feldtruby 0.3.14 → 0.3.16

data/lib/feldtruby/array/permutations_and_subsets.rb

@@ -0,0 +1,11 @@
+class Array
+  def all_pairs
+
+    return [] if length < 2
+
+    x, *rest = self
+
+    rest.map {|e| [x, e]} + rest.all_pairs
+
+  end
+end

data/lib/feldtruby/logger.rb

@@ -0,0 +1,64 @@
+require 'stringio'
+
+module FeldtRuby
+
+class Logger
+  def initialize(io = STDOUT)
+
+    @io = io
+
+    # We save a unique array of events for each type.
+    @events = Hash.new {|h,k| h[k] = Array.new}
+
+  end
+
+  # Return the number of events of type _eventType_.
+  def num_events eventType = nil
+    @events[eventType].length
+  end
+
+  Event = Struct.new(:description, :time_stamp)
+
+  # Count an event of type _eventType_.
+  def log_event eventType = nil, description = ""
+    @events[eventType] << Event.new(description, Time.now)
+  end
+
+  # Return all the events, i.e. descriptions and time stamp, for a given _eventType_.
+  def events(eventType = nil)
+    @events[eventType]
+  end
+
+  # Return all the event descriptions for a given _eventType_.
+  def event_descriptions(eventType = nil)
+    @events[eventType].map {|e| e.description}
+  end
+
+  # Log an event described by a string _str_ optionally with an event type
+  # _eventType_.
+  def log str, eventType = nil
+
+    description = log_entry_description(str, eventType)
+
+    log_event eventType, description
+
+    io_puts description
+
+  end
+
+  # Puts the given _str_ on the io stream.
+  def io_puts str
+    @io.puts str
+  end
+
+  # Map a string and event type to a log string.
+  def log_entry_description str, eventType = nil
+
+    event_tag = eventType ? "{#{eventType.to_s}}: " : ""
+
+    event_tag + str
+
+  end
+end
+
+end

data/lib/feldtruby/minitest_extensions.rb

@@ -20,6 +20,14 @@ module MiniTest::Assertions
   end
 end
 
+NumTestRepetitions = 50
+
+def repeatedly_it(message, &testcode)
+  NumTestRepetitions.times do |i|
+    it("#{i}: " + message, &testcode)
+  end
+end
+
 module MiniTest::Expectations
   infect_an_assertion :assert_similar_proportions, :must_have_similar_proportions
   infect_an_assertion :assert_falsey, :must_be_falsey

data/lib/feldtruby/statistics.rb

@@ -110,8 +110,9 @@ 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.
+  # Convert a Ruby object of one of the types String, Symbol, 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
@@ -119,6 +120,9 @@ class RCommunicator
       when String
         return o.inspect
 
+      when Symbol
+        return o.to_s
+
       when Array
         elems = o.map {|e| ruby_object_to_R_string(e)}.join(", ")
         return "c(#{elems})"
@@ -164,6 +168,10 @@ module Statistics
     res.p_value
   end
 
+  def correlation(ary1, ary2)
+    RC.call("cor", ary1, ary2)
+  end
+
   class DiffusionKDE
     attr_reader :densities, :mesh
 
@@ -246,9 +254,38 @@ module FeldtRuby::Statistics::Plotting
 
   end
 
-  def hexbin_heatmap(csvFilePath, xlabel, ylabel, title = "heatmap", bins = 20)
+  def filled_contour(csvFilePath, xlabel, ylabel, title = "filled.contour")
+    include_library "MASS"
+    #include_library "ggplot2"
+
+    script = <<-EOS
+      data <- read.csv(#{csvFilePath.inspect})
+      k <- with(data, MASS::kde2d(#{xlabel}, #{ylabel}))
+      f <- filled.contour(k, color=topo.colors, 
+             plot.title=title(main = _title_),
+             xlab=_xlabel_, ylab=_ylabel_)
+      f
+    EOS
+
+    subst_eval script, {:title => title,
+      :xlabel => xlabel.to_s, :ylabel => ylabel.to_s}
+
+  end
+
+  def smooth_scatter_plot(csvFilePath, xlabel, ylabel, title = "smoothscatter")
+    include_library "graphics"
+
+    script = <<-EOS
+      f <- ggplot(data, aes(#{xlabel}, #{ylabel})) +
+             geom_point() + geom_smooth( method="loess", se = FALSE )
+    EOS
+
+    plot_2dims(csvFilePath, script, xlabel.to_s, ylabel.to_s, title)
+  end
+
+  def hexbin_heatmap(csvFilePath, xlabel, ylabel, title = "heatmap", bins = 50)
     plot_2dims(csvFilePath,
-      "f <- ggplot(data, aes(#{ylabel}, #{ylabel})) + geom_hex( bins = #{bins} )",
+      "f <- ggplot(data, aes(#{xlabel}, #{ylabel})) + geom_hex( bins = #{bins} )",
       xlabel, ylabel, title)
   end
 
