Skip to content

tokenize

Tokenize

Bases: NumpyOp

Split the sequences into tokens.

Tokenize split the document/sequence into tokens and at the same time perform additional operations on tokens if defined in the passed function object. By default, tokenize only splits the sequences into tokens.

Parameters:

Name Type Description Default
inputs Union[str, Iterable[str]]

Key(s) of sequences to be tokenized.

 required
outputs Union[str, Iterable[str]]

Key(s) of sequences that are tokenized.

 required
mode Union[None, str, Iterable[str]]

What mode(s) to execute this Op in. For example, "train", "eval", "test", or "infer". To execute regardless of mode, pass None. To execute in all modes except for a particular one, you can pass an argument like "!infer" or "!train".

 None
ds_id Union[None, str, Iterable[str]]

What dataset id(s) to execute this Op in. To execute regardless of ds_id, pass None. To execute in all ds_ids except for a particular one, you can pass an argument like "!ds1".

 None
tokenize_fn Union[None, Callable[[str], List[str]]]

Tokenization function object.

 None
to_lower_case bool

Whether to convert tokens to lowercase.

 False
Source code in fastestimator/fastestimator/op/numpyop/univariate/tokenize.py 
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
@traceable()
class Tokenize(NumpyOp):
    """Split the sequences into tokens.

    Tokenize split the document/sequence into tokens and at the same time perform additional operations on tokens if
    defined in the passed function object. By default, tokenize only splits the sequences into tokens.

    Args:
        inputs: Key(s) of sequences to be tokenized.
        outputs: Key(s) of sequences that are tokenized.
        mode: What mode(s) to execute this Op in. For example, "train", "eval", "test", or "infer". To execute
            regardless of mode, pass None. To execute in all modes except for a particular one, you can pass an argument
            like "!infer" or "!train".
        ds_id: What dataset id(s) to execute this Op in. To execute regardless of ds_id, pass None. To execute in all
            ds_ids except for a particular one, you can pass an argument like "!ds1".
        tokenize_fn: Tokenization function object.
        to_lower_case: Whether to convert tokens to lowercase.
    """
    def __init__(self,
                 inputs: Union[str, Iterable[str]],
                 outputs: Union[str, Iterable[str]],
                 mode: Union[None, str, Iterable[str]] = None,
                 ds_id: Union[None, str, Iterable[str]] = None,
                 tokenize_fn: Union[None, Callable[[str], List[str]]] = None,
                 to_lower_case: bool = False) -> None:
        super().__init__(inputs=inputs, outputs=outputs, mode=mode, ds_id=ds_id)
        self.in_list, self.out_list = True, True
        self.tokenize_fn = tokenize_fn
        self.to_lower_case = to_lower_case

    def forward(self, data: List[str], state: Dict[str, Any]) -> List[List[str]]:
        return [self._apply_tokenization(seq) for seq in data]

    def _apply_tokenization(self, data: str) -> List[str]:
        """Split the sequence into tokens and apply lowercase if `do_lower_case` is set.

        Args:
            data: Input sequence.

        Returns:
            A list of tokens.
        """
        if self.tokenize_fn:
            data = self.tokenize_fn(data)
        else:
            data = data.split()
        if self.to_lower_case:
            data = list(map(lambda x: x.lower(), data))
        return data