Contents

#!/usr/bin/env ruby

require 'MESH'

class Numeric
  def duration
    secs, millisecs = self.divmod 1
    # secs = self.to_int
    mins = secs / 60
    hours = mins / 60
    days = hours / 24

    if days > 0
      "#{days} days and #{hours % 24} hours"
    elsif hours > 0
      "#{hours} hours and #{mins % 60} minutes"
    elsif mins > 0
      "#{mins} minutes and #{secs % 60} seconds"
    elsif secs >= 0
      "#{(millisecs + secs).round(3)} seconds"
    end
  end
end

def time_this(name, &block)
  print "#{name}"
  STDOUT.flush
  start = Time.now.to_f
  result = yield
  finish = Time.now.to_f
  puts "\t#{(finish - start).duration}"
  result
end

mesh_tree = time_this('Loading MeSH Tree') { MESH::Tree.new }
time_this('Loading en_gb translation') { mesh_tree.load_translation(:en_gb) }
time_this('Loading wikipedia') { mesh_tree.load_wikipedia }

json_str = File.new('./example.json').read
extracted = JSON.parse(json_str)

title_headings = time_this('Matching in title') { mesh_tree.match_in_text(extracted['title']) }
description_headings = time_this('Matching in description') { mesh_tree.match_in_text(extracted['description']) }
content_headings = time_this('Matching in content') { mesh_tree.match_in_text(extracted['content']) }

classifier = MESH::Classifier.new()
classification = time_this('Classifying from matches') { classifier.classify([
                                                                                 {weight: 10.0, matches: title_headings},
                                                                                 {weight: 5.0, matches: description_headings},
                                                                                 {weight: 1.0, matches: content_headings}
                                                                             ]) }

Version data entries

1 entries across 1 versions & 1 rubygems

Version	Path
mesh-medical-subject-headings-2.3.0	bin/benchmark_match_in_text