def add_categorical_stanford_pcfg(context_skeleton, phoneme):
"""Adds the categorical elements of stanford parse information to a phoneme context."""
c = context_skeleton
#Add basic info
add_basic_stanford_pcfg(c, phoneme)
#Word
rw = phoneme.parent_word.get_next_word()
cw = phoneme.parent_word
lw = phoneme.parent_word.get_prev_word()
utt = phoneme.parent_utt
#Right word categorical position in parent, grandparent and greatgrandparent phrase
if rw == "xx":
c.add("rwcppp", "xx")
c.add("rwcgppp", "xx")
c.add("rwcggppp", "xx")
else:
c.add("rwcppp", get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("rwcgppp", get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("rwcggppp", get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
#Current word categorical position in parent, grandparent and greatgrandparent phrase
c.add("cwcppp", get_pos_cat(cw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("cwcgppp", get_pos_cat(cw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("cwcggppp", get_pos_cat(cw.pos_in_utt(), len(utt.words), with_sil=True))
#Left word categorical position in parent, grandparent and greatgrandparent phrase
if lw == "xx":
c.add("lwcppp", "xx")
c.add("lwcgppp", "xx")
c.add("lwcggppp", "xx")
else:
c.add("lwcppp", get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("lwcgppp", get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("lwcggppp", get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
def add_categorical_stanford_pcfg(context_skeleton, phoneme):
"""Adds the categorical elements of stanford parse information to a phoneme context."""
c = context_skeleton
#Add basic info
add_basic_stanford_pcfg(c, phoneme)
#Word
rw = phoneme.parent_word.get_next_word()
cw = phoneme.parent_word
lw = phoneme.parent_word.get_prev_word()
utt = phoneme.parent_utt
#Right word categorical position in parent, grandparent and greatgrandparent phrase
if rw == "xx":
c.add("rwcppp", "xx")
c.add("rwcgppp", "xx")
c.add("rwcggppp", "xx")
else:
c.add("rwcppp",
get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("rwcgppp",
get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("rwcggppp",
get_pos_cat(rw.pos_in_utt(), len(utt.words), with_sil=True))
#Current word categorical position in parent, grandparent and greatgrandparent phrase
c.add("cwcppp", get_pos_cat(cw.pos_in_utt(), len(utt.words),
with_sil=True))
c.add("cwcgppp", get_pos_cat(cw.pos_in_utt(),
len(utt.words),
with_sil=True))
c.add("cwcggppp",
get_pos_cat(cw.pos_in_utt(), len(utt.words), with_sil=True))
#Left word categorical position in parent, grandparent and greatgrandparent phrase
if lw == "xx":
c.add("lwcppp", "xx")
c.add("lwcgppp", "xx")
c.add("lwcggppp", "xx")
else:
c.add("lwcppp",
get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("lwcgppp",
get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
c.add("lwcggppp",
get_pos_cat(lw.pos_in_utt(), len(utt.words), with_sil=True))
def add_categorical(context_skeleton, phoneme):
"""Adds the categorical context set. Categorical means that positions in segments are categorical."""
utt = phoneme.parent_utt
phoneme_features = utt.phoneme_features
c = context_skeleton
syll = phoneme.parent_syllable
word = phoneme.parent_word
#Add the basic features
add_basic(c, phoneme)
#Get the R and L phonemes
p_pos = phoneme.pos_in_utt()
try:
rp = utt.phonemes[p_pos+1]
except IndexError:
rp = None
if phoneme.pos_in_utt() > 0:
lp = utt.phonemes[p_pos-1]
else:
lp = None
#Add current phoneme syll and word pos
#Phoneme pos in syll
if phoneme.id in phoneme_features.get_sil_phonemes():
c.add("cpsp", "xx")
c.add("cpwp", "xx")
else:
c.add("cpsp", get_pos_cat(phoneme.pos_in_syllable(), syll.num_phonemes()))
c.add("cpwp", get_pos_cat(phoneme.pos_in_word(), word.num_phonemes()))
#Add left phoneme syll and word pos
if lp == None or lp.id in phoneme_features.get_sil_phonemes():
c.add("lpsp", "xx")
c.add("lpwp", "xx")
else:
c.add("lpsp", get_pos_cat(lp.pos_in_syllable(), lp.parent_syllable.num_phonemes()))
c.add("lpwp", get_pos_cat(lp.pos_in_word(), lp.parent_word.num_phonemes()))
#Add right phoneme syll and word pos
