def application_matches(cls, word, applied_str, voicing_allowed):
"""
Checks if a suffix applied word is matched by a surface.
>>> Phonetics.application_matches(u'armudunu', u'armut', True)
True
>>> Phonetics.application_matches(u'armudunu', u'armut', False)
False
>>> Phonetics.application_matches(u'armudunu', u'armudu', True)
True
>>> Phonetics.application_matches(u'armudunu', u'armudu', False)
True
@param word: The full word (surface)
@param applied_str: Suffix applied part of the word
@param voicing_allowed: If voicing should be considered or ignored
@type word: unicode
@type applied_str: unicode
@type voicing_allowed: bool
@rtype: L{bool}
"""
if not applied_str or len(applied_str) > len(word):
return False
elif word == applied_str or word.startswith(applied_str):
return True
if voicing_allowed and word.startswith(applied_str[:-1]):
last_letter_of_application = TurkishAlphabet.get_letter_for_char(applied_str[-1])
last_letter_of_word_part = TurkishAlphabet.get_letter_for_char(word[len(applied_str) - 1])
return TurkishAlphabet.voice(last_letter_of_application) == last_letter_of_word_part
else:
return False
def _handle_phonetics(cls,
word,
phonetic_attributes,
form_str,
lexeme_attributes=None):
lexeme_attributes = lexeme_attributes or []
phonetic_attributes = phonetic_attributes or []
first_letter_of_form = TurkishAlphabet.get_letter_for_char(form_str[0])
# first apply voicing if possible
if LexemeAttribute.NoVoicing not in lexeme_attributes and PhoneticAttributes.LastLetterVoicelessStop in phonetic_attributes and first_letter_of_form.vowel:
voiced_letter = TurkishAlphabet.voice(
TurkishAlphabet.get_letter_for_char(word[-1]))
if voiced_letter:
word = word[:-1] + voiced_letter.char_value
# then try devoicing
if PhoneticAttributes.LastLetterVoiceless in phonetic_attributes and TurkishAlphabet.devoice(
first_letter_of_form):
form_str = TurkishAlphabet.devoice(
first_letter_of_form).char_value + form_str[1:]
applied = u''
for i in range(len(form_str)):
c = form_str[i]
next_c = form_str[i + 1] if i + 1 < len(form_str) else None
if c == '!':
continue
letter = TurkishAlphabet.get_letter_for_char(c)
if letter.vowel and letter.upper_case_char_value == c:
if c == u'A':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
applied += u'a'
else:
applied += u'e'
elif c == u'I':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
if PhoneticAttributes.LastVowelUnrounded in phonetic_attributes or next_c == '!':
applied += u'ı'
else:
applied += u'u'
else:
if PhoneticAttributes.LastVowelUnrounded in phonetic_attributes or next_c == '!':
applied += u'i'
else:
applied += u'ü'
elif c == u'O':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
applied += u'o'
else:
applied += u'ö'
else:
applied = applied + c
return word, applied
def _handle_phonetics(cls, word, phonetic_attributes, form_str, lexeme_attributes=None):
lexeme_attributes = lexeme_attributes or []
phonetic_attributes = phonetic_attributes or []
first_letter_of_form = TurkishAlphabet.get_letter_for_char(form_str[0])
# first apply voicing if possible
if LexemeAttribute.NoVoicing not in lexeme_attributes and PhoneticAttributes.LastLetterVoicelessStop in phonetic_attributes and first_letter_of_form.vowel:
voiced_letter = TurkishAlphabet.voice(TurkishAlphabet.get_letter_for_char(word[-1]))
if voiced_letter:
word = word[:-1] + voiced_letter.char_value
# then try devoicing
if PhoneticAttributes.LastLetterVoiceless in phonetic_attributes and TurkishAlphabet.devoice(first_letter_of_form):
form_str = TurkishAlphabet.devoice(first_letter_of_form).char_value + form_str[1:]
applied = u''
for i in range(len(form_str)):
c = form_str[i]
next_c = form_str[i + 1] if i + 1 < len(form_str) else None
if c == '!':
continue
letter = TurkishAlphabet.get_letter_for_char(c)
if letter.vowel and letter.upper_case_char_value == c:
if c == u'A':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
applied += u'a'
else:
applied += u'e'
elif c == u'I':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
if PhoneticAttributes.LastVowelUnrounded in phonetic_attributes or next_c == '!':
applied += u'ı'
else:
applied += u'u'
else:
if PhoneticAttributes.LastVowelUnrounded in phonetic_attributes or next_c == '!':
applied += u'i'
else:
applied += u'ü'
elif c == u'O':
if PhoneticAttributes.LastVowelBack in phonetic_attributes:
applied += u'o'
else:
applied += u'ö'
else:
applied = applied + c
return word, applied
def application_matches(cls, word, applied_str, voicing_allowed):
"""
Checks if a suffix applied word is matched by a surface.
