def parse(self, word):
"""Parse a word morphologically
e.g.
p = Parser()
word = ['p', 'a', 'n', 'i', 'c', 'k','e','d']
p.parse(word)
---> 'panic+past form'
"""
lexicon = {'panic', 'havoc', 'sync', 'lick', 'want'}
start_state = 'start'
k_insertion = 'k_insertion'
f = FST('parser')
f.add_state(start_state)
f.initial_state = start_state
#Add paths for each word
previous = start_state
for vocab in lexicon:
for char in vocab:
current = vocab + '-' + char #uniquely identify
f.add_state(current)
f.add_arc(previous, current, char, char)
previous = current
f.add_state(k_insertion + '-' + vocab)
f.add_arc(previous, k_insertion + '-' + vocab, 'k', '')
self._add_ending_states(f,
vocab,
k_insertion + '-' + vocab,
k=k_insertion)
self._add_ending_states(f, vocab, previous)
previous = start_state
return ''.join(f.transduce(word)[0])
def generate(self, analysis):
"""Generate the morphologically correct word
e.g.
p = Parser()
analysis = ['p','a','n','i','c','+past form']
p.generate(analysis)
---> 'panicked'
"""
# Let's define our first FST
f1 = FST('morphology-generate')
# Indicate initial and final states
f1.add_state('start')
f1.add_state('vowel')
f1.add_state('consonant')
f1.add_state('c')
f1.add_state('form_1')
f1.add_state('form_2')
f1.initial_state = 'start'
f1.set_final('form_1')
f1.set_final('form_2')
# Generate
vowels = 'aeiou'
for vowel in vowels:
f1.add_arc('start', 'vowel', vowel, vowel)
f1.add_arc('vowel', 'vowel', vowel, vowel)
f1.add_arc('consonant', 'vowel', vowel, vowel)
f1.add_arc('c', 'vowel', vowel, vowel)
for letter in string.ascii_lowercase:
if letter in vowels:
continue
if not letter == 'c':
f1.add_arc('vowel', 'consonant', letter, letter)
f1.add_arc('start', 'consonant', letter, letter)
f1.add_arc('consonant', 'consonant', letter, letter)
f1.add_arc('c', 'consonant', letter, letter)
f1.add_arc('vowel', 'c', 'c', 'c')
f1.add_arc('c', 'form_1', '+past form', 'ked')
f1.add_arc('c', 'form_1', '+present participle form', 'king')
f1.add_arc('consonant', 'form_2', '+past form', 'ed')
f1.add_arc('consonant', 'form_2', '+present participle form', 'ing')
output = f1.transduce(analysis)[0]
return "".join(output)
def generate(self, analysis):
"""Generate the morphologically correct word
e.g.
p = Parser()
analysis = ['p','a','n','i','c','+past form']
p.generate(analysis)
---> 'panicked'
"""
start_state = 'start'
f = FST('generator')
f.add_state(start_state)
f.initial_state = start_state
self._build_generator_fst(f, analysis, start_state)
return ''.join(f.transduce(analysis)[0])
def parse(self, word):
"""Parse a word morphologically
e.g.
p = Parser()
word = ['p','a','n','i','c','k','i','n','g']
p.parse(word)
---> 'panic+present participle form'
"""
# Ok so now let's do the second FST
f2 = FST('morphology-parse')
f2.add_state('start')
f2.initial_state = 'start'
#add states for the word lick
for w in list('lick'):
state_name = 'lick-' + w
f2.add_state(state_name)
#add first letter
f2.add_arc('start', 'lick-l', 'l', 'l')
#add arc for the word lick
lick = list('lick')
for w in range(0, len(lick) - 1):
f2.add_arc('lick-' + lick[w], 'lick-' + lick[w + 1], lick[w + 1],
lick[w + 1])
#add states for the word lick
for w in list('want'):
state_name = 'want-' + w
f2.add_state(state_name)
f2.add_arc('start', 'want-w', 'w', 'w')
#add arc for the word want
want = list('want')
for w in range(0, len(want) - 1):
f2.add_arc('want-' + want[w], 'want-' + want[w + 1], want[w + 1],
want[w + 1])
#add states for the word sync
sync = list('sync')
for w in sync:
state_name = 'sync-' + w
f2.add_state(state_name)
f2.add_arc('start', 'sync-s', 's', 's')
#add arc for the word sync
for w in range(0, len(sync) - 1):
f2.add_arc('sync-' + sync[w], 'sync-' + sync[w + 1], sync[w + 1],
sync[w + 1])
#add states for the word panic
