rust 0.12 → 0.13

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8eb6e3759ef38070603a941ef348ac46e4b6c08b4638c493edb2169acf16c793
-  data.tar.gz: f855ca774695688ee64d7513ac94c8b98d3bae154bfae56b803ed1da567cb282
+  metadata.gz: 56e795fb0a8893df45abd976e2ed91344156f3c3dd4a68e17afd1a0fb317ece3
+  data.tar.gz: 406416738f1ab84fca06edd5cb59efdc623b12cefe03cd51b1a2cd840e218647
 SHA512:
-  metadata.gz: c281de698d8b4750832d77971dac857dbee2c31252b7950abf91777d37763bbb79577098ab48efeee6f8a5ef2a28949ef72816d33703eaad14887d588b4aac32
-  data.tar.gz: d44c0532c19ff8eb2f505d4da62e82b4b4f32dada4b39a05cddf57679a3a725dd38ceecf63aaf2d551a8ae691841196bad7cc0694d3613b928a84a9b6291ef6a
+  metadata.gz: 56854c3ff1bbd64ca8ff9d1201bc16fd37f4d3d465527217ab5c49d5cee0d6f4f34998bdf8a3ebdedf7ea8379909ca0db75d05da1bd46460c3e7066ee882ba7b
+  data.tar.gz: 6b21ba70c7d144384d1647dfa76f8894457bf4d65b74ce32d4e775cc3ecdc660f4596f61edd81e91b08c9d2b3def3491cebe10b7923b9f2988d072f7c5d25674

data/lib/rust/core/csv.rb

@@ -90,9 +90,9 @@ module Rust
             dataframe.column_names.each do |column_name|
                 values = dataframe.column(column_name)
                 
-                if values.all? { |s| !!Integer(s) rescue false }
+                if values.all? { |s| s == nil || !!Integer(s) rescue false }
                     integer_columns << column_name
-                elsif values.all? { |s| !!Float(s) rescue false }
+                elsif values.all? { |s| s == nil || !!Float(s) rescue false }
                     float_columns << column_name
                 end
             end
@@ -103,11 +103,11 @@ module Rust
             end
             
             integer_columns.each do |numeric_column|
-                dataframe.transform_column!(numeric_column) { |v| v.to_i }
+                dataframe.transform_column!(numeric_column) { |v| v != nil ? v.to_i : v }
             end
             
             float_columns.each do |numeric_column|
-                dataframe.transform_column!(numeric_column) { |v| v.to_f }
+                dataframe.transform_column!(numeric_column) { |v| v != nil ? v.to_f : v }
             end
             
             return dataframe

data/lib/rust/core/manual.rb

@@ -0,0 +1,89 @@
+require_relative 'rust'
+
+module Rust
+    class Manual
+        @@manuals = {}
+        
+        def self.about
+            puts "Manuals available:"
+            @@manuals.each do |category, manual|
+                puts "\t- #{manual.name} (:#{category}) → #{manual.description}"
+            end
+            
+            return nil
+        end
+        
+        def self.for(category)
+            category = category.to_sym
+            raise "No manual found for '#{category}'." unless @@manuals[category]
+            
+            return @@manuals[category]
+        end
+        
+        def self.register(category, name, description)
+            category = category.to_sym
+            
+            @@manuals[category] = Manual.new(name, description)
+            
+            return nil
+        end
+        
+        attr_reader     :name
+        attr_reader     :description
+        
+        def initialize(name, description)            
+            @name = name
+            @description = description
+            @voices = {}
+        end
+        
+        def lookup(query)
+            @voices.each do |key, value|
+                if query.match(key[1])
+                    puts "*** #{key[0]} ***"
+                    puts value
+                    return
+                end
+            end
+            
+            puts "Voice not found"
+            
+            return nil
+        end
+        
+        def n_voices
+            @voices.size
+        end
+        
+        def about
+            puts "****** Manual for #@name ******"
+            puts @description
+            puts "Voices in manual #@name:"
+            @voices.keys.each do |key, matcher|
+                puts "\t- #{key}"
+            end
+            
+            return nil
+        end
+        
+        def register(voice, matcher, description)
+            @voices[[voice, matcher]] = description
+        end
+        
+        def inspect
+            return "Manual for #@name with #{self.n_voices} voices"
+        end
+    end
+end
+
+module Rust::RBindings
+    def rust_help(category = nil, query = nil)
+        if !category
+            return Rust::Manual.about
+        elsif !query
+            return Rust::Manual.for(category).about
+        else
+            return Rust::Manual.for(category).lookup(query)
+        end
+    end
+end

data/lib/rust/core/rust.rb

@@ -102,7 +102,10 @@ module Rust
     end
     
     def self._rexec(r_command, return_warnings = false)
-        puts "Calling _rexec with command: #{r_command}" if @@debugging
+        if @@debugging
+            puts "Calling _rexec with command: #{r_command}"
+            puts "\t" + Kernel.caller.select { |v| !v.include?("irb") }.last(3).map { |v| v.sub(/^.*gems\//, "")}.join("\n\t")
+        end
         R_MUTEX.synchronize do
             assert("This command must be executed in an exclusive block") { @@in_client_mutex }
             
@@ -154,10 +157,17 @@ module Rust
     
     ##
     # Installs the given +name+ library and its dependencies.
+    # +github+ indicates whether the package is in GitHub.
     
