File size: 4,081 Bytes
e0d6e3f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 | """Hugging Face tokenizer for CFRD SentencePiece models."""
from __future__ import annotations
import os
import shutil
import unicodedata
from pathlib import Path
import sentencepiece as spm
from transformers import PreTrainedTokenizer
VOCAB_FILES_NAMES = {"vocab_file": "tokenizer.model"}
class CFRDTokenizer(PreTrainedTokenizer):
"""SentencePiece tokenizer with lossless newline preservation."""
vocab_files_names = VOCAB_FILES_NAMES
model_input_names = ["input_ids", "attention_mask"]
def __init__(
self,
vocab_file: str,
newline_token: str = "<|nl|>",
newline_escape_token: str = "<|literal_nl|>",
add_bos_token: bool = True,
add_eos_token: bool = False,
**kwargs,
) -> None:
self.vocab_file = vocab_file
self.newline_token = newline_token
self.newline_escape_token = newline_escape_token
self.add_bos_token = add_bos_token
self.add_eos_token = add_eos_token
self.sp_model = spm.SentencePieceProcessor(model_file=str(vocab_file))
super().__init__(**kwargs)
@property
def vocab_size(self) -> int:
return int(self.sp_model.vocab_size())
def get_vocab(self) -> dict[str, int]:
vocabulary = {self.convert_ids_to_tokens(index): index for index in range(self.vocab_size)}
vocabulary.update(self.get_added_vocab())
return vocabulary
def _tokenize(self, text: str) -> list[str]:
text = unicodedata.normalize("NFC", text)
text = text.replace(self.newline_token, self.newline_escape_token)
text = text.replace("\n", self.newline_token)
return list(self.sp_model.encode(text, out_type=str))
def _convert_token_to_id(self, token: str) -> int:
return int(self.sp_model.piece_to_id(token))
def _convert_id_to_token(self, index: int) -> str:
return str(self.sp_model.id_to_piece(index))
def convert_tokens_to_string(self, tokens: list[str]) -> str:
text = self.sp_model.decode(tokens)
text = text.replace(self.newline_token, "\n")
return text.replace(self.newline_escape_token, self.newline_token)
def build_inputs_with_special_tokens(
self,
token_ids_0: list[int],
token_ids_1: list[int] | None = None,
) -> list[int]:
if token_ids_1 is not None:
raise ValueError("CFRDTokenizer does not support paired sequences")
output = list(token_ids_0)
if self.add_bos_token:
output.insert(0, self.bos_token_id)
if self.add_eos_token:
output.append(self.eos_token_id)
return output
def get_special_tokens_mask(
self,
token_ids_0: list[int],
token_ids_1: list[int] | None = None,
already_has_special_tokens: bool = False,
) -> list[int]:
if already_has_special_tokens:
special_ids = set(self.all_special_ids)
return [1 if token_id in special_ids else 0 for token_id in token_ids_0]
if token_ids_1 is not None:
raise ValueError("CFRDTokenizer does not support paired sequences")
return ([1] if self.add_bos_token else []) + [0] * len(token_ids_0) + (
[1] if self.add_eos_token else []
)
def create_token_type_ids_from_sequences(
self,
token_ids_0: list[int],
token_ids_1: list[int] | None = None,
) -> list[int]:
return [0] * len(self.build_inputs_with_special_tokens(token_ids_0, token_ids_1))
def save_vocabulary(self, save_directory: str, filename_prefix: str | None = None):
directory = Path(save_directory)
directory.mkdir(parents=True, exist_ok=True)
filename = ((filename_prefix + "-") if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"]
destination = directory / filename
if os.path.abspath(self.vocab_file) != os.path.abspath(destination):
shutil.copyfile(self.vocab_file, destination)
return (str(destination),)
CFRDTokenizer.register_for_auto_class("AutoTokenizer")
|