feldtruby 0.3.11 → 0.3.12

data/Gemfile.lock

@@ -1,14 +1,16 @@
 PATH
   remote: .
   specs:
-    feldtruby (0.3.3)
+    feldtruby (0.3.11)
       json
+      nokogiri
       rinruby
 
 GEM
   remote: https://rubygems.org/
   specs:
-    json (1.7.6)
+    json (1.7.7)
+    nokogiri (1.5.6)
     rinruby (2.0.3)
 
 PLATFORMS

data/feldtruby.gemspec

@@ -18,5 +18,8 @@ Gem::Specification.new do |gem|
   gem.require_paths = ["lib"]
 
   gem.add_dependency('rinruby')
+
   gem.add_dependency('json')
+
+  gem.add_dependency('nokogiri')
 end

data/lib/feldtruby/statistics.rb

@@ -39,6 +39,21 @@ class RCommunicator
     @r.eval str
   end
 
+  # Given a script that has variable references in the form "_name_" insert
+  # the ruby objects mapped from these names in scriptNameToRubyValues
+  def subst_eval(script, scriptNameToRubyValues)
+
+    scriptNameToRubyValues.each do |key, value|
+
+      script = script.gsub("_#{key.to_s}_", ruby_object_to_R_string(value))
+
+    end
+
+    #puts "Eval'ing script:\n#{script}"
+    eval script
+
+  end
+
   # This represents a hash returned as JSON from R but mapped to a
   # Ruby object so we can more easily use it as if it was an R object.
   class Rvalue
@@ -85,6 +100,32 @@ class RCommunicator
     end
   end
 
+  # Convert a Ruby object of one of the types String, Array, Integer or Float
+  # to a String that can be used in R code/scripts to represent the object.
+  def ruby_object_to_R_string(o)
+
+      case o
+
+      when String
+        return o.inspect
+
+      when Array
+        elems = o.map {|e| ruby_object_to_R_string(e)}.join(", ")
+        return "c(#{elems})"
+
+      when Integer
+        return o.to_s
+
+      when Float
+        return o.to_s
+
+      else
+        raise "Cannot represent object #{o} in valid R code"
+
+      end
+
+  end
+
   private
 
   def res_name(index = 1)
@@ -156,7 +197,51 @@ end
 # Plotting data sets in R with ggplot2 and save them to files.
 module FeldtRuby::Statistics::Plotting
 
-  def plot_2dims(csvFilePath, graphFilePath, xName, yName, title = "scatterplot", format = "pdf", width = 7, height = 5)
+  GfxFormatToGfxParams = {
+    "pdf" => {:width => 7, :height => 5, :paper => 'special'},
+    "png" => {:units => "cm", :width => 12, :height => 8},
+    "tiff" => {:units => "cm", :width => 12, :height => 8},
+  }
+
+  def gfx_device(format, width = nil, height = nil)
+
+    format = format.to_s            # If given as a symbol instead of a string
+
+    unless GfxFormatToGfxParams.has_key?(format)
+      raise ArgumentError.new("Don't now about gfx format #{format}")
+    end
+
+    params = GfxFormatToGfxParams[format]
+
+    "#{format}(#{hash_to_R_params(params)})"
+
+  end
+
+  # Map a ruby hash of objects to parameters in R code/script.
+  def hash_to_R_params(hash)
+
+    hash.keys.sort.map do |key|
+
+      "#{key.to_s} = #{ruby_object_to_R_string(hash[key])}"
+
+    end.join(", ")
+
+  end
+
+  def set_file_ending(filepath, newEnding)
+
+    dirname = File.dirname(filepath)
+
+    current_ending = filepath.split(".").last
+
+    # Works even if there is no current file ending:
+    basename = File.basename(filepath, "." + current_ending)
+
+    File.join dirname, basename
+
+  end
+
+  def plot_2dims(csvFilePath, graphFilePath, xName, yName, title = "scatterplot", format = "pdf", width = nil, height = nil)
 
     include_library("ggplot2")
 