-    def self.install_library(name)
+    def self.install_library(name, github = false)
+        self.prerequisite("remotes") if github
+
         self.exclusive do
-            self._eval("install.packages(\"#{name}\", dependencies = TRUE)")
+            if github
+                self._eval("remotes::install_github(\"#{name}\", dependencies=TRUE)")
+            else
+                self._eval("install.packages(\"#{name}\", dependencies = TRUE)")
+            end
         end
         
         return nil
@@ -165,9 +175,15 @@ module Rust
     
     ##
     # Installs the +library+ library if it is not available and loads it.
+    # +github+ indicates whether the package appears in GitHub.
     
-    def self.prerequisite(library)
-        self.install_library(library) unless self.check_library(library)
+    def self.prerequisite(library, github = false)
+        full_library = library
+        library = library.split("/").last if github
+        
+        unless self.check_library(library)
+            self.install_library(full_library, github)
+        end
         self.load_library(library)
     end
     
@@ -218,4 +234,11 @@ def bind_r!
     include Rust::RBindings
 end
 
+##
+# Shortcut for requiring rust external libraries
+
+def require_rust(name)
+    require "rust/external/#{name}"
+end
+
 bind_r! if ENV['RUBY_RUST_BINDING'] == '1'

data/lib/rust/core/types/dataframe.rb

@@ -1,4 +1,5 @@
 require_relative 'datatype'
+require 'tempfile'
 
 module Rust
     
@@ -355,8 +356,25 @@ module Rust
         end
         
         def load_in_r_as(variable_name)
-            command = []
+            tempfile = Tempfile.new('rust.dfport')
+            tempfile.close
+            
+            Rust::CSV.write(tempfile.path, self)
+            Rust._eval("#{variable_name} <- read.csv(\"#{tempfile.path}\", header=T)")
+            
+            if Rust.debug?
+                FileUtils.cp(tempfile.path, tempfile.path + ".debug.csv")
+                puts "Debug CSV port file available at: #{tempfile.path + ".debug.csv"}"
+            end
+            
+            tempfile.unlink
             
+            return true
+        end
+        
+        def directly_load_in_r_as(variable_name)
+            command = []
+                        
             command << "#{variable_name} <- data.frame()"
             row_index = 1
             self.each do |row|
@@ -374,6 +392,10 @@ module Rust
             end
             
             Rust._eval_big(command)
+            
+            tempfile.unlink
+            
+            return true
         end
         
         def inspect
@@ -408,16 +430,39 @@ module Rust
             return result
         end
         
+        ##
+        # Merges this data-frame with +other+ in terms of the +by+ column(s) (Array or String). Keeps all the rows in this data frame.
+        # +first_alias+ and +second_alias+ allow to specify the prefix that should be used for the columns not in +by+
+        # for this and the +other+ data-frame, respectively.
+        
+        def left_merge(other, by, first_alias, second_alias, **options)
+            options[:keep_right] = true
+            options[:keep_left]  = false
+            return other.merge(self, by, first_alias, second_alias, **options)
+        end
+        
+        ##
+        # Merges this data-frame with +other+ in terms of the +by+ column(s) (Array or String). Keeps all the rows in the other data frame.
+        # +first_alias+ and +second_alias+ allow to specify the prefix that should be used for the columns not in +by+
+        # for this and the +other+ data-frame, respectively.
+        
+        def right_merge(other, by, first_alias, second_alias, **options)
+            options[:keep_right] = true
+            options[:keep_left]  = false
+            return self.merge(other, by, first_alias, second_alias, **options)
+        end
+        
         ##
         # Merges this data-frame with +other+ in terms of the +by+ column(s) (Array or String).
         # +first_alias+ and +second_alias+ allow to specify the prefix that should be used for the columns not in +by+
         # for this and the +other+ data-frame, respectively.
         
-        def merge(other, by, first_alias = "x", second_alias = "y")
+        def merge(other, by, first_alias = "x", second_alias = "y", **options)
             raise TypeError, "Expected Rust::DataFrame" unless other.is_a?(DataFrame)
             raise TypeError, "Expected list of strings" if !by.is_a?(Array) || !by.all? { |e| e.is_a?(String) }
             raise "This dataset should have all the columns in #{by}" unless (by & self.column_names).size == by.size
             raise "The passed dataset should have all the columns in #{by}" unless (by & other.column_names).size == by.size
+            raise "Either keep_right or keep_left should be provided as options, not both" if options[:keep_right] && options[:keep_left]
             
             if first_alias == second_alias
                 if first_alias == ""
@@ -473,6 +518,28 @@ module Rust
                     end
                     
