def __init__(
self,
examples: List[QAExample],
tokenizer: SentencePieceBPETokenizer,
max_sequence_length: int,
is_train: bool = True,
) -> None:
self.examples = examples
self.tokenizer = tokenizer
self.max_sequence_length = max_sequence_length
self.sos_token = tokenizer.token_to_id("<s>")
self.eos_token = tokenizer.token_to_id("</s>")
self.question_prefix_tokens = self.tokenizer.encode("질문:").ids
self.is_train = is_train
def load_tokenizer(path,
enable_truncation=True,
enable_padding=True,
max_length=512):
tokenizer = SentencePieceBPETokenizer(os.path.join(path, "vocab.json"),
os.path.join(path, "merges.txt"))
tokenizer._tokenizer.post_processor = BertProcessing(
("</s>", tokenizer.token_to_id("</s>")),
("<s>", tokenizer.token_to_id("<s>")),
)
if enable_truncation:
tokenizer.enable_truncation(max_length=max_length)
if enable_padding:
tokenizer.enable_padding(pad_token="<pad>",
pad_id=tokenizer.token_to_id("<pad>"))
return tokenizer
class TokenizerWrapper:
def __init__(self, tok_type, unk_token, sep_token, cls_token, pad_token,
mask_token):
self.tok_type = tok_type
if self.tok_type == 'bpe':
self.tokenizer = ByteLevelBPETokenizer()
elif self.tok_type == 'wordpiece':
self.tokenizer = BertWordPieceTokenizer(unk_token=unk_token,
sep_token=sep_token,
cls_token=cls_token,
pad_token=pad_token,
mask_token=mask_token)
elif self.tok_type == 'sentencepiece':
self.tokenizer = SentencePieceBPETokenizer(unk_token=unk_token)
def train(self, data_file, vocab_size, special_tokens):
if self.tok_type in ['bpe', 'wordpiece', 'sentencepiece']:
self.tokenizer.train([data_file],
vocab_size=vocab_size,
special_tokens=special_tokens)
def tokenize(self, text):
if self.tok_type in ['bpe', 'wordpiece', 'sentencepiece']:
return self.tokenizer.encode(text).tokens
elif self.tok_type == 'word':
return nltk.tokenize.word_tokenize(text)
elif self.tok_type == 'char':
return [c for c in text]
else:
raise Exception('Unknown tokenizer: ' + self.tok_type)
def decode(self, tokens, blank_token):
if self.tok_type in ['bpe', 'wordpiece', 'sentencepiece']:
ids = [self.tokenizer.token_to_id(t) for t in tokens]
ids = [
i if i != None else self.tokenizer.token_to_id(blank_token)
for i in ids
]
return self.tokenizer.decode(ids, skip_special_tokens=False)
elif self.tok_type == 'word':
return ' '.join(tokens)
elif self.tok_type == 'char':
return ''.join(tokens)
else:
raise Exception('Unknown tokenizer: ' + self.tok_type)
class DecodeBySentencePieceBPETokenizer(Preprocessor):
__provider__ = 'decode_by_sentence_piece_bpe_tokenizer'
@classmethod
def parameters(cls):
parameters = super().parameters()
parameters.update({
'vocabulary_file':
PathField(),
'merges_file':
PathField(),
'sos_symbol':
StringField(optional=True, default='<s>'),
'eos_symbol':
StringField(optional=True, default='</s>'),
'add_extra_symbols':
BoolField(optional=True, default=True),
})
return parameters
def configure(self):
if isinstance(SentencePieceBPETokenizer, UnsupportedPackage):
SentencePieceBPETokenizer.raise_error(self.__provider__)
self.tokenizer = SentencePieceBPETokenizer(
str(self.get_value_from_config('vocabulary_file')),
str(self.get_value_from_config('merges_file')))
self.add_extra_symbols = self.get_value_from_config(
'add_extra_symbols')
self.idx = {}
for s in ['sos', 'eos']:
self.idx[s] = self.tokenizer.token_to_id(
str(self.get_value_from_config(s + '_symbol')))
def process(self, image, annotation_meta=None):
sentence = " ".join(image.data)
tokens = self.tokenizer.encode(sentence).ids
if self.add_extra_symbols:
tokens = [self.idx['sos']] + tokens + [self.idx['eos']]
image.data = tokens
image.metadata['decoded'] = True
image.identifier = "tokens"
return image
class Tokenizer:
""" Sentence tokenizer.
