def test_init_from_accents_adds_no_accents(self):
res = AccentsDict.init_from_accents({"a", "b", "c"})
self.assertEqual(3, len(res))
def test_ds_dataset_to_merged_dataset(self):
ds = DsDataset(
name="ds",
accent_ids=AccentsDict.init_from_accents({"c", "d"}),
data=DsDataList([
DsData(
entry_id=1,
basename="basename",
speaker_id=2,
serialized_symbols="0,1",
gender=Gender.MALE,
lang=Language.GER,
serialized_accents="0,0",
wav_path="wav",
speaker_name="speaker",
text="text",
)
]),
mels=MelDataList(
[MelData(
entry_id=1,
mel_path="mel",
n_mel_channels=5,
)]),
speakers=SpeakersDict.fromlist(["sp1", "sp2", "sp3"]),
symbol_ids=SymbolIdDict.init_from_symbols({"a", "b"}),
texts=TextDataList([
TextData(
entry_id=1,
lang=Language.IPA,
serialized_accent_ids="1,0",
serialized_symbol_ids="1,1",
text="text_new",
)
]),
wavs=WavDataList(
[WavData(
entry_id=1,
duration=15,
sr=22050,
wav="wav_new",
)]),
)
res = ds_dataset_to_merged_dataset(ds)
self.assertEqual("ds", res.name)
self.assertEqual(2, len(res.accent_ids))
self.assertEqual(3, len(res.speakers))
self.assertEqual(2, len(res.symbol_ids))
self.assertEqual(1, len(res.items()))
first_entry = res.items()[0]
self.assertEqual(1, first_entry.entry_id)
self.assertEqual(Gender.MALE, first_entry.gender)
self.assertEqual("basename", first_entry.basename)
self.assertEqual(2, first_entry.speaker_id)
self.assertEqual(Language.IPA, first_entry.lang)
self.assertEqual("1,0", first_entry.serialized_accents)
self.assertEqual("1,1", first_entry.serialized_symbols)
self.assertEqual("wav_new", first_entry.wav_path)
self.assertEqual(15, first_entry.duration)
self.assertEqual(22050, first_entry.sampling_rate)
self.assertEqual("mel", first_entry.mel_path)
self.assertEqual(5, first_entry.n_mel_channels)
def test_init_from_accents_is_sorted(self):
res = AccentsDict.init_from_accents({"c", "a", "b"})
self.assertEqual("a", res.get_accent(0))
self.assertEqual("b", res.get_accent(1))
self.assertEqual("c", res.get_accent(2))
def _get_all_accents(l: PreDataList) -> AccentsDict:
accents = set()
for x in l.items():
accents = accents.union(set(x.accents))
return AccentsDict.init_from_accents(accents)