1
0
mirror of https://github.com/QData/TextAttack.git synced 2021-10-13 00:05:06 +03:00
Files
textattack-nlp-transformer/textattack/shared/tokenized_text.py
2020-05-17 18:23:07 -04:00

196 lines
10 KiB
Python

import torch
from copy import deepcopy
from .utils import get_device, words_from_text
from nltk.corpus import stopwords
class TokenizedText:
"""
A helper class that represents a string that can be attacked.
Models that take multiple sentences as input separate them by `SPLIT_TOKEN`.
Attacks "see" the entire input, joined into one string, without the split token.
Args:
text (string): The string that this TokenizedText represents
tokenizer (textattack.Tokenizer): An object that can encode text
"""
SPLIT_TOKEN = '>>>>'
def __init__(self, text, tokenizer, attack_attrs=dict()):
text = text.strip()
self.tokenizer = tokenizer
ids = tokenizer.encode(text)
if not isinstance(ids, tuple):
# Some tokenizers may tokenize text to a single vector.
# In this case, wrap the vector in a tuple to mirror the
# format of other tokenizers.
ids = (ids,)
self.ids = ids
self.words = words_from_text(text, words_to_ignore=[TokenizedText.SPLIT_TOKEN])
self.text = text
self.attack_attrs = attack_attrs
if 'modified_indices' not in attack_attrs:
attack_attrs['modified_indices'] = set()
if 'stopword_indices' not in attack_attrs:
attack_attrs['stopword_indices'] = set()
def __eq__(self, other):
return (self.text == other.text) and (self.attack_attrs == other.attack_attrs)
def __hash__(self):
return hash(self.text)
def identify_stopwords(self, textfooler_stopwords=False):
self.stopwords = set(stopwords.words('english'))
if textfooler_stopwords:
self.stopwords = set(['a', 'about', 'above', 'across', 'after', 'afterwards', 'again', 'against', 'ain', 'all', 'almost', 'alone', 'along', 'already', 'also', 'although', 'am', 'among', 'amongst', 'an', 'and', 'another', 'any', 'anyhow', 'anyone', 'anything', 'anyway', 'anywhere', 'are', 'aren', "aren't", 'around', 'as', 'at', 'back', 'been', 'before', 'beforehand', 'behind', 'being', 'below', 'beside', 'besides', 'between', 'beyond', 'both', 'but', 'by', 'can', 'cannot', 'could', 'couldn', "couldn't", 'd', 'didn', "didn't", 'doesn', "doesn't", 'don', "don't", 'down', 'due', 'during', 'either', 'else', 'elsewhere', 'empty', 'enough', 'even', 'ever', 'everyone', 'everything', 'everywhere', 'except', 'first', 'for', 'former', 'formerly', 'from', 'hadn', "hadn't", 'hasn', "hasn't", 'haven', "haven't", 'he', 'hence', 'her', 'here', 'hereafter', 'hereby', 'herein', 'hereupon', 'hers', 'herself', 'him', 'himself', 'his', 'how', 'however', 'hundred', 'i', 'if', 'in', 'indeed', 'into', 'is', 'isn', "isn't", 'it', "it's", 'its', 'itself', 'just', 'latter', 'latterly', 'least', 'll', 'may', 'me', 'meanwhile', 'mightn', "mightn't", 'mine', 'more', 'moreover', 'most', 'mostly', 'must', 'mustn', "mustn't", 'my', 'myself', 'namely', 'needn', "needn't", 'neither', 'never', 'nevertheless', 'next', 'no', 'nobody', 'none', 'noone', 'nor', 'not', 'nothing', 'now', 'nowhere', 'o', 'of', 'off', 'on', 'once', 'one', 'only', 'onto', 'or', 'other', 'others', 'otherwise', 'our', 'ours', 'ourselves', 'out', 'over', 'per', 'please','s', 'same', 'shan', "shan't", 'she', "she's", "should've", 'shouldn', "shouldn't", 'somehow', 'something', 'sometime', 'somewhere', 'such', 't', 'than', 'that', "that'll", 'the', 'their', 'theirs', 'them', 'themselves', 'then', 'thence', 'there', 'thereafter', 'thereby', 'therefore', 'therein', 'thereupon', 'these', 'they','this', 'those', 'through', 'throughout', 'thru', 'thus', 'to', 'too','toward', 'towards', 'under', 'unless', 'until', 'up', 'upon', 'used', 've', 'was', 'wasn', "wasn't", 'we', 'were', 'weren', "weren't", 'what', 'whatever', 'when', 'whence', 'whenever', 'where', 'whereafter', 'whereas', 'whereby', 'wherein', 'whereupon', 'wherever', 'whether', 'which', 'while', 'whither', 'who', 'whoever', 'whole', 'whom', 'whose', 'why', 'with', 'within', 'without', 'won', "won't", 'would', 'wouldn', "wouldn't", 'y', 'yet', 'you', "you'd", "you'll", "you're", "you've", 'your', 'yours', 'yourself', 'yourselves'])
self.attack_attrs['stopword_indices'] = set()
for i, word in enumerate(self.words):
if word.lower() in self.stopwords:
self.attack_attrs['stopword_indices'].add(i)
def delete_tensors(self):
""" Delete tensors to clear up GPU space. Only should be called
once the TokenizedText is only needed to display.