Arguments:
path (str): path to tokenizer's model folder.
max_tokens (int): max tokens.
"""
def __init__(self, path, max_tokens):
self.logger = log.getLogger("Tokenizer")
self.logger.info("loading tokenizer")
self.logger.info("path: " + path)
self.logger.info("max_tokens: " + str(max_tokens))
self.tokenizer = SentencePieceBPETokenizer(
os.path.join(path, "vocab.json"),
os.path.join(path, "merges.txt")
)
self.max_tokens = max_tokens
self.idx = {}
for s in ['</s>', '<s>', '<pad>']:
self.idx[s] = self.tokenizer.token_to_id(s)
def encode(self, sentence):
""" Encode method for sentence.
Arguments:
sentence (str): sentence.
Returns:
tokens (np.array): encoded sentence in tokneized format.
"""
tokens = self.tokenizer.encode(sentence).ids
return self._extend_tokens(tokens)
def decode(self, tokens, remove_repeats=True):
""" Decode method for tokens.
Arguments:
tokens (np.array): sentence in tokenized format.
remove_repeats (bool): remove repeated words.
Returns:
sentence (str): output sentence.
"""
sentence = self.tokenizer.decode(tokens)
for s in self.idx.keys():
sentence = sentence.replace(s, '')
if remove_repeats:
sentence = self._remove_repeats(sentence)
return sentence.lstrip()
def _extend_tokens(self, tokens):
""" Extend tokens.
Arguments:
tokens (np.array): sentence in tokenized format.
Returns:
tokens (np.array): extended tokens.
"""
tokens = [self.idx['<s>']] + tokens + [self.idx['</s>']]
pad_length = self.max_tokens - len(tokens)
if pad_length > 0:
tokens = tokens + [self.idx['<pad>']] * pad_length
return tokens
def _remove_repeats(self, sentence):
""" Remove repeated words.
Arguments:
sentence (str): sentence.
Returns:
sentence (str): sentence in lowercase without repeated words.
"""
tokens = sentence.lower().split()
return " ".join(key for key, _ in itertools.groupby(tokens))
class BPEVocabulary(Vocabulary):
""" Represents a SentencePiece vocabulary for c2s.
"""
def __init__(self, args: Namespace):
super().__init__()
self.target_encoder = SentencePieceBPETokenizer(
args.target_vocab, args.target_merges)
self.subtoken_encoder = SentencePieceBPETokenizer(
args.subtoken_vocab, args.subtoken_merges)
# self.target_encoder.add_special_tokens(
# [self.EOS_TOKEN, self.SOS_TOKEN, self.PAD_TOKEN]
# )
# self.subtoken_encoder.add_special_tokens([self.EOS_TOKEN, self.PAD_TOKEN])
with open(args.node_dict, "rb") as f:
self.node_to_index = pickle.load(f)
self.index_to_node = {v: k for k, v in self.node_to_index.items()}
def target_vocab_size(self):
# print(self.target_encoder.num_special_tokens_to_add())
return self.target_encoder.get_vocab_size() + 4
def node_vocab_size(self):
# print(self.target_encoder.num_special_tokens_to_add())
return len(self.node_to_index) + 2
def terminal_vocab_size(self):
return self.subtoken_encoder.get_vocab_size() + 4
def add_special_target_token(self, token: str):
self.target_encoder.add_special_tokens([token])
def add_special_terminal_token(self, token: str):
self.subtoken_encoder.add_special_tokens([token])
def encode_node(self, token_or_tokens):
if isinstance(token_or_tokens, str):
return self.node_to_index.get(token_or_tokens,
self.node_to_index[self.UNK_TOKEN])
else:
return list(map(self.encode_node, token_or_tokens))
def decode_node(self, index_or_indices):
if isinstance(index_or_indices, int):
