red_amber 0.1.7 → 0.1.8

data/lib/red_amber/data_frame.rb

@@ -13,10 +13,7 @@ module RedAmber
 
     def initialize(*args)
       @variables = @keys = @vectors = @types = @data_types = nil
-      # bug in gobject-introspection: ruby-gnome/ruby-gnome#1472
-      #  [Arrow::Table] == [nil] shows ArgumentError
-      #  temporary use yoda condition to workaround
-      if args.empty? || args == [[]] || args == [{}] || [nil] == args
+      if args.empty? || args[0] == [] || args[0] == {} || args[0].nil?
         # DataFrame.new, DataFrame.new([]), DataFrame.new({}), DataFrame.new(nil)
         #   returns empty DataFrame
         @table = Arrow::Table.new({}, [])
@@ -34,6 +31,7 @@ module RedAmber
             raise DataFrameTypeError, "invalid argument: #{arg}"
           end
       end
+      name_unnamed_keys
     end
 
     def self.load(path, options = {})
@@ -78,12 +76,12 @@ module RedAmber
     alias_method :var_names, :keys
 
     def key?(key)
-      @keys.include?(key.to_sym)
+      keys.include?(key.to_sym)
     end
     alias_method :has_key?, :key?
 
     def key_index(key)
-      @keys.find_index(key.to_sym)
+      keys.find_index(key.to_sym)
     end
     alias_method :find_index, :key_index
     alias_method :index, :key_index
@@ -144,8 +142,10 @@ module RedAmber
       end
     end
 
-    def group(*group_keys)
-      Group.new(self, group_keys)
+    def group(*group_keys, &block)
+      g = Group.new(self, group_keys)
+      g = g.summarize(&block) if block
+      g
     end
 
     private
@@ -182,5 +182,23 @@ module RedAmber
       html = IRuby::HTML.table(converted.to_h, maxrows: 8, maxcols: 15)
       "#{self.class} <#{size} x #{n_keys} vector#{pl(n_keys)}> #{html}"
     end
+
+    def name_unnamed_keys
+      return unless @table[:'']
+
+      # We can't use #keys because it causes mismatch of @table and @keys
+      keys = @table.schema.fields.map { |f| f.name.to_sym }
+      unnamed = (:unnamed1..).find { |e| !keys.include?(e) }
+      fields =
+        @table.schema.fields.map do |field|
+          if field.name.empty?
+            Arrow::Field.new(unnamed, field.data_type)
+          else
+            field
+          end
+        end
+      schema = Arrow::Schema.new(fields)
+      @table = Arrow::Table.new(schema, @table.columns)
+    end
   end
 end

data/lib/red_amber/data_frame_displayable.rb

@@ -5,6 +5,8 @@ require 'stringio'
 module RedAmber
   # mix-ins for the class DataFrame
   module DataFrameDisplayable
+    INDEX_KEY = :index_key_for_format_table
+
     def to_s
       return '' if empty?
 
@@ -139,7 +141,7 @@ module RedAmber
       original = self
       indices = size > head + tail ? [*0...head, *(size - tail)...size] : [*0...size]
       df = slice(indices).assign do
-        assigner = { '': indices.map { |i| (i + 1).to_s } }
+        assigner = { INDEX_KEY => indices.map { |i| (i + 1).to_s } }
         vectors.each_with_object(assigner) do |v, a|
           a[v.key] = v.to_a.map do |e|
             if e.nil?
@@ -155,12 +157,12 @@ module RedAmber
         end
       end
 
-      df = df.pick { [keys[-1], keys[0..-2]] }
+      df = df.pick { [INDEX_KEY, keys - [INDEX_KEY]] }
       df = size > head + tail ? df[0, 0, 0...head, 0, -tail..-1] : df[0, 0, 0..-1]
       df = df.assign do
         vectors.each_with_object({}) do |v, assigner|
-          vec = v.replace(0, v.key.to_s)
-                 .replace(1, v.key == :'' ? '' : "<#{original[v.key].type}>")
+          vec = v.replace(0, v.key == INDEX_KEY ? '' : v.key.to_s)
+                 .replace(1, v.key == INDEX_KEY ? '' : "<#{original[v.key].type}>")
           assigner[v.key] = size > head + tail ? vec.replace(head + 2, ':') : vec
         end
       end
@@ -197,7 +199,7 @@ module RedAmber
     end
 
     def format_for_column(vector, original, width)
-      if vector.key != :'' && !original[vector.key].numeric?
+      if vector.key != INDEX_KEY && !original[vector.key].numeric?
         "%-#{width}s"
       else
         "%#{width}s"

data/lib/red_amber/group.rb

@@ -16,36 +16,30 @@ module RedAmber
       @group = @table.group(*@group_keys)
     end
 
-    def count(*summary_keys)
-      by(:count, summary_keys)
+    functions = %i[count sum product mean min max stddev variance]
+    functions.each do |function|
+      define_method(function) do |*summary_keys|
+        by(function, summary_keys)
+      end
     end
 