                     result << to_add
+                    
+                elsif options[:keep_right]
+                    to_add = {}
+                    
+                    by.each do |colname|
+                        to_add[colname] = other_row[colname]
+                    end
+                    
+                    merged_column_self.each do |colname|
+                        to_add["#{first_alias}#{colname}"] = nil
+                    end
+                    
+                    merged_column_other.each do |colname|
+                        to_add["#{second_alias}#{colname}"] = other_row[colname]
+                    end
+                    
+                    result << to_add
+                    
+                elsif options[:keep_left]
+                    options[:keep_left]  = false
+                    options[:keep_right] = true
+                    return other.merge(self, by, first_alias, second_alias, **options)
                 end
             end
             

data/lib/rust/core/types/datatype.rb

@@ -36,7 +36,7 @@ module Rust
             if candidates.size > 0
                 type = candidates.max_by { |c| c.pull_priority }
                 
-                puts "Using #{type} to pull #{variable}" if Rust.debug?
+                puts "Using #{type} to pull #{variable} (candidates: #{candidates.map { |c| c.to_s + "=>" + c.pull_priority.to_s}.join(", ")})" if Rust.debug?
                 return type.pull_variable(variable, r_type, r_class)
             else
                 if Rust._pull("length(#{variable})") == 0
@@ -80,7 +80,7 @@ module Rust
         def r_mirror
             varname = self.mirrored_R_variable_name
                         
-            if !Rust._pull("exists(\"#{varname}\")") || Rust._pull("#{varname}.hash") != self.r_hash
+            if !Rust._pull("exists(\"#{varname}\")") || Rust["#{varname}.hash"] != self.r_hash
                 puts "Loading #{varname}" if Rust.debug?
                 Rust[varname] = self
                 Rust["#{varname}.hash"] = self.r_hash

data/lib/rust/core/types/factor.rb

@@ -125,6 +125,10 @@ module Rust
             @level
         end
         
+        def to_str
+           @level.to_s
+        end
+        
         def to_R
             self.to_i
         end

data/lib/rust/core/types/language.rb

@@ -171,6 +171,22 @@ module Rust
         end
     end
     
+    ##
+    # Represents a verbatim R expression
+    
+    class Verbatim
+        ## 
+        #Creates a verbatim R expression
+        
+        def initialize(expression)
+            @expression = expression
+        end
+        
+        def to_R
+            @expression
+        end
+    end
+    
     ##
     # Represents the arguments of a function in R. Works as an Array of objects.
     
@@ -196,4 +212,20 @@ module Rust
             return options
         end
     end
+    
+    def self.verbatim(expression)
+        Verbatim.new(expression)
+    end
+    
+    def self.variable(variable)
+        Variable.new(variable)
+    end
+    
+    def self.function(name)
+        Function.new(name)
+    end
+    
+    def self.formula(left_part, right_part)
+        Formula.new(left_part, right_part)
+    end
 end 

data/lib/rust/core/types/list.rb

@@ -29,6 +29,8 @@ module Rust
             @data.each do |key, value|
                 Rust["#{variable_name}[[#{key + 1}]]"] = value
             end
+            
+            Rust._eval("names(#{variable_name}) <- #{self.names.to_R}")
         end
         
         ##

data/lib/rust/core/types/matrix.rb

@@ -81,6 +81,14 @@ module Rust
             @data.size
         end
         
+        def rownames
+            @row_names
+        end
+        
+        def colnames
+            @column_names
+        end
+        
         ##
         # Returns the number of columns.
         

data/lib/rust/core.rb

@@ -1,7 +1,57 @@
 require_relative 'core/rust'
 require_relative 'core/csv'
+require_relative 'core/manual'
 
 self_path = File.expand_path(__FILE__)
 Dir.glob(File.join(File.dirname(self_path), "core/types/*.rb")).each do |lib|
     require_relative lib
 end
+
+Rust::Manual.register(:base, "Quick intro", "Core philosophy behind Rust.")
+Rust::Manual.for(:base).register('Introduction', /intro/,
+<<-EOS
+Rust is a statistical library. Rust wraps R and its libraries to achieve this goal.
+Rust aims at:
+- Making easier for Ruby developers make all the kinds of operations that are straightforward in R;
+- Providing an object-oriented interface, more familiar than the one in R.
+
+Rust can be used in two ways:
+- By using the object-oriented interface (advised if you are writing a script);
+- By using the R bindings, that allow to use Ruby pretty much like R (handful if you are using it from IRB).
+
+Rust provides wrappers for many elements, including types (e.g., data frames), statistical hypothesis tests, plots, and so on.
+Under the hood, Rust creates an R environment (through rinruby), through which Rust can perform the most advanced operations,
+for which a re-implementation would be impractical.
+EOS
+)
+
+Rust::Manual.for(:base).register('Types', /type/,
+<<-EOS
+Rust provides wrappers for the most commonly-found types in R. Specifically, the following types are available:
+- Data frames → Rust::DataFrame
+- Factors → Rust::Factor
+- Matrices → Rust::Matrix
+- Lists → Rust::List
+- S4 classes → Rust::S4Class
+- Formulas → Rust::Formula
+
+Note that some of them (e.g., data frames and matrices) are not just wrappers, but complete re-implementations of the R
+types (for performance reasons).
+EOS
+)
+
+Rust::Manual.for(:base).register('CSVs', /csv/,
+<<-EOS
+Rust allows to read and write CSV files, mostly like in R.
+To read a CSV file, you can use:
+Rust::CSV.read(filename)
+
+It returns a data frame. You can also specify the option "headers" to tell if the first row in the CSV contains the headers 
+(column names for the data frame). Other options get directly passed to the R function "read.csv".
+
+To write a CSV file, you can use:
+Rust::CSV.write(filename, data_frame)
+
+It writes the given data frame on the file at filename.
+EOS
+)