@@ -164,13 +249,13 @@ module FeldtRuby::Statistics::Plotting
 
     pre = [
       "data <- read.csv(#{csvFilePath.inspect})",
-      "#{format}(#{gfxFile.inspect}, width=#{width}, height=#{height})"
+      gfx_device(format, width, height)
     ]
 
     #plot = ["suppressWarnings( " + yield().join(" ") + " + theme_bw(base_size = 12, base_family = \"\") )"]
-    plot = [yield().join(" ") + " + theme_bw(base_size = 12, base_family = \"\")"]
-    #plot = yield()
-    #plot << " + theme_bw(base_size = 16, base_family = \"\")"
+    #plot = [yield().join(" ") + " + theme_bw(base_size = 12, base_family = \"\")"]
+    plot = yield()
+    plot << " + theme_bw(base_size = 12, base_family = \"\")"
 
     post = [
       "dev.off()"
@@ -182,12 +267,13 @@ module FeldtRuby::Statistics::Plotting
   end
 
   # Scatter plot of columns xName vs yName in csvFilePath is saved to graphFilePath.
-  def scatter_plot(csvFilePath, graphFilePath, xName, yName, title = "scatterplot", smoothFit = true, format = "pdf", width = 7, height = 5)
+  def scatter_plot(csvFilePath, graphFilePath, xName, yName, title = "scatterplot", smoothFit = true, format = nil, width = 7, height = 5)
     plot_2dims(csvFilePath, graphFilePath, xName, yName, title, format, width, height) {
       [
+        "smoothing_method <- if(nrow(data) > 1000) {'gam'} else {'loess'}",
         "ggplot(data, aes(#{xName}, #{yName})) + ",
         "  geom_point(shape = 1) + ", # Each point is non-filled circle
-        (smoothFit ? "  geom_smooth() + " : nil),
+        (smoothFit ? "  geom_smooth(method = smoothing_method) + " : nil),
         "  ggtitle(#{title.inspect})"
       ].compact
     }
@@ -199,6 +285,39 @@ module FeldtRuby::Statistics::Plotting
       [ "ggplot(data, aes(#{xName}, #{yName})) + geom_hex( bins = #{bins} ) + ggtitle(\"#{title}\")"]
     }
   end
+
+  # This is wrapped around a block that draws a diagram and will save that diagram
+  # to the given filename.
+  def save_graph(filename)
+    RC.eval("pdf(#{filename.inspect}, width = 6, height = 4, paper = 'special')")
+    yield() # Just be sure not to nest these save_graph calls within each other...
+    RC.eval("dev.off()")
+  end
+
+  # Overlaid density graph of the observations (sampled distributions) in data1
+  # and data2.
+  def overlaid_densities(data1, data2, xlabel = "x", ylabel = "density")
+    include_library("ggplot2")
+
+    raise ArgumentError.new("Must have same cardinality") unless data1.length == data2.length
+
+    script = <<-EOS
+      df <- data.frame(dc1 = _d1_, dc2 = _d2_)
+      f <- ggplot(df)
+      f <- f + geom_density(aes(dc1, fill = "blue"), alpha = 0.2)
+      f <- f + geom_density(aes(dc2, fill = "red"), alpha = 0.2)
+      f <- f + theme_bw() + ggtitle("Overlaid densities")
+      f <- f + theme(plot.title = element_text(face="bold", size=12), 
+       axis.title.x = element_text(face="bold", size=10)
+      ) + xlab(_xlabel_)
+      # pdf("tmp.pdf", width = 6, height = 4, paper='special')
+      f
+      #dev.off()
+    EOS
+    subst_eval script, {:d1 => data1, :d2 => data2, 
+      :xlabel => xlabel, :ylabel => ylabel}
+
+  end
 end
 
 class FeldtRuby::RCommunicator

data/lib/feldtruby/version.rb

@@ -1,3 +1,3 @@
 module FeldtRuby
-  VERSION = "0.3.11"
+  VERSION = "0.3.12"
 end

data/test/test_statistics.rb

@@ -100,32 +100,89 @@ end
 
 describe "Plotting" do
 
-  it "can do a scatter plot" do
+  it "can map Ruby integers to R code/script strings" do
 