@@ -260,7 +297,7 @@ module FeldtRuby::Statistics::Plotting
       f <- f + stat_smooth()
     EOS
 
-    plot_2dims(csvFilePath, script, xlabel, ylabel, title)
+    plot_2dims(csvFilePath, script, xlabel.to_s, ylabel.to_s, title)
 
   end
 

data/lib/feldtruby/version.rb

@@ -1,3 +1,3 @@
 module FeldtRuby
-  VERSION = "0.3.14"
+  VERSION = "0.3.16"
 end

data/test/helper.rb

@@ -4,4 +4,5 @@ require 'minitest/spec'
 
 $LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
 $LOAD_PATH.unshift(File.dirname(__FILE__))
-require 'feldtruby'
+require 'feldtruby'
+require 'feldtruby/minitest_extensions'

data/test/test_array_permutations_and_subsets.rb

@@ -0,0 +1,43 @@
+require 'feldtruby/array/permutations_and_subsets'
+
+describe "all pairs of elements from an array" do
+
+  it "cannot return any pairs if there are fewer than 2 elements" do
+
+    [].all_pairs.must_equal []
+    [1].all_pairs.must_equal []
+    [:a].all_pairs.must_equal []
+
+  end
+
+  it "returns the only pair if 2 elements" do
+
+    [1, 2].all_pairs.must_equal [[1, 2]]
+    [1, "a"].all_pairs.must_equal [[1, "a"]]
+
+  end
+
+  it "returns the right pairs when 3 elements" do
+
+    [1, 2, 3].all_pairs.sort.must_equal [[1, 2], [1, 3], [2, 3]].sort
+
+  end
+
+  def invariant_all_original_elements_are_at_least_in_one_pair(ary)
+
+    ary.all_pairs.flatten.uniq.sort.must_equal ary.uniq.sort
+
+  end
+
+  repeatedly_it "returns the right number of pairs when > 3 elements" do
+
+    ary = Array.new(4 + rand(10)).map {rand(1000)}
+
+    invariant_all_original_elements_are_at_least_in_one_pair ary
+
+    num_pairs = ary.all_pairs
+    n = ary.length
+
+  end
+
+end

data/test/test_html_doc_getter.rb

@@ -1,16 +1,35 @@
 require 'minitest/spec'
 require 'feldtruby/net/html_doc_getter'
 
-describe "HtmlDocGetter" do
-  it "Can get the html page as a string" do
-    h = FeldtRuby::HtmlDocGetter.new
-    s = h.get("http://www.google.com")
-    s.must_be_instance_of String
+# Skip these network-dependent tests if we have no network connection.
+def has_network_connection?
+  begin
+    open("http://www.google.se")
+    return true
+  rescue Exception => e
+    return false
   end
+end
 
-  it "Can get the html page as a Nokogiri doc" do
-    h = FeldtRuby::HtmlDocGetter.new
-    d = h.get_html_doc("http://www.google.com")
-    d.must_be_instance_of Nokogiri::HTML::Document
+def quicker_has_network_connection?
+  reply = `ping -o google.se`
+  reply != "" # Is empty if no connection since error is printed on STDERR
+end
+
+if quicker_has_network_connection?
+
+  describe "HtmlDocGetter" do
+    it "Can get the html page as a string" do
+      h = FeldtRuby::HtmlDocGetter.new
+      s = h.get("http://www.google.com")
+      s.must_be_instance_of String
+    end
+  
+    it "Can get the html page as a Nokogiri doc" do
+      h = FeldtRuby::HtmlDocGetter.new
+      d = h.get_html_doc("http://www.google.com")
+      d.must_be_instance_of Nokogiri::HTML::Document
+    end
   end
+
 end

