#!/usr/bin/env ruby
#
# Simple implementation of regression tests for xml generated by parse-speeches.rb
# N.B. Need to pre-populate reference xml files with those that have previously been generated.
# In other words, this is only useful for checking that any refactoring has not caused a regression in behaviour.
#
$:.unshift "#{File.dirname(__FILE__)}/../lib"
require 'people'
require 'hansard_parser'
require 'configuration'
# Range of dates to test
from_date = Date.new(2007, 1, 1)
to_date = Date.new(2008, 1, 1) - 1
# Number of items to skip at the beginning
skip = 0
# Dates to test first before anything else
# Update this list with any dates that have shown up problems in the past
test_first = [Date.new(2007,8,8), Date.new(2007,8,14), Date.new(2007,5,8), Date.new(2007,2,14)]
skip_dates = []
#
conf = Configuration.new
# First load people back in so that we can look up member id's
people = PeopleCSVReader.read_members
parser = HansardParser.new(people)
def compare_xml(ref_path, test_path, date, count)
if File.exists?(ref_path) && File.exists?(test_path)
command = "diff -q #{test_path} #{ref_path}"
puts command
system(command)
if $? != 0
test = "regression_failed_text.xml"
ref = "regression_failed_ref.xml"
system("tidy -xml -utf8 -o #{test} #{test_path}")
system("tidy -xml -utf8 -o #{ref} #{ref_path}")
system("opendiff #{test} #{ref}")
puts "ERROR: #{test_path} and #{ref_path} don't match"
puts "Regression tests FAILED on date #{date} at count #{count}!"
# Give the user the option to overwrite the reference file and continue
puts "Press return to exit or 'o' to overwrite reference file and continue"
if gets == "o\n"
system("cp #{test_path} #{ref_path}")
else
exit
end
end
elsif File.exists?(ref_path)
puts "ERROR: #{test_path} is missing"
puts "Regression tests FAILED on date #{date} at count #{count}!"
exit
elsif File.exists?(test_path)
puts "ERROR: #{ref_path} is missing"
puts "Regression tests FAILED on date #{date} at count #{count}!"
exit
end
end
def test_date(date, conf, parser, count)
reps_xml_filename = "debates#{date}.xml"
senate_xml_filename = "daylord#{date}.xml"
new_reps_xml_path = "#{conf.xml_path}/scrapedxml/debates/#{reps_xml_filename}"
new_senate_xml_path = "#{conf.xml_path}/scrapedxml/lordspages/#{senate_xml_filename}"
ref_reps_xml_path = "#{File.dirname(__FILE__)}/../../ref/#{reps_xml_filename}"
ref_senate_xml_path = "#{File.dirname(__FILE__)}/../../ref/#{senate_xml_filename}"
parser.parse_date_house(date, new_reps_xml_path, House.representatives)
compare_xml(ref_reps_xml_path, new_reps_xml_path, date, count)
parser.parse_date_house(date, new_senate_xml_path, House.senate)
compare_xml(ref_senate_xml_path, new_senate_xml_path, date, count)
end
class Array
def randomly_permute
temp = clone
result = []
(1..size).each do
i = Kernel.rand(temp.size)
result << temp[i]
temp.delete_at(i)
end
result
end
end
# Randomly permute array. This means that we will cover a much broader range of dates quickly
srand(42)
dates = (from_date..to_date).to_a.randomly_permute
test_first.each do |date|
# Moves date to the beginning of the array
dates.delete(date)
dates.unshift(date)
end
skip_dates.each { |date| dates.delete(date) }
count = skip
time0 = Time.new
dates[skip..-1].each do |date|
test_date(date, conf, parser, count)
count = count + 1
puts "Regression test progress: Done #{count}/#{dates.size}"
seconds_left = ((Time.new - time0) / (count - skip) * (dates.size - count)).to_i
minutes_left = (seconds_left / 60).to_i
seconds_left = seconds_left - 60 * minutes_left
hours_left = (minutes_left / 60).to_i
minutes_left = minutes_left - 60 * hours_left
if hours_left > 0
puts "Estimated time left to completion: #{hours_left} hours #{minutes_left} mins"
else
puts "Estimated time left to completion: #{minutes_left} mins #{seconds_left} secs"
end
end
puts "Regression tests all passed!"