data/lib/rust/external/ggplot2/plot_builder.rb

@@ -64,14 +64,6 @@ module Rust::Plots::GGPlot
             return self
         end
         
-        def labeled(value)
-            raise "No context for assigning a label" unless @current_context
-            @label_options[@current_context] = value
-            @current_context = nil
-            
-            return self
-        end
-        
         def with_x_label(value)
             @label_options[:x] = value
             
@@ -90,6 +82,76 @@ module Rust::Plots::GGPlot
             return self
         end
         
+        def scale_x_continuous(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:x, :continuous, **options)
+            
+            return self
+        end
+        
+        def scale_y_continuous(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:y, :continuous, **options)
+            
+            return self
+        end
+        
+        def scale_x_discrete(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:x, :discrete, **options)
+            
+            return self
+        end
+        
+        def scale_y_discrete(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:y, :discrete, **options)
+            
+            return self
+        end
+        
+        def scale_x_log10(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:x, :log10, **options)
+            
+            return self
+        end
+        
+        def scale_y_log10(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:y, :log10, **options)
+            
+            return self
+        end
+        
+        def scale_x_reverse(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:x, :reverse, **options)
+            
+            return self
+        end
+        
+        def scale_y_reverse(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:y, :reverse, **options)
+            
+            return self
+        end
+        
+        def scale_x_sqrt(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:x, :sqrt, **options)
+            
+            return self
+        end
+        
+        def scale_y_sqrt(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(:y, :sqrt, **options)
+            
+            return self
+        end
+        
         def with_title(value)
             @label_options[:title] = value
             
@@ -160,6 +222,48 @@ module Rust::Plots::GGPlot
             return self
         end
         
+        def labeled(value)
+            raise "No context for assigning a label" unless @current_context
+            @label_options[@current_context] = value
+            
+            return self
+        end
+        
+        def scale_continuous(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(@current_context, :continuous, **options)
+            
+            return self
+        end
+        
+        def scale_discrete(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(@current_context, :discrete, **options)
+            
+            return self
+        end
+        
+        def scale_log10(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(@current_context, :log10, **options)
+            
+            return self
+        end
+        
+        def scale_reverse(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(@current_context, :reverse, **options)
+            
+            return self
+        end
+        
+        def scale_sqrt(**options)
+            raise "No context for assigning a label" unless @current_context
+            @layers << AxisScaler.new(@current_context, :sqrt, **options)
+            
+            return self
+        end
+        
         def flip_coordinates
             @layers << FlipCoordinates.new
             

data/lib/rust/external/ggplot2/scale.rb

@@ -0,0 +1,12 @@
+require_relative 'core'
+
+module Rust::Plots::GGPlot
+    class AxisScaler < Layer
+        def initialize(axis, type = :continuous, **options)
+            @axis = axis
+            @type = type
+            
+            super("scale_#{@axis}_#{@type}", **options)
+        end
+    end
+end

data/lib/rust/external/ggplot2/themes.rb

@@ -53,7 +53,7 @@ module Rust::Plots::GGPlot
         def to_h
             options = @options.clone
             
-            options['_starting'] = @starting.sub("theme_", "")
+            options['_starting'] = @starting.sub("theme_", "") if @starting
             options = options.map do |key, value|
                 [key, value.is_a?(Theme::Element) ? value.to_h : value]
             end.to_h
@@ -78,6 +78,9 @@ module Rust::Plots::GGPlot
         end
     end
     
+    class ExistingTheme < Layer
+    end
+    
     class Theme::Element
         attr_reader :options
         
