def learn_affixes(self, alignment, ft=features):
#------------------------------------------------------------------------------
def create_affix(aff, aff2):
if not re.match('_', aff):
if self.affixes.has_key(aff):
self.affixes[aff].count += 1
self.affixes[aff].partner += [aff2]
self.affixes[aff].add_enviro(r_edge, 'r', ft)
self.affixes[aff].add_enviro(l_edge, 'l', ft)
else:
self.affixes[aff] = Affix(aff, aff2)
self.affixes[aff].add_enviro(r_edge, 'r', ft)
self.affixes[aff].add_enviro(l_edge, 'l', ft)
# end definition of create_affix()
switch = False
aff1, aff2 = '', ''
for indx in range(len(alignment)):
if indx == 0:
previous = ('','')
else: # collecting enviro
previous = alignment[indx-1]
pair = alignment[indx]
if compare_features(pair[0], pair[1], ft=features) > 1:
nomatch1 = pair[0] not in pair[1]
nomatch2 = pair[1] not in pair[0]
if nomatch1 and nomatch2:
# we check if one string contains another
aff2 =''.join([aff2, pair[1]])
aff1 =''.join([aff1, pair[0]])
if not switch:
l_edge = previous[0]
switch = True
else:
if switch:
r_edge = pair[0]
create_affix(aff1, aff2)
create_affix(aff2, aff1)
aff1, aff2 = '', ''
switch = False
if indx == len(alignment)-1:
if switch:
r_edge = ''
create_affix(aff1, aff2)
create_affix(aff2, aff1)
def verify_affixes(self, alignment, features):
#------------------------------------------------------------------------------
def create_affix(aff):
if not re.match('_', aff):
return (aff, l_edge, r_edge)
#end create_affix() definition
#------------------------------------------------------------------------------
def logic_and(bool1, bool2):
if bool1 == True and bool2 == True:
return True
else:
return False
#end logic_and() definition
def test_env(lib, string):
print string
for indx in lib.keys():
if features[string][indx] != lib[indx]:
print indx, features[string][indx], lib[indx]
return False
else:
pass
return True
#------------------------------------------------------------------------------
def test_affix(aff):
if aff != None:
if self.affixes.has_key(aff[0]):
print 'affix exists'
affix = self.affixes[aff[0]]
right, left = False, False
if affix.suffix and aff[2] == '':
print 'suspect it\'s a suffix'
right = True
left = test_env(affix.on_left, aff[1])
elif affix.prefix and aff[1] == '':
left = True
right = test_env(affix.on_right, aff[1])
else:
right = test_env(affix.on_right, aff[0])
left = test_env(affix.on_left, aff[1])
return (aff, logic_and(right, left))
else:
return (aff, False)
else:
pass
#end test_affix() definition
answers = []
switch = False
aff1, aff2 = '', ''
for indx in range(len(alignment)):
pair = alignment[indx]
if indx == 0:
previous = ('','')
else:
previous = alignment[indx-1]
if compare_features(pair[0], pair[1], ft=features) > 1:
nomatch1 = pair[0] not in pair[1]
nomatch2 = pair[1] not in pair[0]
if nomatch1 and nomatch2:
aff2 =''.join([aff2, pair[1]])
aff1 =''.join([aff1, pair[0]])
if not switch:
l_edge = previous[0]
switch = True
else:
if switch:
r_edge = pair[0]
answers.append(test_affix(create_affix(aff1)))
answers.append(test_affix(create_affix(aff2)))
aff1, aff2 = '', ''
switch = False
if indx == len(alignment)-1 and switch:
r_edge = ''
answers.append(test_affix(create_affix(aff1)))
answers.append(test_affix(create_affix(aff2)))
return [x for x in answers if x != None]