panic = list('panic')
for w in panic:
state_name = 'panic-' + w
f2.add_state(state_name)
f2.add_arc('start', 'panic-p', 'p', 'p')
#add arc for the word panic
for w in range(0, len(panic) - 1):
f2.add_arc('panic-' + panic[w], 'panic-' + panic[w + 1],
panic[w + 1], panic[w + 1])
#add states for the word havoc
havoc = list('havoc')
for w in havoc:
state_name = 'havoc-' + w
f2.add_state(state_name)
f2.add_arc('start', 'havoc-h', 'h', 'h')
#add arc for the word havoc
for w in range(0, len(havoc) - 1):
f2.add_arc('havoc-' + havoc[w], 'havoc-' + havoc[w + 1],
havoc[w + 1], havoc[w + 1])
f2.add_state('intermediate1')
f2.add_state('intermediate2')
f2.add_state('pres1')
f2.add_state('past1')
f2.add_arc('lick-k', 'intermediate1', '', '')
f2.add_arc('want-t', 'intermediate1', '', '')
f2.add_arc('sync-c', 'intermediate1', '', '')
f2.add_arc('panic-c', 'intermediate1', 'k', '')
f2.add_arc('havoc-c', 'intermediate1', 'k', '')
f2.add_arc('intermediate1', 'pres1', 'ing', '+present participle form')
f2.add_arc('intermediate1', 'past1', 'ed', '+past form')
f2.set_final('pres1')
f2.set_final('past1')
if ''.join(word[-3:]) == 'ing':
inputs = word[:-3]
inputs.append('ing')
elif ''.join(word[-2:]) == 'ed':
inputs = word[:-2]
inputs.append('ed')
else:
inputs = word
output = f2.transduce(inputs)[0]
return ''.join(output)
def generate(self, analysis):
"""Generate the morphologically correct word
e.g.
p = Parser()
analysis = ['p','a','n','i','c','+past form']
p.generate(analysis)
---> 'panicked'
"""
#analysis = ['p','a','n','i','c','+past form']
# Let's define our first FST
f1 = FST('morphology-generate')
f1.add_state('1')
f1.add_state('2')
f1.add_state('3')
f1.add_state('4')
f1.add_state('5')
f1.add_state('6') #non-c state
f1.add_state('7') #c state
f1.add_state('8') #add k
f1.add_state('9') #+present
f1.add_state('10') #+past
f1.initial_state = '1'
#f1.set_final('8')
f1.set_final('9')
f1.set_final('10')
#state 1 to 2, and 2 to 3. we don't care about vowel or consonant here
for letter in list(string.ascii_letters):
f1.add_arc('1', '2', letter, letter)
f1.add_arc('2', '3', letter, letter)
#3 to 5 input/output consonants
vowels = ['a', 'e', 'i', 'o', 'u', 'A', 'E', 'I', 'O', 'U']
consonants = [c for c in list(string.ascii_letters) if c not in vowels]
non_c_con = [c for c in consonants if c not in ['c', 'C']]
for letter in consonants:
f1.add_arc('3', '5', letter, letter)
f1.add_arc('5', '5', letter, letter)
#the third and fourth input should be a vowel
for letter in vowels:
f1.add_arc('3', '4', letter, letter)
f1.add_arc('4', '4', letter, letter)
#if the fourth input is a non c consonant, go to 5
for letter in non_c_con:
f1.add_arc('4', '5', letter, letter)
#if the input at state 5 is a vowel, go back to 4
for letter in vowels:
f1.add_arc('5', '4', letter, letter)
#if the second last letter is a c, go to 7
f1.add_arc('4', '7', 'c', 'c')
#add k after 7
f1.add_arc('7', '8', '', 'k')
#output nothing from 5 to 8
f1.add_arc('5', '8', '', '')
f1.add_arc('8', '9', '+present participle form', 'ing')
f1.add_arc('8', '10', '+past form', 'ed')
output = f1.transduce(analysis)[0]
return ''.join(output)
Created on Fri Sep 08 11:21:21 2017
@author: SHILPASHREE RAO
"""
import string
import sys
from fst import FST
from fsmutils import trace
vowels = ['a', 'e', 'i', 'o', 'u']
f = FST('devowelizer')
f.add_state('1')
f.initial_state = '1'
f.set_final('1')
for letter in string.ascii_lowercase:
if letter in vowels:
_ = f.add_arc('1', '1', (letter), ())
else:
_ = f.add_arc('1', '1', (letter), (letter))
print ''.join(f.transduce(['v', 'o', 'w', 'e', 'l']))
print ''.join(f.transduce('e x c e p t i o n'.split()))
print ''.join(f.transduce('c o n s o n a n t'.split()))
t = trace(f, ['v', 'o', 'w', 'e', 'l'])
print t