@@ -152,6 +155,8 @@ module Rust::Plots::GGPlot
                 return value
             elsif value.is_a?(Hash)
                 return Theme::LineElement.new(**value)
+            elsif !value
+                return Theme::BlankElement.new
             else
                 raise "Expected line or hash"
             end
@@ -162,6 +167,8 @@ module Rust::Plots::GGPlot
                 return value
             elsif value.is_a?(Hash)
                 return Theme::RectElement.new(**value)
+            elsif !value
+                return Theme::BlankElement.new
             else
                 raise "Expected rect or hash"
             end
@@ -172,6 +179,8 @@ module Rust::Plots::GGPlot
                 return value
             elsif value.is_a?(Hash)
                 return Theme::TextElement.new(**value)
+            elsif !value
+                return Theme::BlankElement.new
             else
                 raise "Expected text or hash"
             end
@@ -225,7 +234,7 @@ module Rust::Plots::GGPlot
     end
     
     class ThemeBuilder < ThemeComponentBuilder
-        def initialize(starting = 'bw')
+        def initialize(starting = nil)
             super("plot")
             @starting = starting
         end
@@ -417,7 +426,21 @@ module Rust::Plots::GGPlot
         end
     end
     
-    self.default_theme = ThemeBuilder.new.
+    class ThemeCollection
+        def self.ggtech(name = "google")
+            Rust.prerequisite("ricardo-bion/ggtech", true)
+            
+            return ExistingTheme.new("theme_tech", theme: name)
+        end
+        
+        def self.ggdark(style = "classic")
+            Rust.prerequisite("ggdark")
+            
+            return ExistingTheme.new("dark_theme_#{style}")
+        end
+    end
+    
+    self.default_theme = ThemeBuilder.new("bw").
             title(face: 'bold', size: 12).
             legend do |legend|
                 legend.background(fill: 'white', size: 4, colour: 'white')

data/lib/rust/external/ggplot2.rb

@@ -2,4 +2,115 @@ require_relative 'ggplot2/core'
 require_relative 'ggplot2/geoms'
 require_relative 'ggplot2/themes'
 require_relative 'ggplot2/plot_builder'
-require_relative 'ggplot2/helper'
+require_relative 'ggplot2/scale'
+
+Rust::Manual.register(:ggplot2, "ggplot2", "Informations on the wrapper of the popular ggplot2 plotting library for R.")
+
+Rust::Manual.for(:ggplot2).register("Introduction", /intro/,
+<<-EOS
+bind_ggplot! # Avoid using long module names to reach Rust::Plots::GGPlot (simply includes this module)
+
+# Best with a dataframe, but not necessary. If you have it...
+df = Rust.toothgrowth
+plot = PlotBuilder.for_dataframe(df). # Use a dataframe (symbols will be variable names)
+        labeled("Example plot"). # "labeled" sets the label to the last set aesthetic item (x, y, or title, in this case)
+        with_x(:len).labeled("X data from df").  # Set all the aesthetics (x, y, ...)
+        with_y(:dose).labeled("Y data from df").
+        draw_points. # Set the geometries to plot (based on the plot type)
+        build # Returns the plot ready to use
+plot.show # Show the plot in a window
+plot.save("output.pdf", width: 5, height: 4) # Save the plot, width, height etc. are optional
+
+# If you don't have a dataframe...
+plot2 = PlotBuilder.new.
+        with_x([1,2,3]).labeled("X data from df").
+        with_y([3,4,5]).labeled("Y data from df").
+        draw_points.
+        build
+plot2.show
+EOS
+)
+
+Rust::Manual.for(:ggplot2).register("Scatter plots", /scatter/,
+<<-EOS
+bind_ggplot!
+df = Rust.toothgrowth
+plot = PlotBuilder.for_dataframe(df).
+        with_x(:len).labeled("X data").
+        with_y(:dose).labeled("Y data").
+        draw_points. # To draw points
+        draw_lines.  # To draw lines (keep both to draw both)
+        build
+plot.show
+EOS
+)
+
+Rust::Manual.for(:ggplot2).register("Bar plots", /bar/,
+<<-EOS
+bind_ggplot!
+df = Rust.toothgrowth
+plot = PlotBuilder.for_dataframe(df).
+        with_x(:len).labeled("X data").
+        with_fill(:supp).labeled("Legend"). # Use with_fill or with_color for stacked plots
+        draw_bars. # To draw bars
+        build
+plot.show
+EOS
+)
+
+Rust::Manual.for(:ggplot2).register("Box plots", /box/,
+<<-EOS
+bind_ggplot!
+df = Rust.toothgrowth
+plot = PlotBuilder.for_dataframe(df).
+        with_y(:len).labeled("Data to boxplot").
+        with_group(:supp).labeled("Groups"). # Groups to plot
+        draw_boxplot.
+        build
+plot.show
+EOS
+)
+
+Rust::Manual.for(:ggplot2).register("Histograms", /hist/,
+<<-EOS
+bind_ggplot!
+df = Rust.toothgrowth
+plot = PlotBuilder.for_dataframe(df).
+        with_x(:len).labeled("Data to plot").
+        with_fill(:supp).labeled("Color"). # Use with_fill or with_color for multiple plots
+        draw_histogram.
+        build
+plot.show
+EOS
+)
+
+Rust::Manual.for(:ggplot2).register("Themes", /them/,
+<<-EOS
+bind_ggplot!
+df = Rust.toothgrowth
+# The method with_theme allows to change theme options. The method can be called
+# several times, each time the argument does not overwrite the previous options,
+# unless they are specified again (in that case, the last specified ones win).
+plot = PlotBuilder.for_dataframe(df).
+        with_x(:len).labeled("X data").
+        with_y(:dose).labeled("Y data").
+        draw_points.
+        with_theme(
+            ThemeBuilder.new('bw').
+                title(face: 'bold', size: 12). # Each method sets the property for the related element
+                legend do |legend| # Legend and other parts can be set like this
+                    legend.position(:left) # Puts the legend on the left
+                end.
+                axis do |axis| # Modifies the axes
+                    axis.line(Theme::BlankElement.new) # Hides the lines for the axes
+                    axis.text_x(size: 3) # X axis labels
+                end.
+                panel do |panel|
+                    panel.grid_major(colour: 'grey70', size: 0.2) # Sets the major ticks grid
+                    panel.grid_minor(Theme::BlankElement.new) # Hides the minor ticks grid
+                end.
+                build
+        ).build
+plot.show
+EOS
+)

