb
 - lib/charty/vector_adapters.rb
 - lib/charty/vector_adapters/array_adapter.rb
+- lib/charty/vector_adapters/arrow_adapter.rb
 - lib/charty/vector_adapters/daru_adapter.rb
 - lib/charty/vector_adapters/narray_adapter.rb
 - lib/charty/vector_adapters/nmatrix_adapter.rb
 - lib/charty/vector_adapters/numpy_adapter.rb
 - lib/charty/vector_adapters/pandas_adapter.rb
+- lib/charty/vector_adapters/vector_adapter.rb
 - lib/charty/version.rb
 homepage: https://github.com/red-data-tools/charty
 licenses:
@@ -355,7 +375,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.3
+rubygems_version: 3.2.23
 signing_key: 
 specification_version: 4
 summary: Visualizing your data in a simple way.