biopsy 0.2.1 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3f45b9377c3de44caf8abf21b07576a6e89e0c03
-  data.tar.gz: 94e4a81fe98380482adc7b290f8067174684850d
+  metadata.gz: 227832bb75ba1c0114e59d8058a55c2ead937105
+  data.tar.gz: a3f5e6b364d5265878ed4a8f4e7f086282c2a156
 SHA512:
-  metadata.gz: 76d90535589d1f09f74bf424d7437a45855b1065b802bc3ba4d562cc30c1c8fd775583cd856e6de02fee919a9b34619875693a96b228ef686c7e658bb9b29ea5
-  data.tar.gz: 0dbc5e5d21297759affc2199ffff2fe90d25c1958a7380d531811bdded9fbe8caf8eb9fc0aed8b62ecb5d52d6f0f2e2292b480ae8a8e84c79f0e457f44f45c3a
+  metadata.gz: 06e5510f5546f565c080ff405f3406e27d8a69976ac3fe80a89814b85a9d24cbe98aa9fca01c0a38dd26d46af6a7889ebc36e1b250939577594ab4226f0180b9
+  data.tar.gz: 326da0583c62503acdd211ece069434ea63cbaafcce2e6fa5caabb4d3c3ff7e9125ada20dbf6c91e0f62c2f1f4f6bc32098b1300ee78b159a0ffc24807daba7d

data/lib/biopsy/experiment.rb

@@ -23,7 +23,7 @@ module Biopsy
 
     # Returns a new Experiment
     def initialize(target, options:{}, threads:4, start:nil, algorithm:nil,
-                   timelimit:nil, verbosity: :quiet)
+                   timelimit:nil, verbosity: :quiet, id:nil)
       @threads = threads
       @start = start
       @algorithm = algorithm
@@ -40,6 +40,7 @@ module Biopsy
       self.select_starting_point
       @scores = {}
       @iteration_count = 0
+      set_id id
     end
 
     # return the set of parameters to evaluate first
@@ -63,9 +64,9 @@ module Biopsy
       max = Settings.instance.sweep_cutoff
       n = @target.count_parameter_permutations
       if n < max
-        @algorithm = ParameterSweeper.new(@target.parameters)
+        @algorithm = ParameterSweeper.new(@target.parameters, @id)
       else
-        @algorithm = TabuSearch.new(@target.parameters)
+        @algorithm = TabuSearch.new(@target.parameters, @id)
       end
     end
 
@@ -83,6 +84,7 @@ module Biopsy
       in_progress = true
       @algorithm.setup @start
       @current_params = @start
+      max_scores = @target.count_parameter_permutations
       while in_progress
         run_iteration
         # update the best result
@@ -100,7 +102,7 @@ module Biopsy
            end
         end
         # have we finished?
-        in_progress = !@algorithm.finished?
+        in_progress = !@algorithm.finished? && @scores.size < max_scores
         if in_progress && !(@timelimit.nil?)
           in_progress = (Time.now - start_time < @timelimit)
         end
@@ -117,25 +119,27 @@ module Biopsy
     # encompassing the program, objective(s) and optimiser.
     # Returns the output of the optimiser.
     def run_iteration
-      # create temp dir
-      Dir.chdir(self.create_tempdir) do
+      param_key = @current_params.to_s
+      result = nil
+      # lookup the result if possible
+      if @scores.key? param_key
+        result = @scores[param_key]
+      else
+        # create temp dir
+        curdir = Dir.pwd
+        Dir.chdir(self.create_tempdir) unless Settings.instance.no_tempdirs
         # run the target
         raw_output = @target.run @current_params.merge(@options)
         # evaluate with objectives
-        param_key = @current_params.to_s
-        result = nil
-        if @scores.key? param_key
-          result = @scores[param_key]
-        else
-          result = @objective.run_for_output(raw_output, @threads, nil)
-          @iteration_count += 1
-          self.print_progress(@iteration_count, @current_params, result, @best)
-        end
+        result = @objective.run_for_output(raw_output, @threads, nil)
+        @iteration_count += 1
+        self.print_progress(@iteration_count, @current_params, result, @best)
         @scores[@current_params.to_s] = result
-        # get next steps from optimiser
-        @current_params = @algorithm.run_one_iteration(@current_params, result)
+        self.cleanup
+        Dir.chdir(curdir) unless Settings.instance.no_tempdirs
       end
-      self.cleanup
+      # get next steps from optimiser
+      @current_params = @algorithm.run_one_iteration(@current_params, result)
     end
 
     def print_progress(iteration, params, score, best)
@@ -148,6 +152,7 @@ module Biopsy
     end
 
     def cleanup
+      return if Settings.instance.no_tempdirs
       # TODO: make this work
       # remove all but essential files
       essential_files = ""
@@ -183,6 +188,17 @@ module Biopsy
       token
     end
 