data/lib/rust/models/regression.rb

@@ -14,6 +14,11 @@ module Rust::Models::Regression
     # Generic regression model in R.
     
     class RegressionModel < Rust::RustDatatype
+        
+        attr_accessor :data
+        attr_accessor :dependent_variable
+        attr_accessor :options
+        
         def self.can_pull?(type, klass)
             # Can only pull specific sub-types
             return false
@@ -38,22 +43,30 @@ module Rust::Models::Regression
             
             formula = Rust::Formula.new(dependent_variable, independent_variables.join(" + "))
             
+            result = nil
             Rust.exclusive do
                 Rust["#{model_type}.data"] = data
                 
                 Rust._eval("#{model_type}.model.result <- #{model_type}(#{formula.to_R}, data=#{model_type}.data#{mapped})")
                 result = Rust["#{model_type}.model.result"]
-                result.r_mirror_to("#{model_type}.model.result")
                 
-                return result
+                raise "An error occurred while building the model" unless result
+                
+                result.r_mirror_to("#{model_type}.model.result")
             end
+            
+            result.dependent_variable = dependent_variable
+            result.data = data
+            result.options = options
+            
+            return result
         end
         
         ##
-        # Creates a new +model+.
+        # Creates a new model based on +model+.
             
         def initialize(model)
-            raise StandardError if model.is_a?(RegressionModel)
+            raise "Expected a R list, given a #{model.class}" if !model.is_a?(Rust::List)
             @model = model
         end
         
@@ -118,6 +131,58 @@ module Rust::Models::Regression
             a = self.summary|"coefficients"
         end
         
+        ##
+        # Returns object variables for the model with basic data (coefficients and p-values). Use the method `coefficients`
+        # to get more data.
+        
+        def variables
+            unless @variables
+                coefficients = self.coefficients
+                
+                @variables = coefficients.rownames.map do |name|
+                    ModelVariable.new(name, coefficients[name, "Estimate"], coefficients[name, "Pr(>|t|)"])
+                end
+            end
+            
+            return @variables
+        end
+        
+        ##
+        # Returns only the significant variables as ModelVariable instances. See the method `variables`.
+        
+        def significant_variables(a = 0.05)
+            self.variables.select { |v| v.significant?(a) }
+        end
+        
+        ##
+        # Runs backward selection (recursively removes a variable until the best model is found).
+        # Returns both the best model and the list of excluded variable at each step
+        # Note: Not fully tested
+        
+        def backward_selection(excluded = [])
+            candidates = self.variables.select { |v| !v.intercept? && !v.significant? }.sort_by { |v| v.pvalue }.reverse
+            all = self.variables.select { |v| !v.intercept? }
+            
+            candidates.each do |candidate|
+                new_model = RegressionModel.generate(
+                    self.class,
+                    self.class.r_model_name,
+                    self.dependent_variable,
+                    (all - [candidate]).map { |v| v.name },
+                    self.data,
+                    **self.options
+                )
+                
+                if new_model.r_2_adjusted >= self.r_2_adjusted
+                    puts "Excluded #{candidate}" if Rust.debug?
+                    return *new_model.backward_selection(excluded + [candidate])
+                end
+            end
+            
+            return self, excluded
+        end
+        
+        
         def method_missing(name, *args)
             return model|name.to_s
         end
@@ -145,7 +210,11 @@ module Rust::Models::Regression
     
     class LinearRegressionModel < RegressionModel
         def self.can_pull?(type, klass)
-            return type == "list" && klass == "lm"
+            return type == "list" && klass == self.r_model_name
+        end
+        
+        def self.pull_priority
+            1
         end
         
         def self.pull_variable(variable, type, klass)
@@ -154,6 +223,10 @@ module Rust::Models::Regression
             return LinearRegressionModel.new(model)
         end
         