return self.index_to_node[index_or_indices]
else:
return list(map(self.decode_node, index_or_indices))
def encode_target(self, token_or_tokens):
if isinstance(token_or_tokens, str):
return self.target_encoder.token_to_id(token_or_tokens)
else:
return self.target_encoder.encode(" ".join(token_or_tokens)).ids
def decode_target(self, index_or_indices):
if isinstance(index_or_indices, int):
return self.target_encoder.id_to_token(index_or_indices)
else:
return self.target_encoder.decode(index_or_indices)
def encode_terminal(self, token_or_tokens):
if isinstance(token_or_tokens, str):
return self.subtoken_encoder.token_to_id(token_or_tokens)
else:
return self.subtoken_encoder.encode(" ".join(token_or_tokens)).ids
def decode_terminal(self, index_or_indices):
if isinstance(index_or_indices, int):
return self.terminal_encoder.id_to_token(index_or_indices)
else:
return self.terminal_encoder.decode(index_or_indices)
class TextProcessor:
def __init__(self, tok_model_path: Optional[str] = None):
self.languages = {}
if tok_model_path is not None:
self.tokenizer = SentencePieceBPETokenizer(
tok_model_path + "/vocab.json",
tok_model_path + "/merges.txt",
)
with open(os.path.join(tok_model_path, "langs"), "rb") as fp:
self.languages: Dict[str, int] = pickle.load(fp)
self.init_properties(self.languages)
def init_properties(self, languages: Dict[str, int] = None):
self.max_len = 512
self.pad_token = "<pad>"
self.mask_token = "<mask>"
self.unk_token = "<unk>"
self.sep_token = "</s>"
self.bos = "<s>"
self.special_tokens = [
self.pad_token, self.bos, self.unk_token, self.mask_token,
self.sep_token
] + list(languages.keys())
self.languages = languages
def train_tokenizer(self, paths: List[str], vocab_size: int,
to_save_dir: str, languages: Dict[str, int]):
self.tokenizer = SentencePieceBPETokenizer()
self.init_properties(languages)
self.tokenizer.train(files=paths,
vocab_size=vocab_size,
min_frequency=5,
special_tokens=self.special_tokens)
self.save(directory=to_save_dir)
def _tokenize(self, line) -> Encoding:
return self.tokenizer.encode(line)
def save(self, directory):
self.tokenizer.save(directory)
with open(os.path.join(directory, "langs"), "wb") as fp:
pickle.dump(self.languages, fp)
def tokenize_one_line(self,
line,
ignore_middle_eos: bool = False) -> List[int]:
tokenized = []
spl = [sen for sen in line.split("</s>") if len(sen.strip()) > 0]
if spl[0].startswith("<"):
words = spl[0].strip().split(" ")
spl[0] = " ".join(words[1:])
tokenized += [self.token_id(words[0])]
for sen in spl:
tokenized += self._tokenize(sen).ids
if not ignore_middle_eos:
tokenized += [self.sep_token_id()]
if ignore_middle_eos:
tokenized += [self.sep_token_id()]
return tokenized
def tokenize_one_sentence(self, line) -> List[int]:
"""
Assume that the sentence has language id in the first token and end of sentence as the end!
:param line:
:return:
"""
spl = line.strip().split(" ")
lang_id, sen, eos = spl[0], " ".join(spl[1:-1]), spl[-1]
tokenized = [self.token_id(lang_id)
] + self._tokenize(sen).ids + [self.token_id(eos)]
return tokenized
def tokenize_lines(self,
line,
blind_split: bool = False,
split_len: int = 512) -> List[List[int]]:
"""
:param line:
:param blind_split: If True, just splits the tokenized data into chunks without considering that every vector
should start with a first word in sentence.
