def test_count_vocab_items_case_insensitive(self):
indexer = SingleIdTokenIndexer("words", lowercase_tokens=True)
counter = defaultdict(lambda: defaultdict(int))
indexer.count_vocab_items(Token("Hello"), counter)
indexer.count_vocab_items(Token("hello"), counter)
assert counter["words"] == {"hello": 2}
def test_count_vocab_items_respect_casing(self):
indexer = SingleIdTokenIndexer("words")
counter = defaultdict(lambda: defaultdict(int))
indexer.count_vocab_items(Token("Hello"), counter)
indexer.count_vocab_items(Token("hello"), counter)
assert counter["words"] == {"hello": 1, "Hello": 1}
def __init__(self,
namespace='tokens',
lowercase_tokens: bool = False,
start_tokens: List[str] = None,
end_tokens: List[str] = None,
token_min_padding_length: int = 0):
super(SingleIdTokenIndexer, self).__init__(token_min_padding_length)
self.namespace = namespace
self.lowercase_tokens = lowercase_tokens
self.start_tokens = [Token(t) for t in (start_tokens or [])]
self.end_tokens = [Token(t) for t in (end_tokens or [])]
def __init__(
self,
namespace: str = "token_characters",
character_tokenizer: CharacterTokenizer = CharacterTokenizer(),
start_tokens: List[str] = [],
end_tokens: List[str] = [],
min_padding_length: int = 0,
token_min_padding_length: int = 0) -> None:
# Call super class
super(TokenCharacterIndexer, self).__init__(token_min_padding_length)
self.min_padding_length = min_padding_length
self.namespace = namespace
self.character_tokenizer = character_tokenizer
self.start_tokens = [Token(t) for t in start_tokens]
self.end_tokens = [Token(t) for t in end_tokens]
def text_to_instance(self,
source_string: str,
target_string: str = None) -> Instance:
tokenized_source = self.source_tokenizer.tokenize(source_string)
if self.source_add_start_token:
tokenized_source.insert(0, Token(START_SYMBOL))
tokenized_source.append(Token(END_SYMBOL))
source_field = TextField(tokenized_source, self.source_token_indexers)
if target_string is not None:
tokenized_target = self.target_tokenizer.tokenize(target_string)
tokenized_target.insert(0, Token(START_SYMBOL))
tokenized_target.append(Token(END_SYMBOL))
target_field = TextField(tokenized_target,
self.target_token_indexers)
return Instance({
"source_tokens": source_field,
"target_tokens": target_field
})
else:
return Instance({"source_tokens": source_field})
def split_words(self, sentence: str) -> List[Token]:
tokens = re.findall(r"(?:\d+,\d+)|(?:[\w'\u0080-\u9999]+(?:[-]+[\w'\u0080-\u9999]+)+)|(?:[\w\u0080-\u9999]+(?:[']+[\w\u0080-\u9999]+)+)|\b[_]|(?:[_]*[\w\u0080-\u9999]+(?=_\b))|(?:[\w\u0080-\u9999]+)|[^\w\s\p{Z}]", sentence, re.UNICODE)
return [Token(t) for t in tokens]
def split_words(self, sentence: str) -> List[Token]: return [Token(t) for t in sentence.split()]