if rp == None or rp.id in phoneme_features.get_sil_phonemes():
c.add("rpsp", "xx")
c.add("rpwp", "xx")
else:
c.add("rpsp", get_pos_cat(rp.pos_in_syllable(), rp.parent_syllable.num_phonemes()))
c.add("rpwp", get_pos_cat(rp.pos_in_word(), rp.parent_word.num_phonemes()))
#Add syllable word pos
s_pos = phoneme.parent_syllable.pos_in_utt()
try:
rs = utt.syllables[s_pos+1]
except IndexError:
rs = None
if syll.pos_in_utt() > 0:
ls = utt.syllables[s_pos-1]
else:
ls = None
#C Syllable pos in word
c.add("cswp", get_pos_cat(syll.pos_in_word(), syll.parent_word.num_syllables()))
#L Syllable pos in word
if ls == None:
c.add("lswp", "xx")
else:
c.add("lswp", get_pos_cat(ls.pos_in_word(), ls.parent_word.num_syllables()))
#R Syllable pos in word
if rs == None:
c.add("rswp", "xx")
else:
c.add("rswp", get_pos_cat(rs.pos_in_word(), rs.parent_word.num_syllables()))
#Add pos in utt
#Word pos in utterance
w_pos = phoneme.parent_word.pos_in_utt()
try:
rw = utt.words[w_pos+1]
except IndexError:
rw = None
if word.pos_in_utt() > 0:
lw = utt.words[w_pos-1]
else:
lw = None
#C Word pos in utt
c.add("cwup", get_pos_cat(word.pos_in_utt(), word.parent_utt.num_words(), True))
#L Word pos in utt
if lw == None:
c.add("lwup", "xx")
else:
c.add("lwup", get_pos_cat(lw.pos_in_utt(), lw.parent_utt.num_words(), True))
#R Word pos in utt
if rw == None:
c.add("rwup", "xx")
else:
c.add("rwup", get_pos_cat(rw.pos_in_utt(), rw.parent_utt.num_words(), True))
def add_categorical(context_skeleton, phoneme):
"""Adds the categorical context set. Categorical means that positions in segments are categorical."""
utt = phoneme.parent_utt
phoneme_features = utt.phoneme_features
c = context_skeleton
syll = phoneme.parent_syllable
word = phoneme.parent_word
#Add the basic features
add_basic(c, phoneme)
#Get the R and L phonemes
p_pos = phoneme.pos_in_utt()
try:
rp = utt.phonemes[p_pos + 1]
except IndexError:
rp = None
if phoneme.pos_in_utt() > 0:
lp = utt.phonemes[p_pos - 1]
else:
lp = None
#Add current phoneme syll and word pos
#Phoneme pos in syll
if phoneme.id in phoneme_features.get_sil_phonemes():
c.add("cpsp", "xx")
c.add("cpwp", "xx")
else:
c.add("cpsp",
get_pos_cat(phoneme.pos_in_syllable(), syll.num_phonemes()))
c.add("cpwp", get_pos_cat(phoneme.pos_in_word(), word.num_phonemes()))
#Add left phoneme syll and word pos
if lp == None or lp.id in phoneme_features.get_sil_phonemes():
c.add("lpsp", "xx")
c.add("lpwp", "xx")
else:
c.add(
"lpsp",
get_pos_cat(lp.pos_in_syllable(),
lp.parent_syllable.num_phonemes()))
c.add("lpwp",
get_pos_cat(lp.pos_in_word(), lp.parent_word.num_phonemes()))
#Add right phoneme syll and word pos
if rp == None or rp.id in phoneme_features.get_sil_phonemes():
c.add("rpsp", "xx")
c.add("rpwp", "xx")
else:
c.add(
"rpsp",
get_pos_cat(rp.pos_in_syllable(),
rp.parent_syllable.num_phonemes()))
c.add("rpwp",
get_pos_cat(rp.pos_in_word(), rp.parent_word.num_phonemes()))
#Add syllable word pos
s_pos = phoneme.parent_syllable.pos_in_utt()
try:
rs = utt.syllables[s_pos + 1]
except IndexError:
rs = None
if syll.pos_in_utt() > 0:
ls = utt.syllables[s_pos - 1]
else:
ls = None
#C Syllable pos in word
c.add("cswp",
get_pos_cat(syll.pos_in_word(), syll.parent_word.num_syllables()))
#L Syllable pos in word
if ls == None:
c.add("lswp", "xx")
else:
c.add("lswp",
get_pos_cat(ls.pos_in_word(), ls.parent_word.num_syllables()))
#R Syllable pos in word
if rs == None:
c.add("rswp", "xx")
else:
c.add("rswp",
get_pos_cat(rs.pos_in_word(), rs.parent_word.num_syllables()))
#Add pos in utt
#Word pos in utterance
w_pos = phoneme.parent_word.pos_in_utt()
try:
rw = utt.words[w_pos + 1]
except IndexError:
rw = None
if word.pos_in_utt() > 0:
lw = utt.words[w_pos - 1]
else:
lw = None
#C Word pos in utt
c.add("cwup",
get_pos_cat(word.pos_in_utt(), word.parent_utt.num_words(), True))
#L Word pos in utt
if lw == None:
c.add("lwup", "xx")
else:
c.add("lwup",
get_pos_cat(lw.pos_in_utt(), lw.parent_utt.num_words(), True))
#R Word pos in utt
if rw == None:
c.add("rwup", "xx")
else:
c.add("rwup",
get_pos_cat(rw.pos_in_utt(), rw.parent_utt.num_words(), True))