-    def sum(*summary_keys)
-      by(:sum, summary_keys)
+    def inspect
+      tallys = @dataframe.pick(@group_keys).vectors.map.with_object({}) do |v, h|
+        h[v.key] = v.tally
+      end
+      "#<#{self.class}:#{format('0x%016x', object_id)}\n#{tallys}>"
     end
 
-    def product(*summary_keys)
-      by(:product, summary_keys)
-    end
-
-    def mean(*summary_keys)
-      by(:mean, summary_keys)
-    end
-
-    def min(*summary_keys)
-      by(:min, summary_keys)
-    end
-
-    def max(*summary_keys)
-      by(:max, summary_keys)
-    end
-
-    def stddev(*summary_keys)
-      by(:stddev, summary_keys)
-    end
-
-    def variance(*summary_keys)
-      by(:variance, summary_keys)
+    def summarize(&block)
+      agg = instance_eval(&block)
+      case agg
+      when DataFrame
+        agg
+      when Array
+        agg.reduce { |aggregated, df| aggregated.assign(df.to_h) }
+      else
+        raise GroupArgumentError, "Unknown argument: #{agg}"
+      end
     end
 
     private
@@ -55,7 +49,11 @@ module RedAmber
       d = summary_keys - @dataframe.keys
       raise GroupArgumentError, "#{d} is not a key of\n #{@dataframe}." unless summary_keys.empty? || d.empty?
 
-      RedAmber::DataFrame.new(@group.send(func, *summary_keys))
+      df = RedAmber::DataFrame.new(@group.send(func, *summary_keys))
+      df = df[df.keys[-1], df.keys[0...-1]]
+      # if counts are the same (no nil included), aggregate count columns.
+      df = df[df.keys[0..1]].rename(df.keys[1], :count) if func == :count && df.to_h.values[1..].uniq.size == 1
+      df
     end
   end
 end

data/lib/red_amber/vector_selectable.rb

@@ -64,6 +64,8 @@ module RedAmber
         return filter_by_array(arg)
       when Arrow::Array
         array = arg
+      when Range
+        array = normalize_element(arg)
       else
         unless arg.is_a?(Numeric) || booleans?([arg])
           raise VectorArgumentError, "Argument must be numeric or boolean: #{args}"

data/lib/red_amber/vector_updatable.rb

@@ -12,7 +12,15 @@ module RedAmber
     # @param replacer [Array, Vector, Arrow::Array] new data to replace for.
     # @return [Vector] Replaced new Vector
     def replace(args, replacer)
-      args = args.is_a?(Array) ? args : Array(args)
+      args =
+        case args
+        when Array
+          args
+        when Range
+          normalize_element(args)
+        else
+          Array(args)
+        end
       replacer = Array(replacer)
       return self if args.empty? || args[0].nil?
 
@@ -22,6 +30,7 @@ module RedAmber
         if vector.boolean?
           vector
         elsif vector.numeric?
+          replacer.sort_by! { |x| args[replacer.index(x)] } if replacer # rubocop:disable Style/SafeNavigation
           Vector.new(indices).is_in(vector)
         else
           raise VectorArgumentError, "Invalid data type #{args}"
@@ -50,6 +59,18 @@ module RedAmber
       is_nil.if_else(false, self).invert
     end
 
+    def shift(amount = 1, fill: nil)
+      raise VectorArgumentError, 'Shift amount is too large' if amount.abs > size
+
+      if amount.positive?
+        replace(amount..-1, self[0...-amount]).replace(0...amount, fill)
+      elsif amount.negative?
+        replace(0...amount, self[-amount..]).replace(amount..-1, fill)
+      else # amount == 0
+        self
+      end
+    end
+
     private
 
     # [Ternary]: replace_with(booleans, replacements) => vector

data/lib/red_amber/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RedAmber
-  VERSION = '0.1.7'
+  VERSION = '0.1.8'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: red_amber
 version: !ruby/object:Gem::Version
-  version: 0.1.7
+  version: 0.1.8
 platform: ruby
 authors:
 - Hirokazu SUZUKI (heronshoes)
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-07-15 00:00:00.000000000 Z
+date: 2022-08-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: red-arrow
@@ -34,6 +34,7 @@ extra_rdoc_files: []
 files:
 - ".rubocop.yml"
 - ".rubocop_todo.yml"
+- ".yardopts"
 - CHANGELOG.md
 - Gemfile
 - LICENSE
@@ -41,10 +42,10 @@ files:
 - Rakefile
 - benchmark/csv_load_penguins.yml
 - benchmark/drop_nil.yml
-- doc/47_examples_of_red_amber.ipynb
 - doc/CODE_OF_CONDUCT.md
 - doc/DataFrame.md
 - doc/Vector.md
+- doc/examples_of_red_amber.ipynb
 - doc/image/arrow_table_new.png
 - doc/image/dataframe/assign.png
 - doc/image/dataframe/drop.png