+    # set experiment ID with either user provided value, or date-time
+    # as fallback
+    def set_id id
+      @id = id
+      if @id.nil?
+        t = Time.now
+        parts = %w[y m d H M S Z].map{ |p| t.strftime "%#{p}" }
+        @id = "experiment_#{parts.join('_')}"
+      end
+    end
+
   end # end of class RunHandler
 
 end # end of module Biopsy

data/lib/biopsy/optimisers/parameter_sweeper.rb

@@ -24,11 +24,12 @@ module Biopsy
   class Combinator
 
     include Enumerable
-   
-    def initialize parameters
+
+    def initialize(parameters, id)
       @parameters = parameters
+      @id = id
     end
-   
+
     def generate_combinations(index, opts, &block)
       if index == @parameters.length
         block.call opts.clone
@@ -80,7 +81,7 @@ module Biopsy
       @current = { :parameters => parameters, :score => score }
       self.update_best?
       return @combinator.next
-    rescue 
+    rescue
       @is_finished = true
       return nil
     end

data/lib/biopsy/optimisers/tabu_search.rb

@@ -90,7 +90,7 @@ module Biopsy
 
     def initialize(distributions, max_size, tabu)
       # tabu
-      @tabu = tabu 
+      @tabu = tabu
       # neighbourhood
       @max_size = max_size
       @members = []
@@ -108,7 +108,7 @@ module Biopsy
       n = 0
       begin
         if n >= 100
-          # taking too long to generate a neighbour, 
+          # taking too long to generate a neighbour,
           # loosen the neighbourhood structure so we explore further
           # debug("loosening distributions")
           @distributions.each do |param, dist|
@@ -162,15 +162,16 @@ module Biopsy
     attr_accessor :max_hood_size, :sd_increment_proportion
     attr_accessor :starting_sd_divisor, :backtrack_cutoff, :jump_cutoff
 
-    Thread = Struct.new(:best, :tabu, :distributions, 
-                        :standard_deviations, :recent_scores, 
+    Thread = Struct.new(:best, :tabu, :distributions,
+                        :standard_deviations, :recent_scores,
                         :iterations_since_best, :backtracks,
                         :current, :current_hood, :loaded,
                         :score_history, :best_history)
 
-    def initialize(parameter_ranges, threads=8, limit=nil)
+    def initialize(parameter_ranges, id, threads=8, limit=nil)
 
       @ranges = parameter_ranges
+      @id = id
 
       # solution tracking
       @best = nil
@@ -194,7 +195,7 @@ module Biopsy
       # logging
       @score_history = []
       @best_history = []
-      @log_data = false
+      @log_data = true
       @logfiles = {}
       self.log_setup
 
@@ -323,7 +324,7 @@ module Biopsy
       mean = @ranges[param].index(value)
       range = @ranges[param]
       sd = self.sd_for_param(param, range)
-      @distributions[param] = Biopsy::Distribution.new(mean, 
+      @distributions[param] = Biopsy::Distribution.new(mean,
                                                       range,
                                                       @sd_increment_proportion,
                                                       sd)
@@ -348,14 +349,13 @@ module Biopsy
       end
       if best[:parameters].nil?
         # this should never happen!
-        best = @best        
+        best = @best
       end
       best
     end
 
     def backtrack
       @backtracks += 1.0
-      # debug('backtracked to best')
       @distributions.each_pair { |k, d| d.tighten }
     end
 
@@ -368,8 +368,8 @@ module Biopsy
     # use the gradient of recent best scores to update the distributions
     def adjust_distributions_using_gradient
       return if @recent_scores.length < 3
-      vx = (1..@recent_scores.length).to_a.to_scale
-      vy = @recent_scores.reverse.to_scale
+      vx = (1..@recent_scores.length).to_a.to_numeric
+      vy = @recent_scores.reverse.to_numeric
       r = Statsample::Regression::Simple.new_from_vectors(vx,vy)
       slope = r.b
       if slope > 0
@@ -398,15 +398,17 @@ module Biopsy
       end
     end
 