>>> Phonetics.application_matches(u'armudunu', u'armut', True)
True
>>> Phonetics.application_matches(u'armudunu', u'armut', False)
False
>>> Phonetics.application_matches(u'armudunu', u'armudu', True)
True
>>> Phonetics.application_matches(u'armudunu', u'armudu', False)
True
@param word: The full word (surface)
@param applied_str: Suffix applied part of the word
@param voicing_allowed: If voicing should be considered or ignored
@type word: unicode
@type applied_str: unicode
@type voicing_allowed: bool
@rtype: L{bool}
"""
if not applied_str or len(applied_str) > len(word):
return False
elif word == applied_str or word.startswith(applied_str):
return True
if voicing_allowed and word.startswith(applied_str[:-1]):
last_letter_of_application = TurkishAlphabet.get_letter_for_char(
applied_str[-1])
last_letter_of_word_part = TurkishAlphabet.get_letter_for_char(
word[len(applied_str) - 1])
return TurkishAlphabet.voice(
last_letter_of_application) == last_letter_of_word_part
else:
return False
def _generate_modified_root_nodes(cls, lexeme):
if LexemeAttribute.RootChange in lexeme.attributes:
special_roots = cls._handle_special_roots(lexeme)
if special_roots:
return special_roots
modified_seq = lexeme.root
original_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(
lexeme.root)
modified_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(
lexeme.root)
original_phonetic_expectations = set()
modified_phonetic_expectations = set()
if LexemeAttribute.Voicing in lexeme.attributes or LexemeAttribute.VoicingOpt in lexeme.attributes:
last_letter = TurkishAlphabet.get_letter_for_char(modified_seq[-1])
modified_letter = TurkishAlphabet.voice(last_letter)
assert modified_letter is not None
if lexeme.lemma.endswith(u"nk"):
modified_letter = TurkishAlphabet.L_g
modified_seq = modified_seq[:-1] + modified_letter.char_value
if PhoneticAttributes.LastLetterVoicelessStop in modified_attributes:
modified_attributes.remove(
PhoneticAttributes.LastLetterVoicelessStop)
if modified_letter.continuant:
if PhoneticAttributes.LastLetterNotContinuant in modified_attributes:
modified_attributes.remove(
PhoneticAttributes.LastLetterNotContinuant)
modified_attributes.add(
PhoneticAttributes.LastLetterContinuant)
else:
if PhoneticAttributes.LastLetterContinuant in modified_attributes:
modified_attributes.remove(
PhoneticAttributes.LastLetterContinuant)
modified_attributes.add(
PhoneticAttributes.LastLetterNotContinuant)
if LexemeAttribute.VoicingOpt not in lexeme.attributes:
original_phonetic_expectations.add(
PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.Doubling in lexeme.attributes:
modified_seq = modified_seq + modified_seq[-1]
original_phonetic_expectations.add(
PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.LastVowelDrop in lexeme.attributes:
modified_seq = modified_seq[:-2] + modified_seq[-1]
if lexeme.syntactic_category != SyntacticCategory.VERB:
original_phonetic_expectations.add(
PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.InverseHarmony in lexeme.attributes:
original_attributes.add(PhoneticAttributes.LastVowelFrontal)
if PhoneticAttributes.LastVowelBack in original_attributes:
original_attributes.remove(PhoneticAttributes.LastVowelBack)
modified_attributes.add(PhoneticAttributes.LastVowelFrontal)
if PhoneticAttributes.LastVowelBack in modified_attributes:
modified_attributes.remove(PhoneticAttributes.LastVowelBack)
if LexemeAttribute.ProgressiveVowelDrop in lexeme.attributes:
modified_seq = modified_seq[:-1]
if RootGenerator._has_vowel(modified_seq):
modified_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(
