File size: 1,660 Bytes
6a62ffb |
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 |
# Copyright (c) Facebook, Inc. and its affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from dataclasses import dataclass, field
from fairseq.data.encoders import register_tokenizer
from fairseq.dataclass import FairseqDataclass
@dataclass
class MosesTokenizerConfig(FairseqDataclass):
source_lang: str = field(default="en", metadata={"help": "source language"})
target_lang: str = field(default="en", metadata={"help": "target language"})
moses_no_dash_splits: bool = field(
default=False, metadata={"help": "don't apply dash split rules"}
)
moses_no_escape: bool = field(
default=False,
metadata={"help": "don't perform HTML escaping on apostrophe, quotes, etc."},
)
@register_tokenizer("moses", dataclass=MosesTokenizerConfig)
class MosesTokenizer(object):
def __init__(self, cfg: MosesTokenizerConfig):
self.cfg = cfg
try:
from sacremoses import MosesTokenizer, MosesDetokenizer
self.tok = MosesTokenizer(cfg.source_lang)
self.detok = MosesDetokenizer(cfg.target_lang)
except ImportError:
raise ImportError(
"Please install Moses tokenizer with: pip install sacremoses"
)
def encode(self, x: str) -> str:
return self.tok.tokenize(
x,
aggressive_dash_splits=(not self.cfg.moses_no_dash_splits),
return_str=True,
escape=(not self.cfg.moses_no_escape),
)
def decode(self, x: str) -> str:
return self.detok.detokenize(x.split())
|