calendar.rb

# encoding: utf-8
require 'open-uri'
require 'pdf/reader'

io = open('http://www.nctu.edu.tw/campus/bulletin/Calendar/calendar2011.pdf')
reader = PDF::Reader.new(io)

sems = reader.pages.map do |page|
  sem = page.text.scan(
    /本學期自(\d+)年(\d+)月(\d+)日.+至(\d+)年(\d+)月(\d+)日/
  )[0]
  
  if sem
    sem.map(&:to_i)
  end
end

fall, spring = sems.select {|s| s}
p fall
p spring