tokenization_roberta.py

# coding=utf-8
# Copyright 2018 The Open AI Team Authors and The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Tokenization classes for RoBERTa."""
from __future__ import (absolute_import, division, print_function,
                        unicode_literals)

import sys
import json
import logging
import os
import regex as re
from io import open

from .tokenization_gpt2 import bytes_to_unicode, get_pairs
from .tokenization_utils import PreTrainedTokenizer

try:
    from functools import lru_cache
except ImportError:
    # Just a dummy decorator to get the checks to run on python2
    # because honestly I don't want to support a byte-level unicode BPE tokenizer on python 2 right now.
    def lru_cache():
        return lambda func: func

logger = logging.getLogger(__name__)

VOCAB_FILES_NAMES = {
    'vocab_file': 'vocab.json',
    'merges_file': 'merges.txt',
}

PRETRAINED_VOCAB_FILES_MAP = {
    'vocab_file':
    {
        'roberta-base': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-vocab.json",
        'roberta-large': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-vocab.json",
        'roberta-large-mnli': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-mnli-vocab.json",
    },
    'merges_file':
    {
        'roberta-base': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-base-merges.txt",
        'roberta-large': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-merges.txt",
        'roberta-large-mnli': "https://s3.amazonaws.com/models.huggingface.co/bert/roberta-large-mnli-merges.txt",
    },
}

PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
    'roberta-base': 512,
    'roberta-large': 512,
    'roberta-large-mnli': 512,
}


class RobertaTokenizer(PreTrainedTokenizer):
    """
    RoBERTa BPE tokenizer, derived from the GPT-2 tokenizer. Peculiarities: Byte-level BPE
    """
    vocab_files_names = VOCAB_FILES_NAMES
    pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
    max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES

    def __init__(self, vocab_file, merges_file, errors='replace', bos_token="<s>", eos_token="</s>", sep_token="</s>",
                 cls_token="<s>", unk_token="<unk>", pad_token='<pad>', mask_token='<mask>', **kwargs):
        super(RobertaTokenizer, self).__init__(bos_token=bos_token, eos_token=eos_token, unk_token=unk_token,
                                               sep_token=sep_token, cls_token=cls_token, pad_token=pad_token,
                                               mask_token=mask_token, **kwargs)

        self.encoder = json.load(open(vocab_file, encoding="utf-8"))
        self.decoder = {v: k for k, v in self.encoder.items()}
        self.errors = errors  # how to handle errors in decoding
        self.byte_encoder = bytes_to_unicode()
        self.byte_decoder = {v: k for k, v in self.byte_encoder.items()}
        bpe_data = open(merges_file, encoding='utf-8').read().split('\n')[1:-1]
        bpe_merges = [tuple(merge.split()) for merge in bpe_data]
        self.bpe_ranks = dict(zip(bpe_merges, range(len(bpe_merges))))
        self.cache = {}

        # Should haved added re.IGNORECASE so BPE merges can happen for capitalized versions of contractions
        self.pat = re.compile(r"""'s|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+""")

    @property
    def vocab_size(self):
        return len(self.encoder)

    def bpe(self, token):
        if token in self.cache:
            return self.cache[token]
        word = tuple(token)
        pairs = get_pairs(word)

        if not pairs:
            return token

        while True:
            bigram = min(pairs, key = lambda pair: self.bpe_ranks.get(pair, float('inf')))
            if bigram not in self.bpe_ranks:
                break
            first, second = bigram
            new_word = []
            i = 0
            while i < len(word):
                try:
                    j = word.index(first, i)
                    new_word.extend(word[i:j])
                    i = j
                except:
                    new_word.extend(word[i:])
                    break

                if word[i] == first and i < len(word)-1 and word[i+1] == second:
                    new_word.append(first+second)
                    i += 2
                else:
                    new_word.append(word[i])
                    i += 1
            new_word = tuple(new_word)
            word = new_word
            if len(word) == 1:
                break
            else:
                pairs = get_pairs(word)
        word = ' '.join(word)
        self.cache[token] = word
        return word

    def _tokenize(self, text):
        """ Tokenize a string. """
        bpe_tokens = []
        for token in re.findall(self.pat, text):
            if sys.version_info[0] == 2:
                token = ''.join(self.byte_encoder[ord(b)] for b in token)
            else:
                token = ''.join(self.byte_encoder[b] for b in token.encode('utf-8'))
            bpe_tokens.extend(bpe_token for bpe_token in self.bpe(token).split(' '))
        return bpe_tokens

    def _convert_token_to_id(self, token):
        """ Converts a token (str/unicode) in an id using the vocab. """
        return self.encoder.get(token, self.encoder.get(self.unk_token))

    def _convert_id_to_token(self, index):
        """Converts an index (integer) in a token (string/unicode) using the vocab."""
        return self.decoder.get(index)

    def convert_tokens_to_string(self, tokens):
        """ Converts a sequence of tokens (string) in a single string. """
        text = ''.join(tokens)
        text = bytearray([self.byte_decoder[c] for c in text]).decode('utf-8', errors=self.errors)
        return text

    def add_special_tokens_single_sentence(self, token_ids):
        """
        Adds special tokens to a sequence for sequence classification tasks.
        A RoBERTa sequence has the following format: [CLS] X [SEP]
        """
        return [self._convert_token_to_id(self.cls_token)] + token_ids + [self._convert_token_to_id(self.sep_token)]

    def add_special_tokens_sentences_pair(self, token_ids_0, token_ids_1):
        """
        Adds special tokens to a sequence pair for sequence classification tasks.
        A RoBERTa sequence pair has the following format: [CLS] A [SEP][SEP] B [SEP]
        """
        sep = [self._convert_token_to_id(self.sep_token)]
        cls = [self._convert_token_to_id(self.cls_token)]
        return cls + token_ids_0 + sep + sep + token_ids_1 + sep

    def save_vocabulary(self, save_directory):
        """Save the tokenizer vocabulary and merge files to a directory."""
        if not os.path.isdir(save_directory):
            logger.error("Vocabulary path ({}) should be a directory".format(save_directory))
            return
        vocab_file = os.path.join(save_directory, VOCAB_FILES_NAMES['vocab_file'])
        merge_file = os.path.join(save_directory, VOCAB_FILES_NAMES['merges_file'])

        with open(vocab_file, 'w', encoding='utf-8') as f:
            f.write(json.dumps(self.encoder, ensure_ascii=False))

        index = 0
        with open(merge_file, "w", encoding="utf-8") as writer:
            writer.write(u'#version: 0.2\n')
            for bpe_tokens, token_index in sorted(self.bpe_ranks.items(), key=lambda kv: kv[1]):
                if index != token_index:
                    logger.warning("Saving vocabulary to {}: BPE merge indices are not consecutive."
                                   " Please check that the tokenizer is not corrupted!".format(merge_file))
                    index = token_index
                writer.write(' '.join(bpe_tokens) + u'\n')
                index += 1

        return vocab_file, merge_file