+        def self.r_model_name
+            "lm"
+        end
+        
         ##
         # Generates a linear regression model, given its +dependent_variable+ and +independent_variables+ and its +data+. 
         # +options+ can be specified and directly passed to the model. 
@@ -161,7 +234,7 @@ module Rust::Models::Regression
         def self.generate(dependent_variable, independent_variables, data, **options)
             RegressionModel.generate(
                 LinearRegressionModel,
-                "lm", 
+                self.r_model_name, 
                 dependent_variable, 
                 independent_variables, 
                 data, 
@@ -175,13 +248,17 @@ module Rust::Models::Regression
     
     class LinearMixedEffectsModel < RegressionModel
         def self.can_pull?(type, klass)
-            return type == "S4" && klass == "lmerModLmerTest"
+            return type == "S4" && klass == self.r_model_name
         end
         
         def self.pull_priority
             1
         end
         
+        def self.r_model_name
+            "lmerModLmerTest"
+        end
+        
         def self.pull_variable(variable, type, klass)
             model = Rust::RustDatatype.pull_variable(variable, Rust::S4Class)
             
@@ -213,7 +290,7 @@ module Rust::Models::Regression
             
             RegressionModel.generate(
                 LinearMixedEffectsModel,
-                "lmer", 
+                self.r_model_name, 
                 dependent_variable, 
                 fixed_effects + random_effects, 
                 data, 
@@ -235,18 +312,44 @@ module Rust::Models::Regression
             end
         end
     end
+    
+    ##
+    # Slim representation for a variable in a model, with just the variable name, its coefficient and its p-value.
+    
+    class ModelVariable
+        attr_accessor   :name
+        attr_accessor   :coefficient
+        attr_accessor   :pvalue
+        
+        def initialize(name, coefficient, pvalue)
+            @name = name
+            @coefficient = coefficient
+            @pvalue = pvalue
+        end
+        
+        def intercept?
+            @name == "(Intercept)"
+        end
+        
+        ##
+        # Checks whether the variable is significant w.r.t. a given +a+ (0.05 by default)
+        
+        def significant?(a = 0.05)
+            @pvalue <= a
+        end
+    end
 end
 
 module Rust::RBindings
     def lm(formula, data, **options)
         independent = formula.right_part.split("+").map { |v| v.strip }
-        return LinearRegressionModel.generate(formula.left_part, independent, data, **options)
+        return Rust::Models::Regression::LinearRegressionModel.generate(formula.left_part, independent, data, **options)
     end
     
     def lmer(formula, data, **options)
         independent = formula.right_part.split("+").map { |v| v.strip }
         
-        RegressionModel.generate(
+        Rust::Models::Regression::RegressionModel.generate(
             LinearMixedEffectsModel,
             "lmer", 
             formula.left_part, 

data/lib/rust/stats/probabilities.rb

@@ -85,6 +85,20 @@ module Rust
             @values.map { |k, v| k*v }.sum
         end
         
+        ##
+        # Returns the variance for this slice.
+        
+        def variance
+            @values.map { |k, v| k**2 * v }.sum - (self.expected ** 2)
+        end
+        
+        ##
+        # Returns the standard deviation for this slice.
+        
+        def sd
+            Math.sqrt(self.variance)
+        end
+        
         ##
         # Returns a slice with the values that are greater than +n+.
         
@@ -124,7 +138,7 @@ module Rust
         # Returns a slice with the values between +a+ and +b+.
         
         def between(a, b)
-            self.so_that { |k| k.between(a, b) }
+            self.so_that { |k| k.between?(a, b) }
         end
         
         ##
@@ -133,6 +147,13 @@ module Rust
         def so_that
             RandomVariableSlice.new(@values.select { |k, v| yield(k) })
         end
+        
+        ##
+        # Creates a bar plot of the distribution
+        
+        def plot
+            Rust::Plots::BarPlot.new(@values.sort_by { |k, v| k }.to_h)
+        end
     end
     
     ##

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rust
 version: !ruby/object:Gem::Version
-  version: '0.12'
+  version: '0.13'
 platform: ruby
 authors:
 - Simone Scalabrino
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-08-17 00:00:00.000000000 Z
+date: 2024-11-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rinruby
@@ -61,6 +61,7 @@ files:
 - lib/rust.rb
 - lib/rust/core.rb
 - lib/rust/core/csv.rb
+- lib/rust/core/manual.rb
 - lib/rust/core/rust.rb
 - lib/rust/core/types/all.rb
 - lib/rust/core/types/dataframe.rb
@@ -74,8 +75,8 @@ files:
 - lib/rust/external/ggplot2.rb
 - lib/rust/external/ggplot2/core.rb
 - lib/rust/external/ggplot2/geoms.rb
-- lib/rust/external/ggplot2/helper.rb
 - lib/rust/external/ggplot2/plot_builder.rb
+- lib/rust/external/ggplot2/scale.rb
 - lib/rust/external/ggplot2/themes.rb
 - lib/rust/external/robustbase.rb
 - lib/rust/models/all.rb
@@ -110,7 +111,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.15
+rubygems_version: 3.5.16
 signing_key:
 specification_version: 4
 summary: Ruby advanced statistical library