:return:
"""
tokenized = []
if len(self.languages) > 0:
spl = [sen for sen in line.split("</s>") if len(sen.strip()) > 0]
lang_id = []
if spl[0].startswith("<"):
words = spl[0].strip().split(" ")
lang_id = [self.token_id(words[0])]
spl[0] = " ".join(words[1:])
max_len = 0
for sen in spl:
toks = self._tokenize(sen).ids
tokenized += lang_id + toks + [self.sep_token_id()]
max_len = max(max_len, len(toks) + 1)
else:
tokenized = self._tokenize(line.strip()).ids
if blind_split:
num_pads = (split_len - (len(tokenized) % split_len))
pad_arr = [self.pad_token_id()] * num_pads
tokenized = np.array(tokenized + pad_arr)
reshaped = tokenized.reshape((-1, split_len))
return reshaped
else:
return self.split_tokenized(tokenized, min(max_len, self.max_len))
def tokenize(self, lines) -> List[List[int]]:
lines = [
line.strip() for line in lines.strip().split("\n")
if len(line.strip()) > 0
]
tokenized = self.tokenizer.encode_batch(lines)
return [tok.ids for tok in tokenized]
def pad_token_id(self) -> int:
return self.tokenizer.token_to_id(self.pad_token)
def mask_token_id(self) -> int:
return self.tokenizer.token_to_id(self.mask_token)
def unk_token_id(self) -> int:
return self.tokenizer.token_to_id(self.unk_token)
def bos_token_id(self) -> int:
return self.tokenizer.token_to_id(self.bos)
def sep_token_id(self) -> int:
return self.tokenizer.token_to_id(self.sep_token)
def token_id(self, token: str) -> int:
tok_id = self.tokenizer.token_to_id(token)
if tok_id is None:
return 0
return tok_id
def id2token(self, id: int) -> str:
return self.tokenizer.id_to_token(id)
def vocab_size(self) -> int:
return self.tokenizer.get_vocab_size()
def is_lang(self, id) -> bool:
return self.tokenizer.id_to_token(id) in self.languages
def lang_id(self, tok):
if tok in self.languages:
return self.languages[tok]
return 0
def split_tokenized(self,
tokenized: List[int],
max_length: int = 512) -> List[List[int]]:
"""
Based on self.max_len, splits very long sequences to smaller ones.
Here we assume to not have any overlapping sequences.
If the first token is a language, we add it to every new sequence.
:return:
"""
if len(tokenized) <= max_length:
sequences = [tokenized]
sequences[-1] = sequences[-1] + (
max_length - len(sequences[-1])) * [self.pad_token_id()]
return sequences
has_lang = self.is_lang(tokenized[0])
sequence = tokenized[0:] if has_lang else tokenized
seq_len = len(sequence)
sep_id = self.sep_token_id()
max_len = max_length - 1 if has_lang else max_length
cur_start = 0
sequences = []
built_seq = []
truncated = False # Shows if previous sequence is truncated due to its length.
used_ends = set()
while cur_start < seq_len:
if not truncated or not has_lang:
cur_end = min(seq_len, cur_start + max_len)
else:
cur_end = min(seq_len, cur_start + max_len + 1)
subseq = sequence[cur_start:cur_end]
built_seq += subseq
sep_positions = [
i for i, id in enumerate(built_seq) if id == sep_id
]
if len(sep_positions) > 0:
if sep_positions[-1] in used_ends:
truncated = True
else:
built_seq = built_seq[:sep_positions[-1] + 1]
truncated = False
else:
truncated = True
assert built_seq[-1] == sequence[len(built_seq) - 1]
if has_lang and len(subseq) < max_len + 1:
subseq = [tokenized[0]] + subseq
sequences.append(subseq)
cur_start = len(built_seq)
used_ends.add(cur_start - 1)
if len(sequences[-1]) < max_length:
sequences[-1] = sequences[-1] + (
max_length - len(sequences[-1])) * [self.pad_token_id()]
assert built_seq[-1] == sequence[len(built_seq) - 1]
return sequences
