Python port of @dancrew32's Ermagherd translator

ermagherd.py

# To use:
#
# erma = ermagerd("oh my god, a refrigerator1")

import re
class ermagerd:

	def __init__(self, arg1):
		print self.gherd(arg1)

	def translate(self, word):

		# Don't translate short words
		if len(word) == 1:
			return word

		# Handle specific words
		if word == 'AWESOME':
			return 'ERSUM'
		elif word == 'BANANA':
			return 'BERNERNER'
		elif word == 'BAYOU':
			return 'BERU'
		elif word == 'FAVORITE' or word == 'FAVOURITE':
			return 'FRAVRIT'
		elif word == 'GOOSEBUMPS':
			return 'GERSBERMS'
		elif word == 'LONG':
			return 'LERNG'
		elif word == 'MY':
			return 'MAH'
		elif word == 'THE':
			return 'DA'
		elif word == 'THEY':
			return 'DEY'
		elif word == "WE'RE":
			return 'WER'
		elif word == 'YOU':
			return 'U'
		elif word == "YOU'RE":
			return 'YER'

		# Before translating, keep a reference of the original word
		original = word

		# Drop vowel from the end of words
		if len(original) > 2: # Keep it at the end of short words like WE
			word = re.sub('[AEIOUY]$', '', word)

		# Reduce duplicate letters
		word = re.sub('[^\w\s]|(.)(?=\1)', '', word)

		# Reduce adjacent vowels to one
		word = re.sub('[AEIOUY]{2,}', 'E', word)

		# DOWN => DERN
		word = re.sub('OW', 'ER', word)

		# Replace vowels with ER
		word = re.sub('[AEIOUY]+', 'ER', word)

		# OH -> ER
		word = re.sub('ERH', 'ER', word)

		# MY -> MAH
		word = re.sub('MER', 'MAH', word)

		# FALLING -> FERLIN
		word = re.sub('ERNG', 'IN', word)

		# POOPED -> PERPERED -> PERPED
		word = re.sub('ERPERD', 'ERPED', word)

		# MEME -> MAHM -> MERM
		word = re.sub('MAHM', 'MERM', word)

		# Keep Y as first character
		# YES -> ERS -> YERS
		if original[0] == 'Y':
			word = 'Y' + word

		# reduce duplicate letters
		word = re.sub('[^\w\s]|(.)(?=\1)', '', word)

		# YELLOW -> YERLER -> YERLO
		if original[:3] == 'LOW' and word[:3] == 'LER':
			word = word[0:3] + 'LO'

		return word

	def gherd (self, text):

		if len(text) == 0:
			return

		text = text.upper()

		words = text.split(' ')

		translated = []

		for word in words:
			prefix = re.match('/^\W+/', word)
			suffix = re.match('/\W+$/', word)

			if word:
				# Is translateable
				new_word = ''
				if prefix:
					new_word = prefix
				new_word += self.translate(word)

				if suffix:
					new_word += suffix
				translated.append(new_word)
			else:
				# Is punctuation
				translated.append(word)


		return ' '.join(translated)

Inspired by @dancrew32's Ermagherd translator: https://gist.github.com/3185161