Lemmatizer

Lemmatizer module. Provides classes for lemmatizing token and full texts.

• Lemmatizer: Class for performing token and full text lemmatization.
• is_known(): A legacy function that wraps the Lemmatizer's [is_known()][simplemma.lemmatizer.Lemmatizer.is_known] method.
• lemmatize(): A legacy function that wraps the Lemmatizer's lemmatize() method.
• text_lemmatizer(): A legacy function that wraps the Lemmatizer's text_lemmatizer() method.
• lemma_iterator(): A legacy function that wraps the Lemmatizer's lemma_iterator() method.

Classes

Lemmatizer

Lemmatizer class for performing token lemmatization.

Source code in simplemma/lemmatizer.py

class Lemmatizer:
    """Lemmatizer class for performing token lemmatization."""

    __slots__ = [
        "_cached_lemmatize",
        "_fallback_lemmatization_strategy",
        "_lemmatization_strategy",
        "_tokenizer",
    ]

    def __init__(
        self,
        cache_max_size: int = 1048576,
        tokenizer: Tokenizer = RegexTokenizer(),
        lemmatization_strategy: LemmatizationStrategy = DefaultStrategy(),
        fallback_lemmatization_strategy: LemmatizationFallbackStrategy = ToLowercaseFallbackStrategy(),
    ) -> None:
        """
        Initialize the Lemmatizer.

        Args:
            cache_max_size (int, optional): The maximum size of the cache for the lemmatization results.
                Defaults to `1048576`.
            tokenizer (Tokenizer, optional): The tokenizer to use for tokenization.
                Defaults to `RegexTokenizer()`.
            lemmatization_strategy (LemmatizationStrategy, optional): The lemmatization strategy to use.
                Defaults to `DefaultStrategy()`.
            fallback_lemmatization_strategy (LemmatizationFallbackStrategy, optional): The fallback lemmatization strategy to use.
                Defaults to `ToLowercaseFallbackStrategy()`.

        """
        self._tokenizer = tokenizer
        self._lemmatization_strategy = lemmatization_strategy
        self._fallback_lemmatization_strategy = fallback_lemmatization_strategy
        self._cached_lemmatize = lru_cache(maxsize=cache_max_size)(self._lemmatize)

    def lemmatize(
        self,
        token: str,
        lang: Union[str, Tuple[str, ...]],
    ) -> str:
        """Get the lemmatized form of a given word in the specified language(s).

        Args:
            token: The token to lemmatize.
            lang: The language or languages for lemmatization.

        Returns:
            str: The lemmatized form of the token.
        """
        return self._cached_lemmatize(token, lang)

    def _lemmatize(
        self,
        token: str,
        lang: Union[str, Tuple[str, ...]],
    ) -> str:
        """Internal method to lemmatize a token in the specified language(s).

        Args:
            token: The token to lemmatize.
            lang: The language or languages for lemmatization.

        Returns:
            str: The lemmatized form of the token.
        """
        _control_input_type(token)
        lang = validate_lang_input(lang)

        for lang_code in lang:
            candidate = self._lemmatization_strategy.get_lemma(token, lang_code)
            if candidate is not None:
                return candidate

        return self._fallback_lemmatization_strategy.get_lemma(token, next(iter(lang)))

    def get_lemmas_in_text(
        self,
        text: str,
        lang: Union[str, Tuple[str, ...]],
    ) -> Iterator[str]:
        """Get an iterator over lemmatized tokens in a text.

        Args:
            text: The text to process.
            lang: The language or languages for lemmatization.

        Yields:
            str: The lemmatized tokens in the text.
        """
        initial = True
        for token in self._tokenizer.split_text(text):
            yield self.lemmatize(token.lower() if initial else token, lang)
            initial = token in PUNCTUATION

Functions

__init__(cache_max_size=1048576, tokenizer=RegexTokenizer(), lemmatization_strategy=DefaultStrategy(), fallback_lemmatization_strategy=ToLowercaseFallbackStrategy())

Initialize the Lemmatizer.

