statsample 0.6.1 → 0.6.2

data/lib/statsample/factor/pca.rb

@@ -17,8 +17,7 @@ module Factor
     
     
     def initialize(matrix ,opts=Hash.new)
-      if matrix.is_a? ::Matrix
-        require 'matrix_extension'
+      if matrix.respond_to? :to_gsl
         matrix=matrix.to_gsl
       end
       @name=""

data/lib/statsample/factor/principalaxis.rb

@@ -40,8 +40,8 @@ module Factor
       @iterations=0
       t.times do |i|
         @iterations+=1
-        prev_com.each_with_index{|v,i|
-          work_matrix[i][i]=v
+        prev_com.each_with_index{|v,it|
+          work_matrix[it][it]=v
         }
         pca=Statsample::PCA.new(::Matrix.rows(work_matrix))
         

data/lib/statsample/graph/svghistogram.rb

@@ -1,208 +1,206 @@
 module Statsample
-    module Graph
-class SvgHistogram < SVG::Graph::BarBase
-    attr_accessor :inner_margin, :mean, :sigma, :show_normal
-        def initialize(config)
-            config[:fields]=[:dummy]
-            super(config)
-            
-            @histogram=nil
-        end
-		include REXML
-		
-			# In addition to the defaults set in Graph::initialize, sets
-			# [inner_margin] 14
-			# [key] false
-		def set_defaults 
-		 super
-		 self.top_align = self.top_font = 0
-         init_with({
-            :inner_margin=>16,
-            :key=>false,
-            :show_normal=>false
-            })
-         
-		end
-		
-        def histogram=(h)
-            @histogram=h
-            @data=[{:data=>(0...@histogram.bins).to_a.collect {|i|
-                @histogram[i]
-            }}]
+  module Graph
+    class SvgHistogram < SVG::Graph::BarBase
+      attr_accessor :inner_margin, :mean, :sigma, :show_normal
+      def initialize(config)
+        config[:fields]=[:dummy]
+        super(config)
+        
+        @histogram=nil
+      end
+      include REXML
+      
+      # In addition to the defaults set in Graph::initialize, sets
+      # [inner_margin] 14
+      # [key] false
+      def set_defaults 
+      super
+      self.top_align = self.top_font = 0
+      init_with({
+        :inner_margin=>16,
+        :key=>false,
+        :show_normal=>false
+        })
+      
+      end
+      
+      def histogram=(h)
+        @histogram=h
+        @data=[{:data=>(0...@histogram.bins).to_a.collect {|i|
+            @histogram[i]
+        }}]
+      end
+      def get_x_labels
+        [""]
+      end
+      
+      def get_y_labels
+        maxvalue = max_value
+        minvalue = min_value
+        range = maxvalue - minvalue
+        
+        top_pad = range == 0 ? 10 : range / 20.0
+        scale_range = (maxvalue + top_pad) - minvalue
+        
+        scale_division = scale_divisions || (scale_range / 10.0)
+        
+        if scale_integers
+        scale_division = scale_division < 1 ? 1 : scale_division.round
         end
-		def get_x_labels
-            [""]
-		end
         