"""
self.ids = None
for key in self.attack_attrs:
if isinstance(self.attack_attrs[key], torch.Tensor):
del self.attack_attrs[key]
def text_window_around_index(self, index, window_size):
""" The text window of `window_size` words centered around `index`. """
length = len(self.words)
half_size = (window_size - 1) // 2
if index - half_size < 0:
start = 0
end = min(window_size, length-1)
elif index + half_size > length - 1:
start = max(0, length - window_size)
end = length - 1
else:
start = index - half_size
end = index + half_size
text_idx_start = self._text_index_of_word_index(start)
text_idx_end = self._text_index_of_word_index(end) + len(self.words[end])
return self.text[text_idx_start:text_idx_end]
def _text_index_of_word_index(self, i):
""" Returns the index of word `i` in self.text. """
pre_words = self.words[:i+1]
lower_text = self.text.lower()
# Find all words until `i` in string.
look_after_index = 0
for word in pre_words:
look_after_index = lower_text.find(word.lower(), look_after_index)
return look_after_index
def text_until_word_index(self, i):
""" Returns the text before the beginning of word at index `i`. """
look_after_index = self._text_index_of_word_index(i)
return self.text[:look_after_index]
def text_after_word_index(self, i):
""" Returns the text after the end of word at index `i`. """
# Get index of beginning of word then jump to end of word.
look_after_index = self._text_index_of_word_index(i) + len(self.words[i])
return self.text[look_after_index:]
def first_word_diff(self, other_tokenized_text):
""" Returns the first word in self.words that differs from
other_tokenized_text. Useful for word swap strategies. """
w1 = self.words
w2 = other_tokenized_text.words
for i in range(min(len(w1), len(w2))):
if w1[i] != w2[i]:
return w1
return None
def first_word_diff_index(self, other_tokenized_text):
""" Returns the index of the first word in self.words that differs
from other_tokenized_text. Useful for word swap strategies. """
w1 = self.words
w2 = other_tokenized_text.words
for i in range(min(len(w1), len(w2))):
if w1[i] != w2[i]:
return i
return None
def all_words_diff(self, other_tokenized_text):
""" Returns the set of indices for which this and other_tokenized_text
have different words. """
indices = set()
w1 = self.words
w2 = other_tokenized_text.words
for i in range(min(len(w1), len(w2))):
if w1[i] != w2[i]:
indices.add(i)
return indices
def ith_word_diff(self, other_tokenized_text, i):
""" Returns whether the word at index i differs from other_tokenized_text
"""
w1 = self.words
w2 = other_tokenized_text.words
if len(w1) - 1 < i or len(w2) - 1 < i:
return True
return w1[i] != w2[i]
def replace_words_at_indices(self, indices, new_words):
""" This code returns a new TokenizedText object where the word at
`index` is replaced with a new word."""
if len(indices) != len(new_words):
raise ValueError(f'Cannot replace {len(new_words)} words at {len(indices)} indices.')
words = self.words[:]
for i, new_word in zip(indices, new_words):
words[i] = new_word
return self.replace_new_words(words)
def replace_word_at_index(self, index, new_word):
""" This code returns a new TokenizedText object where the word at
`index` is replaced with a new word."""
return self.replace_words_at_indices([index], [new_word])
def replace_new_words(self, new_words):
""" This code returns a new TokenizedText object and replaces old list
of words with a new list of words, but preserves the punctuation
and spacing of the original message.
"""
final_sentence = ''
text = self.text
new_attack_attrs = deepcopy(self.attack_attrs)
new_attack_attrs['stopword_indices'] = set()
new_attack_attrs['modified_indices'] = set()
new_attack_attrs['newly_modified_indices'] = set()
new_i = 0
for i, (input_word, adv_word) in enumerate(zip(self.words, new_words)):
if input_word == '[DELETE]': continue
word_start = text.index(input_word)
word_end = word_start + len(input_word)
final_sentence += text[:word_start]
final_sentence += adv_word
text = text[word_end:]
if i in self.attack_attrs['stopword_indices']:
new_attack_attrs['stopword_indices'].add(new_i)
if i in self.attack_attrs['modified_indices'] or input_word != adv_word:
new_attack_attrs['modified_indices'].add(new_i)
if input_word != adv_word:
new_attack_attrs['newly_modified_indices'].add(new_i)
new_i += 1
final_sentence += text # Add all of the ending punctuation.
return TokenizedText(final_sentence, self.tokenizer,
attack_attrs=new_attack_attrs)
def clean_text(self):
""" Represents self in a clean, printable format. Joins text with multiple
inputs separated by `TokenizedText.SPLIT_TOKEN` with a line break.
"""
return self.text.replace(TokenizedText.SPLIT_TOKEN, '\n\n')
def __repr__(self):
return f'<TokenizedText "{self.text}">'