def test_instantiate(self):
assert Metaspace() is not None
assert Metaspace(replacement="-") is not None
with pytest.raises(Exception, match="replacement must be a character"):
Metaspace(replacement="")
assert Metaspace(add_prefix_space=True) is not None
assert isinstance(Metaspace(), PreTokenizer)
def test_can_modify(self):
pretok = Metaspace(replacement="$", add_prefix_space=False)
assert pretok.replacement == "$"
assert pretok.add_prefix_space == False
# Modify these
pretok.replacement = "%"
assert pretok.replacement == "%"
pretok.add_prefix_space = True
assert pretok.add_prefix_space == True
def test_instantiate(self):
assert Metaspace() is not None
assert Metaspace(replacement="-") is not None
with pytest.raises(ValueError, match="expected a string of length 1"):
Metaspace(replacement="")
assert Metaspace(add_prefix_space=True) is not None
assert isinstance(Metaspace(), PreTokenizer)
assert isinstance(Metaspace(), Metaspace)
assert isinstance(pickle.loads(pickle.dumps(Metaspace())), Metaspace)
def converted(self):
tokenizer = self.tokenizer(self.proto)
# Tokenizer assemble
tokenizer.normalizer = self.normalizer(self.proto)
replacement = "▁"
add_prefix_space = True
tokenizer.pre_tokenizer = Metaspace(replacement=replacement,
add_prefix_space=add_prefix_space)
tokenizer.decoder = decoders.Metaspace(
replacement=replacement, add_prefix_space=add_prefix_space)
post_processor = self.post_processor(tokenizer)
if post_processor:
tokenizer.post_processor = post_processor
# TODO what parameters should we give ?
parameters = {}
return BaseTokenizer(tokenizer, parameters)