#!/usr/bin/env python
"""Parse HTML files and encode email addresses to hide them from spammers.

This program uses numerical HTML entities, either in decimal (default) or
hexadecimal encoding.

Author: Andreas Neudecker
Created: 2003-07-30
Changed: 2003-10-01
Version: 0.0.3

Copyright (c) 2003 by Andreas Neudecker.
Licensed under the GNU General Public License (GPL),
see http://www.gnu.org/copyleft/ for details.

"""

# Changelog:
# (Add change notes chronologially. Include name, email, date)
# [2003-09-28] Added terrific new feature. (E. Xample, e.xample@some-email.net)
# [2003-10-01] Much better regular expression for finding emails, thanks to 
#              Michael Wurzel; more minor changes. (Andreas, zapyon@gmx.net)

# Comment conventions:
# [???] = Problem to be solved
# [!!!] = Needs testing/editing
# [2do] = Needs to be implemented/changed/optimised
# You should add a usefull explanation.

# Execute always --------------------------------------------------------------#

# Use Python JIT Psyco to speed up execution. (http://psyco.sourceforge.net/)
#
# Needs to be imported BEFORE other modules to allow them to be optimised at
# runtime. Makes most sense to import and run Psyco at the start of your
# program's main module!
#   Set 'usePsyco' to false while testing. Leave it on for finished code.
# It is NOT necessary to switch off (False) on systems without Psyco installed.
usePsyco = False
# set to true, if you want to use it AND have verbosity.py
#useVerbosity = True

if usePsyco and __name__ == '__main__':
	try:
		import psyco
		psyco.profile()
	except:
		print 'Psyco not found, ignoring it'



# Modules ---------------------------------------------------------------------#
#from optik import *
#import string
import re
import random as R


# Globals ---------------------------------------------------------------------#
verboseTesting = True

# Classes ---------------------------------------------------------------------#

# Functions -------------------------------------------------------------------#

def obfuscate (string, mode = 'decimal'):
	"""Encode string using numeric HTML entities.

	Possible encoding modes for the entities are:
	    'decimal',     'dec', 'd': decimal encoding (default)
	    'hexadecimal', 'hex', 'h': hexadecimal encoding
	    'mixed',       'mix', 'm': mixed hex and dec encoding
	    'none',               'n': do not encode
	"""

	modeUse = {
		'decimal'     : decHtmlEnt,
		'dec'         : decHtmlEnt,
		'd'           : decHtmlEnt,
		'hexadecimal' : hexHtmlEnt,
		'hex'         : hexHtmlEnt,
		'h'           : hexHtmlEnt,
		'mixed'       : mixHtmlEnt,
		'mix'         : mixHtmlEnt,
		'm'           : mixHtmlEnt,
	}

	modeKeys = modeUse.keys ()

	if mode in modeKeys:
		return ''.join (map (modeUse [mode], string))
	else: # wrong mode or "don't encode" selected.
		return string


def decHtmlEnt (string):
	"""Encode letters in 'string' as decimal HTML entities &<nn>;"""
	return ''.join ( ['&#' + str (ord (ch)) + ';' for ch in string] )


def hexHtmlEnt (string):
	"""Encode letters in 'string' as hexadecimal HTML entities &<nn>;"""
	return ''.join ( ['&#' + hex (ord (ch)) [1:] + ';' for ch in string] )


def mixHtmlEnt (string):
	"""Encode 'string' as random mix of hex and decimal encoded HTML entities."""
	return ''.join ( [ (posRandInt (1) and [decHtmlEnt (ch)] or [hexHtmlEnt (ch)])[0] for ch in string ] )


def noEncoding (string):
	"""Do not manipulate 'string'."""
	return string


def posRandInt (n, count = None):
	"""Pick a random positive integer in the range of [0, n].

	If count is given, return a list of 'count' random positive integers.
	Else, return just one.
	"""
	if count:
		return [int (round (R.random () * n)) for i in range (count)]
	else:
		return int (round (R.random () * n))


def encLine (line, mode = 'dec'):
	"""Encode email addresses and URIs in a line as desired.

		line:   a 'string', usually one line of a HTML file.
		mode:   the email address string is encoded using numerical HTML
		        entities ('&[#]<nn[nn]>;'). Allowed values are:
		            'decimal',     'dec', 'd': decimal encoding (default)
		            'hexadecimal', 'hex', 'h': hexadecimal encoding
		            'mixed',       'mix', 'm': mixed hex and dec encoding
		        For all other values, including None the original string
		        is returned.
	"""
	# Prepare regular expressions
	charClass = r"[a-zA-Z_\-.0-9]"
	email = re.compile (r"(%s * @ %s * ) | mailto:" % (charClass, charClass), re.VERBOSE)

	return replExpr (email, line, mode)


def replExpr (regx, line, mode):
	"""Use regular expression 'regx' to perform replacements in line."""
	def repl(mtch):
		return obfuscate(mtch.group(0),mode)
	return regx.subn(repl,line)[0]


# main () ---------------------------------------------------------------------#
def __main ():
	a = '''
	a.aa@bbb.ccc, def
		(<a href="mailto:ddd.ddd@eee.ff">ddd.ddd@eee.ff</a>),
	ghi (g@h.i)
	'''

	print encLine (a)

if __name__ == '__main__':
	__main ()

# EOF -------------------------------------------------------------------------#