plot_from_tsv.rb

#!/usr/bin/env ruby

require 'time'
require 'gnuplot'
require 'launchy'
require 'tempfile'
require 'fileutils'

ARGV.sort!.reverse!
raise ArgumentError.new("USAGE: #{__FILE__} input.tsv [--without-items] [--without-capacity] [--save]") if ARGV.empty?

data = {}

File.foreach(ARGV[0]) do |line|
  next unless line.include?('DynamoDB query')
  columns = line.split("\t")

  timestamp = Time.parse("#{columns[1]} +0000").to_i
  captures = columns[9].match(/DynamoDB query \((?<items>[\d]+) items, .*, (?<capacity>[\d\.]+) capacity/)

  data[timestamp] ||= {}
  %i(capacity items).each do |key|
    data[timestamp][key] = data[timestamp].fetch(key, 0) + captures[key].to_f
  end
end

tempfile = Tempfile.new(['plot', '.png'])

Gnuplot.open do |gp|
  Gnuplot::Plot.new(gp) do |plot|
    plot.term('png size 1200,600')
    plot.output(tempfile.path)
    plot.xlabel('Time')
    plot.ylabel('Value')
    plot.xdata(:time)
    plot.timefmt('"%s"')

    x = data.keys

    %i(capacity items).each do |property|
      unless ARGV.include?("--without-#{property.to_s}")
        y = data.values.map { |hash| hash.fetch(property) }
        plot.data.push Gnuplot::DataSet.new([x, y]) { |dataset|
          dataset.using = "1:2 title '#{property.to_s.capitalize}'"
          dataset.with = 'impulses'
        }
      end
    end
  end
end

if File.exist?(tempfile.path)
  if ARGV.include?('--save')
    path = File.join(File.dirname(__FILE__), File.basename(tempfile.path))
    FileUtils.cp(tempfile.path, path)
    Launchy.open(path)
  else
    Launchy.open(tempfile.path)
    sleep(1) # Make sure that Ruby won't remove temporary file before opening it
  end
end