XML Integrity Checking

xml_testing.rb

# frozen_string_literal: true

# https://gist.github.com/remomueller/c853734a2f85937b69b373c46760fcdd
# gem install xml-simple colorize --no-document

# ruby xml_testing.rb

# XML Integrity Checking

# I think checking through all XML for the following would be a good place to
# start on a list to check:

# 1. No scored events at all (<ScoredEvent> area essentially empty, save for
#    staging data)
# 2. Arousal <3 seconds in duration (<ScoredEvent> with "Arousal" in <Name> and
#    <Duration> less than 3 seconds)
# 3. Respiratory event <10 seconds in duration (<ScoredEvent> with "Apnea",
#    "Hypopnea", or "Unsure" in <Name> and <Duration> less than 10 seconds)

# This could sort of work like "edfize" where it checks each individual file for
# any failures and then spits out a warning when it encounters an issue. For #2
# and #3 it might also be nice to have the start time of the event(s) flagged in
# the output to know where to look when inspecting the XML manually.

require "rubygems"
require "colorize"
require "json"
require "xmlsimple"

def extract(hash, key, convert: nil)
  value = hash[key] && !hash[key][0].empty? ? hash[key][0] : nil
  convert && value ? value.send(convert) : value
end

def starts_with?(search, word)
  !(/^#{search}/i =~ word).nil?
end

def contains_word?(search, word)
  !(/#{search}/i =~ word).nil?
end

good_count = 0
bad_count = 0

xmls = Dir.glob("**/*.xml", File::FNM_CASEFOLD)
xml_count = xmls.size

xmls.each_with_index do |xml_path, index|
  print "\r#{index + 1} of #{xml_count} (#{(index + 1) * 100 / xml_count}%)"

  xml = XmlSimple.xml_in(xml_path)

  scored_events = \
    if xml["ScoredEvents"].nil? || xml["ScoredEvents"][0].nil? || xml["ScoredEvents"][0]["ScoredEvent"].nil?
      []
    else
      xml["ScoredEvents"][0]["ScoredEvent"].collect do |hash|
        {
          name: extract(hash, "Name"),
          lowest_spo2: extract(hash, "LowestSpO2", convert: :to_f),
          desaturation: extract(hash, "Desaturation", convert: :to_f),
          start: extract(hash, "Start", convert: :to_f),
          duration: extract(hash, "Duration", convert: :to_f),
          input: extract(hash, "Input")
        }
      end
    end

  arousal_events = scored_events.select do |event|
    starts_with?("Arousal", event[:name]) && event[:duration] < 3
  end

  respiratory_events = scored_events.select do |event|
    (
      contains_word?("Apnea", event[:name]) ||
      contains_word?("Hypopnea", event[:name]) ||
      contains_word?("Unsure", event[:name])
    ) && event[:duration] < 10
  end

  if !scored_events.empty? && arousal_events.empty? && respiratory_events.empty?
    good_count += 1
  else
    bad_count += 1
    puts "\n#{xml_path}"
    puts "Empty: No Scored Events" if scored_events.empty?
    unless arousal_events.empty?
      puts "Detected: Arousal Events <3 seconds"
      puts "   Start: #{arousal_events.collect { |event| event[:start] }.join(', ')}"
    end
    unless respiratory_events.empty?
      puts "Detected: Respiratory Events <10 seconds"
      puts "   Start: #{respiratory_events.collect { |event| event[:start] }.join(', ')}"
    end
    puts "\n"
  end
end

puts "\n" + good_count.to_s.colorize(good_count.positive? ? :green : nil) + " XML#{good_count == 1 ? ' ' : 's'} passed"
puts bad_count.to_s.colorize(bad_count.positive? ? :red : nil) + " XML#{bad_count == 1 ? ' ' : 's'} failed"