def test_french_sampa():
text = u'bonjour le monde'
backend = EspeakMbrolaBackend('mb-fr1')
sep = separator.Separator(word=None, phone=' ')
expected = 'b o~ Z u R l @ m o~ d '
out = backend.phonemize(text, separator=sep, strip=False)
assert out == expected
expected = 'b o~ Z u R l @ m o~ d'
out = backend.phonemize(text, separator=sep, strip=True)
assert out == expected
assert '' == backend.phonemize('', separator=sep, strip=True)
assert '' == backend.phonemize('"', separator=sep, strip=True)
def version():
"""Return version information for front and backends"""
# version of the phonemizer
_version = (
'phonemizer-' + pkg_resources.get_distribution('phonemizer').version)
# for each backend, check if it is available or not. If so get its version
available = []
unavailable = []
if EspeakBackend.is_available():
available.append(
'espeak-' + ('ng-' if EspeakBackend.is_espeak_ng() else '')
+ _version_as_str(EspeakBackend.version()))
else: # pragma: nocover
unavailable.append('espeak')
if EspeakMbrolaBackend.is_available():
available.append('espeak-mbrola')
else: # pragma: nocover
unavailable.append('espeak-mbrola')
if FestivalBackend.is_available():
available.append(
'festival-' + _version_as_str(FestivalBackend.version()))
else: # pragma: nocover
unavailable.append('festival')
if SegmentsBackend.is_available():
available.append(
'segments-' + _version_as_str(SegmentsBackend.version()))
else: # pragma: nocover
unavailable.append('segments')
# resumes the backends status in the final version string
if available:
_version += '\navailable backends: ' + ', '.join(available)
if unavailable: # pragma: nocover
_version += '\nuninstalled backends: ' + ', '.join(unavailable)
return _version
out = phonemize(' '.join(text),
language='en-us',
backend='espeak',
strip=False,
njobs=njobs)
assert out == ' '.join(['wʌn tuː', 'θɹiː', 'foːɹ faɪv '])
out = phonemize(os.linesep.join(text),
language='en-us',
backend='espeak',
strip=False,
njobs=njobs)
assert out == os.linesep.join(['wʌn tuː ', 'θɹiː ', 'foːɹ faɪv '])
@pytest.mark.skipif(not EspeakMbrolaBackend.is_available()
or not EspeakMbrolaBackend.is_supported_language('mb-fr1'),
reason='mbrola or mb-fr1 voice not installed')
@pytest.mark.parametrize('njobs', [2, 4])
def test_espeak_mbrola(caplog, njobs):
text = ['un deux', 'trois', 'quatre cinq']
out = phonemize(text,
language='mb-fr1',
backend='espeak-mbrola',
njobs=njobs,
preserve_punctuation=True)
assert out == ['9~d2', 'tRwa', 'katRse~k']
messages = [msg[2] for msg in caplog.record_tuples]
assert 'espeak-mbrola backend cannot preserve punctuation' in messages
def test_mbrola_bad_language():
assert not EspeakMbrolaBackend.is_supported_language('foo-bar')
def test_sampa_fr(text, expected):
assert expected == EspeakMbrolaBackend('mb-fr1').phonemize(
text, strip=True, separator=Separator(phone=''))
with pytest.raises(RuntimeError):
EspeakBackend.version()
os.environ['PHONEMIZER_ESPEAK_PATH'] = __file__
with pytest.raises(ValueError):
EspeakBackend.version()
finally:
try:
del os.environ['PHONEMIZER_ESPEAK_PATH']
except KeyError:
pass
@pytest.mark.skipif(
not EspeakMbrolaBackend.is_available() or
not EspeakMbrolaBackend.is_supported_language('mb-fr1'),
reason='mbrola or mb-fr1 voice not installed')
@pytest.mark.parametrize(
'text, expected',
[
# plosives
('pont', 'po~'),
('bon', 'bo~'),
('temps', 'ta~'),
('dans', 'da~'),
('quand', 'ka~'),
('gant', 'ga~'),
# fricatives
('femme', 'fam'),
('vent', 'va~'),
def recognize(self, wav=None):
if not self.eng.find("fest") == -1:
# ======= festival english us only =====================================
from phonemizer.backend import FestivalBackend
out1 = FestivalBackend(
'en-us', preserve_punctuation=False).phonemize(self.text,
strip=True)
elif not self.eng.find("esp") == -1:
# ======= espeak =======================================================
from phonemizer.backend import EspeakBackend
if not self.lan.find("en") == -1:
backend = EspeakBackend('en-us')
elif not self.lan.find("fr") == -1:
backend = EspeakBackend('fr-fr')
elif not self.lan.find("de") == -1 or not self.lan.find(
"ger") == -1:
backend = EspeakBackend('de-de')
elif not self.lan.find("ita") == -1:
backend = EspeakBackend('it-it')
elif not self.lan.find("esp") == -1 or not self.lan.find(
"spa") == -1:
backend = EspeakBackend('es-es')
else:
print("Invalid Language specified")
exit(1)
sep = separator.Separator(word=';eword ', syllable=None, phone=' ')
out1 = backend.phonemize(self.text, sep, False)
elif not self.eng.find("mbr") == -1:
# ====== mbrola ========================================================
from phonemizer.backend import EspeakBackend, EspeakMbrolaBackend
if EspeakMbrolaBackend.is_available():
if not self.lan.find("en") == -1:
EspeakMbrolaBackend('mb-en1')
elif not self.lan.find("fr") == -1:
EspeakMbrolaBackend('mb-fr2')
elif not self.lan.find("de") == -1 or not self.lan.find(
"ger") == -1:
EspeakMbrolaBackend('mb-de1')
elif not self.lan.find("ita") == -1:
EspeakMbrolaBackend('mb-es1')
elif not self.lan.find("esp") == -1 or not self.lan.find(
"spa") == -1:
EspeakMbrolaBackend('mb-it3')
else:
print("Invalid Language specified")
exit(1)
sep = separator.Separator(word=';eword ',
syllable=None,
phone=' ')
out1 = backend.phonemize(self.text, sep, False)
elif not self.eng.find("seg") == -1:
# ===== segment ========================================================
from phonemizer.backend import SegmentsBackend
if not self.lan.find("ja") == -1:
backend = SegmentsBackend('japanese')
else:
print("Invalid Language specified")
exit(1)
sep = separator.Separator(word=';eword ', syllable=None, phone=' ')
out1 = backend.phonemize(self.text, sep, False)
else:
print("Invalid Language specified")
exit(1)
print(out1) # return the result string to the speach engine
def backend():
return EspeakMbrolaBackend('mb-fr1')