lemma_postprocessing

This module contains various rule-based components aiming to improve on baseline lemmatization tools.

LemmaSmoother #

Bases: Pipe

Smooths lemma by fixing common errors of the edit-tree lemmatizer.

create_lemma_smoother(nlp, name)

lemma_case_smoother(doc)

Smooth lemma casing by POS.

DEPRECATED: This is not needed anymore, as the lemmatizer is now case-insensitive.

Parameters:

Name	Type	Description	Default
`doc`	`Doc`	Input document.	required

Returns:

Name	Type	Description
`Doc`	`Doc`	Output document.