Parameters:

Name	Type	Description	Default
cache_max_size	int	The maximum size of the cache for the lemmatization results. Defaults to 1048576.	1048576
tokenizer	Tokenizer	The tokenizer to use for tokenization. Defaults to RegexTokenizer().	RegexTokenizer()
lemmatization_strategy	LemmatizationStrategy	The lemmatization strategy to use. Defaults to DefaultStrategy().	DefaultStrategy()
fallback_lemmatization_strategy	LemmatizationFallbackStrategy	The fallback lemmatization strategy to use. Defaults to ToLowercaseFallbackStrategy().	ToLowercaseFallbackStrategy()

Source code in simplemma/lemmatizer.py

def __init__(
    self,
    cache_max_size: int = 1048576,
    tokenizer: Tokenizer = RegexTokenizer(),
    lemmatization_strategy: LemmatizationStrategy = DefaultStrategy(),
    fallback_lemmatization_strategy: LemmatizationFallbackStrategy = ToLowercaseFallbackStrategy(),
) -> None:
    """
    Initialize the Lemmatizer.

    Args:
        cache_max_size (int, optional): The maximum size of the cache for the lemmatization results.
            Defaults to `1048576`.
        tokenizer (Tokenizer, optional): The tokenizer to use for tokenization.
            Defaults to `RegexTokenizer()`.
        lemmatization_strategy (LemmatizationStrategy, optional): The lemmatization strategy to use.
            Defaults to `DefaultStrategy()`.
        fallback_lemmatization_strategy (LemmatizationFallbackStrategy, optional): The fallback lemmatization strategy to use.
            Defaults to `ToLowercaseFallbackStrategy()`.

    """
    self._tokenizer = tokenizer
    self._lemmatization_strategy = lemmatization_strategy
    self._fallback_lemmatization_strategy = fallback_lemmatization_strategy
    self._cached_lemmatize = lru_cache(maxsize=cache_max_size)(self._lemmatize)

get_lemmas_in_text(text, lang)

Get an iterator over lemmatized tokens in a text.

Parameters:

Name	Type	Description	Default
text	str	The text to process.	required
lang	Union[str, Tuple[str, ...]]	The language or languages for lemmatization.	required

Yields:

Name	Type	Description
str	str	The lemmatized tokens in the text.

Source code in simplemma/lemmatizer.py

def get_lemmas_in_text(
    self,
    text: str,
    lang: Union[str, Tuple[str, ...]],
) -> Iterator[str]:
    """Get an iterator over lemmatized tokens in a text.

    Args:
        text: The text to process.
        lang: The language or languages for lemmatization.

    Yields:
        str: The lemmatized tokens in the text.
    """
    initial = True
    for token in self._tokenizer.split_text(text):
        yield self.lemmatize(token.lower() if initial else token, lang)
        initial = token in PUNCTUATION

lemmatize(token, lang)

Get the lemmatized form of a given word in the specified language(s).

Parameters:

Name	Type	Description	Default
token	str	The token to lemmatize.	required
lang	Union[str, Tuple[str, ...]]	The language or languages for lemmatization.	required

Returns:

Name	Type	Description
str	str	The lemmatized form of the token.

Source code in simplemma/lemmatizer.py

def lemmatize(
    self,
    token: str,
    lang: Union[str, Tuple[str, ...]],
) -> str:
    """Get the lemmatized form of a given word in the specified language(s).

    Args:
        token: The token to lemmatize.
        lang: The language or languages for lemmatization.

    Returns:
        str: The lemmatized form of the token.
    """
    return self._cached_lemmatize(token, lang)

Functions

is_known(token, lang)

Check if a token is known in the specified language(s).

Parameters:

Name	Type	Description	Default
token	str	The token to check.	required
lang	Union[str, Tuple[str, ...]]	The language or languages to check in.	required

Returns:

Name	Type	Description
bool	bool	True if the token is known, False otherwise.