modified_seq)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
original_phonetic_expectations = original_phonetic_expectations or None
modified_phonetic_expectations = modified_phonetic_expectations or None
original = Root(lexeme.root, lexeme, original_phonetic_expectations,
original_attributes)
modified = Root(modified_seq, lexeme, modified_phonetic_expectations,
modified_attributes)
if original == modified:
return [original]
else:
return [original, modified]
def _generate_modified_root_nodes(cls, lexeme):
if LexemeAttribute.RootChange in lexeme.attributes:
special_roots = cls._handle_special_roots(lexeme)
if special_roots:
return special_roots
modified_seq = lexeme.root
original_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(lexeme.root)
modified_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(lexeme.root)
original_phonetic_expectations = set()
modified_phonetic_expectations = set()
if LexemeAttribute.Voicing in lexeme.attributes or LexemeAttribute.VoicingOpt in lexeme.attributes:
last_letter = TurkishAlphabet.get_letter_for_char(modified_seq[-1])
modified_letter = TurkishAlphabet.voice(last_letter)
assert modified_letter is not None
if lexeme.lemma.endswith(u"nk"):
modified_letter = TurkishAlphabet.L_g
modified_seq = modified_seq[:-1] + modified_letter.char_value
if PhoneticAttributes.LastLetterVoicelessStop in modified_attributes:
modified_attributes.remove(PhoneticAttributes.LastLetterVoicelessStop)
if modified_letter.continuant:
if PhoneticAttributes.LastLetterNotContinuant in modified_attributes :
modified_attributes.remove(PhoneticAttributes.LastLetterNotContinuant)
modified_attributes.add(PhoneticAttributes.LastLetterContinuant)
else:
if PhoneticAttributes.LastLetterContinuant in modified_attributes:
modified_attributes.remove(PhoneticAttributes.LastLetterContinuant)
modified_attributes.add(PhoneticAttributes.LastLetterNotContinuant)
if LexemeAttribute.VoicingOpt not in lexeme.attributes:
original_phonetic_expectations.add(PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.Doubling in lexeme.attributes:
modified_seq = modified_seq + modified_seq[-1]
original_phonetic_expectations.add(PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.LastVowelDrop in lexeme.attributes:
modified_seq = modified_seq[:-2] + modified_seq[-1]
if lexeme.syntactic_category!=SyntacticCategory.VERB:
original_phonetic_expectations.add(PhoneticExpectation.ConsonantStart)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
if LexemeAttribute.InverseHarmony in lexeme.attributes:
original_attributes.add(PhoneticAttributes.LastVowelFrontal)
if PhoneticAttributes.LastVowelBack in original_attributes:
original_attributes.remove(PhoneticAttributes.LastVowelBack)
modified_attributes.add(PhoneticAttributes.LastVowelFrontal)
if PhoneticAttributes.LastVowelBack in modified_attributes:
modified_attributes.remove(PhoneticAttributes.LastVowelBack)
if LexemeAttribute.ProgressiveVowelDrop in lexeme.attributes:
modified_seq = modified_seq[:-1]
if RootGenerator._has_vowel(modified_seq):
modified_attributes = Phonetics.calculate_phonetic_attributes_of_plain_sequence(modified_seq)
modified_phonetic_expectations.add(PhoneticExpectation.VowelStart)
original_phonetic_expectations = original_phonetic_expectations or None
modified_phonetic_expectations = modified_phonetic_expectations or None
original = Root(lexeme.root, lexeme, original_phonetic_expectations, original_attributes)
modified = Root(modified_seq, lexeme, modified_phonetic_expectations, modified_attributes)
if original==modified:
return [original]
else:
return [original, modified]