-		def get_y_labels
-            maxvalue = max_value
-            minvalue = min_value
-            range = maxvalue - minvalue
-            
-            top_pad = range == 0 ? 10 : range / 20.0
-            scale_range = (maxvalue + top_pad) - minvalue
-            
-            scale_division = scale_divisions || (scale_range / 10.0)
-            
-            if scale_integers
-            scale_division = scale_division < 1 ? 1 : scale_division.round
+        rv = []
+        maxvalue = maxvalue%scale_division == 0 ? 
+        maxvalue : maxvalue + scale_division
+        minvalue.step( maxvalue, scale_division ) {|v| rv << v}
+        rv
+      end
+      
+      def unit_width
+        (@graph_width-(@inner_margin*2)) / (@histogram.max-@histogram.min).to_f
+      end
+      def draw_x_label(v)
+        left = (v - @histogram.min)*unit_width                    
+        x=@inner_margin+left
+        text = @graph.add_element( "text" )
+        text.attributes["class"] = "xAxisLabels"
+        text.text = sprintf("%0.2f",v)
+        y = @graph_height + x_label_font_size + 3
+        text.attributes["x"] = x.to_s
+        text.attributes["y"] = y.to_s
+      end
+      
+      def draw_x_labels
+        if show_x_labels
+          (0...@histogram.bins).each do |i|
+            value = @histogram[i]
+            range = @histogram.get_range(i)
+            draw_x_label(range[0])
+            if(i==(@histogram.bins)-1)
+              draw_x_label(range[1])
             end
-            
-            rv = []
-            maxvalue = maxvalue%scale_division == 0 ? 
-            maxvalue : maxvalue + scale_division
-            minvalue.step( maxvalue, scale_division ) {|v| rv << v}
-            return rv
-		end
-		
-        def unit_width
-            (@graph_width-(@inner_margin*2)) / (@histogram.max-@histogram.min).to_f
+          end
         end
-        def draw_x_label(v)
-            left = (v - @histogram.min)*unit_width                    
-            x=@inner_margin+left
-            text = @graph.add_element( "text" )
-            text.attributes["class"] = "xAxisLabels"
-            text.text = sprintf("%0.2f",v)
-            y = @graph_height + x_label_font_size + 3
-            text.attributes["x"] = x.to_s
-            text.attributes["y"] = y.to_s
-            
+      end
+      def draw_data
+        minvalue = min_value
+        fieldwidth = field_width
+        unit_size =  (@graph_height.to_f - font_size*2*top_font) / 
+                    (get_y_labels.max - get_y_labels.min)		
+        bottom = @graph_height
+        field_count = 0
+        hist_min=@histogram.min
+        hist_max=@histogram.max
+        range_hist=hist_max-hist_min
+        total=0
+        
+        (0...@histogram.bins).each do |i|
+          dataset_count = 0
+          value = @histogram[i]
+          total=total+value
+          range = @histogram.get_range(i)
+          left = (range[0] - hist_min)*unit_width
+          bar_width = (range[1] - hist_min)*unit_width - left
+          length = (value.abs - (minvalue > 0 ? minvalue : 0)) * unit_size
+          # top is 0 if value is negative
+          top = bottom - (((value < 0 ? 0 : value) - minvalue) * unit_size)
+
+          @graph.add_element( "rect", {
+            "x" => (@inner_margin+left).to_s,
+            "y" => top.to_s,
+            "width" => bar_width.to_s,
+            "height" => length.to_s,
+            "class" => "fill#{dataset_count+1}"
+          })
+
+          make_datapoint_text(left + @inner_margin+ (bar_width/2), top - 6, value.to_s)
+          field_count += 1
         end
-	    def draw_x_labels
-            if show_x_labels
-                (0...@histogram.bins).each { |i|
-                    value = @histogram[i]
-                    range = @histogram.get_range(i)
-                    draw_x_label(range[0])
-                    if(i==(@histogram.bins)-1)
-                        draw_x_label(range[1])
-                    end
-                }
+        if(show_normal)
+          divs=30
+          path=""
+          0.upto(divs) do |i|
+            x_abs=hist_min+(range_hist/divs)*i
+            y=GSL::Ran::gaussian_pdf((x_abs-mean) / sigma)*total
+            xg=@inner_margin+((x_abs-hist_min)*unit_width)
+            yg=bottom-(y-minvalue)*unit_size
+            if i==0
+              path="M#{xg} #{yg} "
+            else
+              path+="L#{xg} #{yg} "
             end
+          end
+          @graph.add_element("path", 
+            { "d"=>path, "style"=>"stroke:black;fill:none" })
         end
-		def draw_data
-            minvalue = min_value
-            fieldwidth = field_width
-    
-            unit_size =  (@graph_height.to_f - font_size*2*top_font) / 
-                          (get_y_labels.max - get_y_labels.min)		
-            bottom = @graph_height
-            field_count = 0
-            hist_min=@histogram.min
-            hist_max=@histogram.max
-            range_hist=hist_max-hist_min
-            total=0
-            (0...@histogram.bins).each { |i|
-                dataset_count = 0
-                value = @histogram[i]
-                total=total+value
-                range = @histogram.get_range(i)
-                left = (range[0] - hist_min)*unit_width
-                bar_width = (range[1] - hist_min)*unit_width - left
-                length = (value.abs - (minvalue > 0 ? minvalue : 0)) * unit_size
-                # top is 0 if value is negative
-                top = bottom - (((value < 0 ? 0 : value) - minvalue) * unit_size)
-                @graph.add_element( "rect", {
-                    "x" => (@inner_margin+left).to_s,
-                    "y" => top.to_s,
-                    "width" => bar_width.to_s,
-                    "height" => length.to_s,
-                    "class" => "fill#{dataset_count+1}"
-                })
-                make_datapoint_text(left + @inner_margin+ (bar_width/2), top - 6, value.to_s)
-                field_count += 1
-            }
-            if(show_normal)
-                divs=30
-                path=""
-                0.upto(divs) {|i|
-                    x_abs=hist_min+(range_hist/divs)*i
-                    y=GSL::Ran::gaussian_pdf((x_abs-mean) / sigma)*total
-                    xg=@inner_margin+((x_abs-hist_min)*unit_width)
-                    yg=bottom-(y-minvalue)*unit_size
-                    if i==0
-                        path="M#{xg} #{yg} "
-                    else
-                        path+="L#{xg} #{yg} "
-                    end
-                }
-                @graph.add_element("path", {
-                        "d"=>path,
-                        "style"=>"stroke:black;fill:none"
-                }
-                )
-            end
-	end
-    
-    
-     def get_css
-        return <<EOL
+      end
+      
+      
+      def get_css
+      return <<EOL
 /* default fill styles for multiple datasets (probably only use a single dataset on this graph though) */
 