-    # check termination conditions 
+    # check termination conditions
     # and return true if met
     def finished?
-      return false unless @threads.all? { |t| t.recent_scores.size == @jump_cutoff }
+      return false unless @threads.all? do |t|
+        t.recent_scores.size == @jump_cutoff
+      end
       probabilities = self.recent_scores_combination_test
       n_significant = 0
-      probabilities.each do |mann_u, levene| 
+      probabilities.each do |mann_u, levene|
         if mann_u <= @adjusted_alpha && levene <= @convergence_alpha
-          n_significant += 1 
+          n_significant += 1
         end
       end
       finish = n_significant >= probabilities.size * 0.5
@@ -415,8 +417,8 @@ module Biopsy
     # returns a matrix of correlation probabilities for recent
     # scores between all threads
     def recent_scores_combination_test
-      combinations = 
-      @threads.map{ |t| t.recent_scores.to_scale }.combination(2).to_a
+      combinations =
+      @threads.map{ |t| t.recent_scores.to_numeric }.combination(2).to_a
       combinations.map do |a, b|
         [Statsample::Test.u_mannwhitney(a, b).probability_exact,
          Statsample::Test::Levene.new([a,b]).probability]
@@ -431,10 +433,10 @@ module Biopsy
     def log_setup
       if @log_data
         require 'csv'
-        @logfiles[:standard_deviations] = CSV.open('standard_deviations.csv', 'w')
-        @logfiles[:best] = CSV.open('best.csv', 'w')
-        @logfiles[:score] = CSV.open('score.csv', 'w')
-        @logfiles[:params] = CSV.open('params.csv', 'w')
+        @logfiles[:standard_deviations] = CSV.open("#{@id}_standard_deviations.csv", 'w')
+        @logfiles[:best] = CSV.open("#{@id}_best.csv", 'w')
+        @logfiles[:score] = CSV.open("#{@id}_score.csv", 'w')
+        @logfiles[:params] = CSV.open("#{@id}_params.csv", 'w')
       end
     end
 
@@ -466,13 +468,14 @@ module Biopsy
     end
 
     def random_start_point
-      Hash[@ranges.map { |p, r| [p, r.sample] }] 
+      Hash[@ranges.map { |p, r| [p, r.sample] }]
     end
 
     def write_data
       require 'csv'
-      now = Time.now.to_i
-      CSV.open("../#{now}_scores.csv", "w") do |c|
+      pathmod = Settings.instance.no_tempdirs ? '' : '../'
+      path = File.expand_path("#{pathmod}#{@id}_scores.csv")
+      CSV.open(path, "w") do |c|
         c << %w(iteration thread score best)
         @threads.each_with_index do |t, t_idx|
           sh = t.score_history
@@ -482,8 +485,7 @@ module Biopsy
           end
         end
       end
-      path = File.expand_path("../#{now}_scores.csv")
-      puts "wrote TabuSearch run data to #{path}"
+      # puts "wrote TabuSearch run data to #{path}"
     end
 
   end # TabuSearch

data/lib/biopsy/settings.rb

@@ -30,6 +30,7 @@ module Biopsy
     attr_accessor :sweep_cutoff
     attr_accessor :keep_intermediates
     attr_accessor :gzip_intermediates
+    attr_accessor :no_tempdirs
 
     def initialize
       self.set_defaults
@@ -45,6 +46,7 @@ module Biopsy
       @sweep_cutoff = 100
       @keep_intermediates = false
       @gzip_intermediates = false
+      @no_tempdirs = false
     end
 
     # Loads settings from a YAML config file. If no file is

data/lib/biopsy/version.rb

@@ -5,8 +5,8 @@ module Biopsy
 
   module VERSION
     MAJOR = 0
-    MINOR = 2
-    PATCH = 1
+    MINOR = 3
+    PATCH = 0
 
     STRING = [MAJOR, MINOR, PATCH].compact.join('.')
   end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: biopsy
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.3.0
 platform: ruby
 authors:
 - Richard Smith-Unna
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-06-02 00:00:00.000000000 Z
+date: 2015-11-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -206,7 +206,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.2.2
+rubygems_version: 2.4.6
 signing_key: 
 specification_version: 4
 summary: framework for optimising any computational pipeline or program