-    d = File.dirname(__FILE__) + "/"
-    filename = d + "tmp.csv"
-    out = d + "scatterplot.pdf"
-    RC.scatter_plot(filename, out, "size", "height", 
-      "Scatterplot", true)
+    RC.ruby_object_to_R_string(1).must_equal "1"
+    RC.ruby_object_to_R_string(42).must_equal "42"
 
-    File.exist?(out).must_equal true
+  end
 
-    File.delete out
+  it "can map Ruby floats to R code/script strings" do
+
+    RC.ruby_object_to_R_string(3.675).must_equal "3.675"
+    RC.ruby_object_to_R_string(1e10).must_equal "10000000000.0"
 
   end
 
-  it "can do a hexbin heatmap plot" do
+  it "can map Ruby arrays to R code/script strings" do
 
-    d = File.dirname(__FILE__) + "/"
-    filename = d + "tmp.csv"
-    out = d + "heatmap.pdf"
+    RC.ruby_object_to_R_string([1,2,3]).must_equal "c(1, 2, 3)"
+    RC.ruby_object_to_R_string([10, 1.65]).must_equal "c(10, 1.65)"
 
-    RC.hexbin_heatmap(filename, out, "size", "height", 
-      "Hexbin heatmap", 30)
+  end
 
-    File.exist?(out).must_equal true
+  it "can map Ruby strings to R code/script strings" do
 
-    File.delete out
+    RC.ruby_object_to_R_string("loess").must_equal '"loess"'
+    RC.ruby_object_to_R_string("gam").must_equal '"gam"'
+
+  end
+
+  it "can convert a hash of Ruby objects into a R parameter script" do
+
+    RC.hash_to_R_params({:a => 1, :b => 42.5}).must_equal "a = 1, b = 42.5"
 
+    s = RC.hash_to_R_params({:b => "b", :height => [5, 7.2]})
+    s.must_equal 'b = "b", height = c(5, 7.2)'
+
+  end
+
+  it "can change the file ending if is not what is expected" do
+
+  end
+
+#  it "can do a scatter plot" do
+#
+#    d = File.dirname(__FILE__) + "/"
+#    filename = d + "tmp.csv"
+#    out = d + "scatterplot.pdf"
+#    RC.scatter_plot(filename, out, "size", "height", 
+#      "Scatterplot", true)
+#
+#    File.exist?(out).must_equal true
+#
+#    File.delete out
+#
+#  end
+#
+#  it "can do a hexbin heatmap plot" do
+#
+#    d = File.dirname(__FILE__) + "/"
+#    filename = d + "tmp.csv"
+#    out = d + "heatmap.pdf"
+#
+#    RC.hexbin_heatmap(filename, out, "size", "height", 
+#      "Hexbin heatmap", 30)
+#
+#    File.exist?(out).must_equal true
+#
+#    File.delete out
+#
+#  end
+
+  it "can do overlaid density plot" do
+
+    d1 = Array.new(100) {rand(10)}
+    d2 = Array.new(100) {2 + rand(5)}
+
+    out = "tmp.pdf"
+
+    RC.save_graph(out) do
+      RC.overlaid_densities(d1, d2)
+    end
+
+    File.exist?(out).must_equal true
+    File.delete out
+    
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: feldtruby
 version: !ruby/object:Gem::Version
-  version: 0.3.11
+  version: 0.3.12
   prerelease: 
 platform: ruby
 authors:
@@ -43,6 +43,22 @@ dependencies:
     - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Robert Feldt's Common Ruby Code lib
 email:
 - robert.feldt@gmail.com
@@ -57,6 +73,7 @@ files:
 - R/diffusion_kde.R
 - README.md
 - Rakefile
+- Rplots.pdf
 - TODO
 - feldtruby.gemspec
 - lib/feldtruby.rb