 .key1,.fill1{
-	fill: #ff0000;
-	stroke: black;
-	stroke-width: 1px;	
+fill: #ff0000;
+stroke: black;
+stroke-width: 1px;	
 }
 .key2,.fill2{
-	fill: #0000ff;
-	stroke: black;
-	stroke-width: 1px;	
+fill: #0000ff;
+stroke: black;
+stroke-width: 1px;	
 }
 .key3,.fill3{
-	fill: #00ff00;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #00ff00;
+stroke: none;
+stroke-width: 1px;	
 }
 .key4,.fill4{
-	fill: #ffcc00;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #ffcc00;
+stroke: none;
+stroke-width: 1px;	
 }
 .key5,.fill5{
-	fill: #00ccff;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #00ccff;
+stroke: none;
+stroke-width: 1px;	
 }
 .key6,.fill6{
-	fill: #ff00ff;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #ff00ff;
+stroke: none;
+stroke-width: 1px;	
 }
 .key7,.fill7{
-	fill: #00ffff;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #00ffff;
+stroke: none;
+stroke-width: 1px;	
 }
 .key8,.fill8{
-	fill: #ffff00;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #ffff00;
+stroke: none;
+stroke-width: 1px;	
 }
 .key9,.fill9{
-	fill: #cc6666;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #cc6666;
+stroke: none;
+stroke-width: 1px;	
 }
 .key10,.fill10{
-	fill: #663399;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #663399;
+stroke: none;
+stroke-width: 1px;	
 }
 .key11,.fill11{
-	fill: #339900;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #339900;
+stroke: none;
+stroke-width: 1px;	
 }
 .key12,.fill12{
-	fill: #9966FF;
-	stroke: none;
-	stroke-width: 1px;	
+fill: #9966FF;
+stroke: none;
+stroke-width: 1px;	
 }
 EOL
-      	end
-
-end
-end
+      end
+    end
+  end
 end

data/lib/statsample/matrix.rb

@@ -0,0 +1,79 @@
+require 'matrix'
+if RUBY_VERSION<="1.9.0"
+  class ::Vector
+  alias_method :old_coerce, :coerce
+    def coerce(other)
+      case other
+      when Numeric
+        return Matrix::Scalar.new(other), self
+      else
+        raise TypeError, "#{self.class} can't be coerced into #{other.class}"
+      end
+    end
+  end
+end
+
+class ::Matrix
+  def to_gsl
+    out=[]
+    self.row_size.times{|i|
+      out[i]=self.row(i).to_a
+    }
+    GSL::Matrix[*out]
+  end
+
+  # Calculate marginal of rows
+  def rows_sum
+  (0...row_size).collect {|i|
+    row(i).to_a.inject(0) {|a,v| a+v}
+  }
+  end
+  # Calculate marginal of columns
+  def cols_sum
+  (0...column_size).collect {|i|
+    column(i).to_a.inject(0) {|a,v| a+v}
+  }
+  end
+  # Calculate sum of cells
+  def total_sum
+    rows_sum.inject(0){|a,v| a+v}
+  end
+end
+
+module GSL
+  class Matrix
+    def to_matrix
+      rows=self.size1
+      cols=self.size2
+      out=(0...rows).collect{|i| (0...cols).collect {|j| self[i,j]} }
+      ::Matrix.rows(out)
+    end
+  end
+end
+
+module Statsample
+  attr :labels
+  attr :name
+  module CorrelationMatrix
+    def summary
+      rp=ReportBuilder.new()
+      rp.add(self)
+      rp.to_text
+    end
+    def labels=(v)
+      @labels=v
+    end
+    def name=(v)
+      @name=v
+    end
+    def to_reportbuilder(generator)
+      @name||="Correlation Matrix"
+      @labels||=row_size.times.collect {|i| i.to_s} 
+      t=ReportBuilder::Table.new(:name=>@name, :header=>[""]+@labels)
+      row_size.times {|i|
+        t.add_row([@labels[i]]+@rows[i].collect {|i| sprintf("%0.3f",i).gsub("0.",".")})
+      }
+      generator.parse_element(t)
+    end
+  end
+end