Source code in simplemma/lemmatizer.py

def is_known(token: str, lang: Union[str, Tuple[str, ...]]) -> bool:
    """Check if a token is known in the specified language(s).

    Args:
        token: The token to check.
        lang: The language or languages to check in.

    Returns:
        bool: True if the token is known, False otherwise.
    """

    _control_input_type(token)
    lang = validate_lang_input(lang)

    dictionary_lookup = DictionaryLookupStrategy(_legacy_dictionary_factory)
    return any(
        dictionary_lookup.get_lemma(token, lang_code) is not None for lang_code in lang
    )

lemma_iterator(text, lang, greedy=False)

Iterate over lemmatized tokens in a text.

Parameters:

Name	Type	Description	Default
text	str	The text to iterate over.	required
lang	Union[str, Tuple[str, ...]]	The language or languages for lemmatization.	required
greedy	bool	A flag indicating whether to use greedy lemmatization (default: False).	False
tokenizer		The tokenizer to use (default: RegexTokenizer()).	required

Yields:

Name	Type	Description
str	str	The lemmatized tokens in the text.

Source code in simplemma/lemmatizer.py

def lemma_iterator(
    text: str, lang: Union[str, Tuple[str, ...]], greedy: bool = False
) -> Iterator[str]:
    """Iterate over lemmatized tokens in a text.

    Args:
        text: The text to iterate over.
        lang: The language or languages for lemmatization.
        greedy: A flag indicating whether to use greedy lemmatization (default: False).
        tokenizer: The tokenizer to use (default: RegexTokenizer()).

    Yields:
        str: The lemmatized tokens in the text.
    """
    lemmatizer = _legacy_lemmatizer if not greedy else _legacy_greedy_lemmatizer
    return lemmatizer.get_lemmas_in_text(text, lang)

lemmatize(token, lang, greedy=False)

Lemmatize a token in the specified language(s).

Parameters:

Name	Type	Description	Default
token	str	The token to lemmatize.	required
lang	Union[str, Tuple[str, ...]]	The language or languages for lemmatization.	required
greedy	bool	A flag indicating whether to use greedy lemmatization (default: False).	False

Returns:

Name	Type	Description
str	str	The lemmatized form of the token.

Source code in simplemma/lemmatizer.py

def lemmatize(
    token: str, lang: Union[str, Tuple[str, ...]], greedy: bool = False
) -> str:
    """Lemmatize a token in the specified language(s).

    Args:
        token: The token to lemmatize.
        lang: The language or languages for lemmatization.
        greedy: A flag indicating whether to use greedy lemmatization (default: False).

    Returns:
        str: The lemmatized form of the token.
    """
    lemmatizer = _legacy_lemmatizer if not greedy else _legacy_greedy_lemmatizer
    return lemmatizer.lemmatize(token, lang)

text_lemmatizer(text, lang, greedy=False)

Lemmatize a text in the specified language(s).

Parameters:

Name	Type	Description	Default
text	str	The text to lemmatize.	required
lang	Union[str, Tuple[str, ...]]	The language or languages for lemmatization.	required
greedy	bool	A flag indicating whether to use greedy lemmatization (default: False).	False
tokenizer		The tokenizer to use (default: RegexTokenizer()).	required

Returns:

Type	Description
List[str]	List[str]: The list of lemmatized tokens.

Source code in simplemma/lemmatizer.py

def text_lemmatizer(
    text: str, lang: Union[str, Tuple[str, ...]], greedy: bool = False
) -> List[str]:
    """Lemmatize a text in the specified language(s).

    Args:
        text: The text to lemmatize.
        lang: The language or languages for lemmatization.
        greedy: A flag indicating whether to use greedy lemmatization (default: False).
        tokenizer: The tokenizer to use (default: RegexTokenizer()).

    Returns:
        List[str]: The list of lemmatized tokens.
    """

    return list(
        lemma_iterator(
            text,
            lang,
            greedy,
        )
    )