def set_dict(self, dictfilename):
"""
Set the dictionary.
@param dictfilename (str- IN) The pronunciation dictionary in HTK-ASCII
format with UTF-8 encoding.
"""
pdict = DictPron(dictfilename, unkstamp=UNKSTAMP, nodump=False)
self.phonetizer = DictPhon( pdict, self.maptable )
def convert(self, tier):
"""
Phonetize all labels of a tier.
@param tier (Tier) contains the orthohraphic transcription (previously tokenized)
@return A tier named "Phonetization"
"""
t = Tier("Phonetization")
if tier is None: return t
for a in tier:
phon = ''
# Do not phonetize an unlabelled interval!
if not a.GetLabel().IsEmpty():
try:
# Do not phonetize silences
if a.GetLabel().IsSilence() is True:
phon = a.GetLabel().GetValue()
else:
_label = a.GetLabel().GetValue()
# Do not phonetize empty intervals!
if len(_label)>0:
phonetizer = DictPhon(self.pdict,self.logfile)
phon = phonetizer.phonetize( _label, self.opt_phonunk )
del phonetizer
except Exception as e:
raise Exception('Phonetization error of %s. Error: %s'%(a,str(e)))
new_phon = a.Copy()
new_phon.GetLabel().SetValue( phon )
t.Append(new_phon)
return t
class sppasPhon( sppasBase ):
"""
@author: Brigitte Bigi
@organization: Laboratoire Parole et Langage, Aix-en-Provence, France
@contact: [email protected]
@license: GPL, v3
@copyright: Copyright (C) 2011-2016 Brigitte Bigi
@summary: SPPAS integration of the Phonetization automatic annotation.
See DictPhon class for details about automatic phonetization.
How to use sppasPhon?
>>> p = sppasPhon( dict )
>>> p.run(inputtrsname, outputphonfile, outputtokensfile)
"""
def __init__(self, dictfilename, mapfile=None, logfile=None):
"""
Constructor.
@param dictfilename (str) is the pronunciation dictionary file name
(HTK-ASCII format, utf8).
@param mapfile (str) is the filename of a mapping table. It is used
to generate new pronunciations by mapping phonemes of the dictionary.
@param logfile (sppasLog) is a log file utility class member.
"""
sppasBase.__init__(self, logfile)
# Pronunciation dictionary
self.maptable = None
if mapfile is not None:
self.maptable = Mapping( mapfile )
self.set_dict( dictfilename )
# List of options to configure this automatic annotation
self._options = {}
self._options['phonunk'] = False # Phonetize missing tokens
self._options['usestdtokens'] = False # Phonetize standard spelling
# -----------------------------------------------------------------------
# Methods to fix options
# -----------------------------------------------------------------------
def fix_options(self, options):
"""
Fix all options.
Available options are:
- unk
- usesstdtokens
@param options (option)
"""
for opt in options:
key = opt.get_key()
if key == "unk":
self.set_unk( opt.get_value() )
elif key == "usestdtokens":
self.set_usestdtokens( opt.get_value() )
else:
raise Exception('Unknown key option: %s'%key)
# -----------------------------------------------------------------------
def set_unk(self, unk):
"""
Fix the unk option value.
@param unk (bool - IN) If unk is set to True, the system will attempt
to phonetize unknown entries (i.e. tokens missing in the dictionary).
Otherwise, the phonetization of an unknown entry unit is set to the
default string.
"""
self._options['phonunk'] = unk
# -----------------------------------------------------------------------
def set_usestdtokens(self,stdtokens):
"""
Fix the stdtokens option.
@param stdtokens (bool - IN) If it is set to True, the phonetization
will use the standard transcription as input, instead of the faked
transcription. This option does make sense only for an Enriched
Orthographic Transcription.
"""
self._options['usestdtokens'] = stdtokens
# -----------------------------------------------------------------------
# Methods to phonetize series of data
# -----------------------------------------------------------------------
def set_dict(self, dictfilename):
"""
Set the dictionary.
@param dictfilename (str- IN) The pronunciation dictionary in HTK-ASCII
format with UTF-8 encoding.
"""
pdict = DictPron(dictfilename, unkstamp=UNKSTAMP, nodump=False)
self.phonetizer = DictPhon( pdict, self.maptable )
# -----------------------------------------------------------------------
def phonetize(self, entry):
"""
Phonetize a text.
Because we absolutely need to match with the number of tokens, this
method will always return a string: either the automatic phonetization
(from dict or from phonunk) or the UNKSTAMP.
@param entry (str- IN) The string to be phonetized.
@return phonetization of `label`.
"""
tab = self.phonetizer.get_phon_tokens( entry.split(), phonunk=self._options['phonunk'])
tabphon = []
for t,p,s in tab:
message = None
if s == ERROR_ID:
message = "%s is missing of the dictionary and wasn't phonetized."%t
return UNKSTAMP
else:
if s == WARNING_ID:
message = "%s is missing of the dictionary, and "%t
if len(p) > 0:
message = message +"was automatically phonetized as: %s"%p
else:
message = message +"wasn't phonetized."
p = UNKSTAMP
tabphon.append( p )
if message:
self.print_message(message, indent=3, status=s)
return " ".join(tabphon)
# -----------------------------------------------------------------------
def convert(self, tier, variants=True):
"""
Phonetize all annotation of a tokenized tier.
@param tier (Tier - IN) contains the orthographic transcription
previously tokenized.
@param variants (bool - IN)
@return A tier with name "Phonetization"
"""
t = Tier("Phonetization")
if tier is None:
return t
for a in tier:
af = a.Copy()
for text in af.GetLabel().GetLabels():
if text.IsPause() is True:
# In case the pronunciation dictionary were not properly fixed.
text.SetValue( "sil" )
elif text.IsEmpty() is False and text.IsSilence() is False:
phon = self.phonetize( text.GetValue() )
text.SetValue( phon )
t.Append( af )
return t
# -----------------------------------------------------------------------
def get_input_tier(self, trsinput):
"""
Return the tier to be phonetized.
@param trsinput (Transcription)
@return Tier
"""
tierinput = None
pattern = "fake"
if self._options['usestdtokens'] is True: # Phonetize standard spelling
pattern = "standard"
for tier in trsinput:
if "tok" in tier.GetName().lower() and pattern in tier.GetName().lower():
tierinput = tier
break
if tierinput is None:
for tier in trsinput:
if "tok" in tier.GetName().lower():
tierinput = tier
break
if tierinput is None:
for tier in trsinput:
if "trans" in tier.GetName().lower():
tierinput = tier
break
return tierinput
# ------------------------------------------------------------------------
def run( self, inputfilename, outputfile ):
"""
Run the Phonetization process on an input file.
@param inputfilename (str - IN) the input file name with tokenization
@param outputfile (str - IN) the output file name of the phonetization
"""
self.print_options()
self.print_diagnosis(inputfilename)
# Get the tier to be phonetized.
trsinput = annotationdata.io.read( inputfilename )
tierinput = self.get_input_tier(trsinput)
if tierinput is None:
raise Exception("No tier found with tokenization. "
"One of the tier names must contain 'tok' (or 'trans').")
# Phonetize the tier
tierphon = self.convert( tierinput )
# Save
trsoutput = Transcription("sppasPhon")
trsoutput.Append( tierphon )
# Save in a file
annotationdata.io.write( outputfile,trsoutput )