data/test/test_logger.rb

@@ -0,0 +1,96 @@
+require 'feldtruby/logger'
+
+describe 'Logger' do
+
+  before do
+    @sio = StringIO.new
+    @l = FeldtRuby::Logger.new @sio
+  end
+
+  it 'has a event count of 0 when no events has been logged' do
+
+    @l.num_events.must_equal 0
+
+  end
+
+  it 'increases default event count when default events are logged' do
+
+    @l.log "1"
+    @l.num_events.must_equal 1
+
+    @l.log "2"
+    @l.num_events.must_equal 2
+
+  end
+
+  it 'can log default events (described in strings)' do
+
+    @l.log "event 1"
+    @sio.string.must_equal "event 1\n"
+    @l.num_events.must_equal 1
+
+    @l.log "event 2"
+    @sio.string.must_equal "event 1\nevent 2\n"
+    @l.num_events.must_equal 2
+
+  end
+
+  it 'can log events of a given type' do
+
+    @l.log "1", :increase
+    @sio.string.must_equal "{increase}: 1\n"
+    @l.num_events.must_equal 0
+    @l.num_events(:increase).must_equal 1
+
+    @l.log "2", :increase
+    @sio.string.must_equal "{increase}: 1\n{increase}: 2\n"
+    @l.num_events.must_equal 0
+    @l.num_events(:increase).must_equal 2
+
+  end
+
+  it 'can return old default events' do
+
+    @l.log "event 1"
+    @l.event_descriptions.must_equal ["event 1"]
+
+    @l.log "event 2"
+    @l.event_descriptions.must_equal ["event 1", "event 2"]
+
+  end
+
+  it 'can return old events of given type' do
+
+    @l.log "1", :increase
+    @l.log "2", :increase
+    @l.log "0.4", :alpha
+    @l.log "1", :increase
+
+    @l.event_descriptions(:increase).must_equal ["{increase}: 1", "{increase}: 2", "{increase}: 1"]
+
+    @l.event_descriptions(:alpha).must_equal ["{alpha}: 0.4"]
+
+    @l.event_descriptions(:beta).must_equal []
+
+  end
+
+  it 'time stamps each log entry' do
+
+    @l.log "1", :a
+    @l.log "2", :b
+    @l.log "2", :a
+
+    time_stamps_a = @l.events(:a).map {|e| e.time_stamp}
+    time_stamps_b = @l.events(:b).map {|e| e.time_stamp}
+
+    time_stamps_a[0].must_be_instance_of Time
+    time_stamps_a[1].must_be_instance_of Time
+    time_stamps_b[0].must_be_instance_of Time
+
+    time_stamps_a[0].must_be :<, time_stamps_a[1]
+    time_stamps_a[0].must_be :<, time_stamps_b[0]
+    time_stamps_b[0].must_be :<, time_stamps_a[1]
+
+  end
+
+end

data/test/test_statistics.rb

@@ -82,6 +82,18 @@ describe "Statistics" do
       kde.densities[3].must_be_close_to 0.2728
     end
   end
+
+  describe "correlation between Ruby arrays" do
+
+    it "works for a simple example from the R doc for cor" do
+      correlation( (1..10).to_a, (2..11).to_a ).must_equal 1
+    end
+
+    it "works for the inverse of the simple example from the R doc for cor" do
+      correlation( (1..10).to_a, (-11..-2).to_a.reverse ).must_equal -1
+    end
+
+  end
 end
 
 require 'feldtruby/minitest_extensions'
@@ -154,7 +166,7 @@ describe "Plotting" do
 
     File.exist?(out).must_equal true
 
-    #File.delete out
+    File.delete out
 
   end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: feldtruby
 version: !ruby/object:Gem::Version
-  version: 0.3.14
+  version: 0.3.16
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-03-11 00:00:00.000000000 Z
+date: 2013-03-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rinruby
@@ -79,9 +79,11 @@ files:
 - lib/feldtruby/array.rb
 - lib/feldtruby/array/basic_stats.rb
 - lib/feldtruby/array/count_by.rb
+- lib/feldtruby/array/permutations_and_subsets.rb
 - lib/feldtruby/file/file_change_watcher.rb
 - lib/feldtruby/file/tempfile.rb
 - lib/feldtruby/float.rb
+- lib/feldtruby/logger.rb
 - lib/feldtruby/math/rand.rb
 - lib/feldtruby/minitest_extensions.rb
 - lib/feldtruby/net/html_doc_getter.rb
@@ -115,11 +117,13 @@ files:
 - test/test_array.rb
 - test/test_array_basic_stats.rb
 - test/test_array_count_by.rb
+- test/test_array_permutations_and_subsets.rb
 - test/test_clustering.rb
 - test/test_euclidean_distance.rb
 - test/test_fastmap.rb
 - test/test_float.rb
 - test/test_html_doc_getter.rb
+- test/test_logger.rb
 - test/test_normalization.rb
 - test/test_optimize.rb
 - test/test_optimize_differential_evolution.rb
@@ -165,11 +169,13 @@ test_files:
 - test/test_array.rb
 - test/test_array_basic_stats.rb
 - test/test_array_count_by.rb
+- test/test_array_permutations_and_subsets.rb
 - test/test_clustering.rb
 - test/test_euclidean_distance.rb
 - test/test_fastmap.rb
 - test/test_float.rb
 - test/test_html_doc_getter.rb
+- test/test_logger.rb
 - test/test_normalization.rb
 - test/test_optimize.rb
 - test/test_optimize_differential_evolution.rb