data/lib/rust/external/ggplot2/helper.rb

@@ -1,122 +0,0 @@
-require_relative 'core'
-
-GGPLOT_EXAMPLES = {}
-
-GGPLOT_EXAMPLES[["Quick introduction", /intro/]] = <<-EOS
-bind_ggplot! # Avoid using long module names to reach Rust::Plots::GGPlot (simply includes this module)
-
-# Best with a dataframe, but not necessary. If you have it...
-df = Rust.toothgrowth
-plot = PlotBuilder.for_dataframe(df). # Use a dataframe (symbols will be variable names)
-        labeled("Example plot"). # "labeled" sets the label to the last set aesthetic item (x, y, or title, in this case)
-        with_x(:len).labeled("X data from df").  # Set all the aesthetics (x, y, ...)
-        with_y(:dose).labeled("Y data from df").
-        draw_points. # Set the geometries to plot (based on the plot type)
-        build # Returns the plot ready to use
-plot.show # Show the plot in a window
-plot.save("output.pdf", width: 5, height: 4) # Save the plot, width, height etc. are optional
-
-# If you don't have a dataframe...
-plot2 = PlotBuilder.new.
-        with_x([1,2,3]).labeled("X data from df").
-        with_y([3,4,5]).labeled("Y data from df").
-        draw_points.
-        build
-plot2.show
-EOS
-
-GGPLOT_EXAMPLES[["Scatter plots", /scatter/]] = <<-EOS
-bind_ggplot!
-df = Rust.toothgrowth
-plot = PlotBuilder.for_dataframe(df).
-        with_x(:len).labeled("X data").
-        with_y(:dose).labeled("Y data").
-        draw_points. # To draw points
-        draw_lines.  # To draw lines (keep both to draw both)
-        build
-plot.show
-EOS
-
-GGPLOT_EXAMPLES[["Bar plots", /bar/]] = <<-EOS
-bind_ggplot!
-df = Rust.toothgrowth
-plot = PlotBuilder.for_dataframe(df).
-        with_x(:len).labeled("X data").
-        with_fill(:supp).labeled("Legend"). # Use with_fill or with_color for stacked plots
-        draw_bars. # To draw bars
-        build
-plot.show
-EOS
-
-GGPLOT_EXAMPLES[["Box plots", /box/]] = <<-EOS
-bind_ggplot!
-df = Rust.toothgrowth
-plot = PlotBuilder.for_dataframe(df).
-        with_y(:len).labeled("Data to boxplot").
-        with_group(:supp).labeled("Groups"). # Groups to plot
-        draw_boxplot.
-        build
-plot.show
-EOS
-
-GGPLOT_EXAMPLES[["Histograms", /hist/]] = <<-EOS
-bind_ggplot!
-df = Rust.toothgrowth
-plot = PlotBuilder.for_dataframe(df).
-        with_x(:len).labeled("Data to plot").
-        with_fill(:supp).labeled("Color"). # Use with_fill or with_color for multiple plots
-        draw_histogram.
-        build
-plot.show
-EOS
-
-GGPLOT_EXAMPLES[["Themes", /them/]] = <<-EOS
-bind_ggplot!
-df = Rust.toothgrowth
-# The method with_theme allows to change theme options. The method can be called
-# several times, each time the argument does not overwrite the previous options,
-# unless they are specified again (in that case, the last specified ones win).
-plot = PlotBuilder.for_dataframe(df).
-        with_x(:len).labeled("X data").
-        with_y(:dose).labeled("Y data").
-        draw_points.
-        with_theme(
-            ThemeBuilder.new('bw').
-                title(face: 'bold', size: 12). # Each method sets the property for the related element
-                legend do |legend| # Legend and other parts can be set like this
-                    legend.position(:left) # Puts the legend on the left
-                end.
-                axis do |axis| # Modifies the axes
-                    axis.line(Theme::BlankElement.new) # Hides the lines for the axes
-                    axis.text_x(size: 3) # X axis labels
-                end.
-                panel do |panel|
-                    panel.grid_major(colour: 'grey70', size: 0.2) # Sets the major ticks grid
-                    panel.grid_minor(Theme::BlankElement.new) # Hides the minor ticks grid
-                end.
-                build
-        ).build
-plot.show
-EOS
-
-module Rust::Plots::GGPlot
-    def self.help!(topic = nil)
-        unless topic
-            puts "Topics:"
-            GGPLOT_EXAMPLES.keys.each do |key, matcher|
-                puts "- #{key}"
-            end
-            puts "Call again specifying the topic of interest."
-        else
-            GGPLOT_EXAMPLES.each do |key, value|
-                if topic.match(key[1])
-                    puts "*** #{key[0]} ***"
-                    puts value
-                    return
-                end
-            end
-            
-            puts "Topic not found"
-        end
-    end
-end