pages.py

import sys
from os.path import exists
import subprocess
import re

# Calls the R system specifying that commands come from file commands.R
# The commands.R provided with this assignment will read the file named
# data and will output a histogram of that data to the file pageshist,pdf
def runR( ):
    res = subprocess.call(['R', '-f', 'commands.R'])

# log2hist analyzes a log file to calculate the total number of pages
# printed by each user during the period represented by this log file,
# and uses R to produce a pdf file pageshist.pdf showing a histogram
# of these totals.  logfilename is a string which is the name of the
# log file to analyze.
#
def log2hist(logfilename):
  log = open(logfilename)
  users = {}
  print_regex = re.compile('user:\s*(?P<user>.*)\s*printer:.*?pages:\s*(?P<pages>\d+)')

  for line in log:
    match = print_regex.search(line)

    if match:
      user = match.group('user').strip()
      pages = int(match.group('pages'))

      if user not in users:
        users[user] = pages
      else:
        users[user] += pages

  runR()

  data = open('data', 'w+')
  for pages in users.values():
    data.write("%d\n" % pages)

  data.close()

if __name__ == '__main__':
  script, logfilename = sys.argv

  if not exists(logfilename):
    print "%r does not exist" % logfilename
    sys.exit()

  log2hist(logfilename)