Change segmentaion rule

[YaminoYuki]

I want to change (not to add to the pipeline) the segmentation rule

steps:

import spacy
nlp = spacy.load('en_core_web_sm')
myString = "This is a sentrance. This is another. \n\n This is \n third one."

The rule I want to use (segment depending on new line only)

from spacy.pipeline import Sentencizer
def split_on_newlines(doc):
    start= 0
    seen_newline = False
    for word in doc:
        if seen_newline:
            yield doc[start:word.i]
            start=word.i
            seen_newline = False
        elif word.text.startwith('\n'):
            seen_newline= True
    yield doc[start:]

in previous versions there was SentenceSegmenter model which it was replaced by Sentencizer model

with SentendSegmenter it was straightforward to change the rule:

sbd = SentenceSegmenter(nlp.vocab, strategy=split_on_newlines)
nlp.add_pipe(sbd)

but now with Sentencizer I do not know the exact steps:

@Language.component("split_on_newlines")
def split_on_newlines(doc):
    start= 0
    seen_newline = False
    for word in doc:
        if seen_newline:
            yield doc[start:word.i]
            start=word.i
            seen_newline = False
        elif word.text.startwith('\n'):
            seen_newline= True
    yield doc[start:]

sbd = Sentencizer('split_on_newlines', overwrite= True)

is this correct ? if yes then what is next ?

I have posted the question on Stackoverflow but there was no answer for that

again my question is not how to add I have already done with that

spaCy verison

3.7.2

python version

3.12.2

Used IDE

Anaconda + jupyter notebook

[YaminoYuki]

I was able to do it in a different way

nlp = spacy.load('en_core_web_sm') #reload the model 
myString = "This is a sentrance. This is another. ; This is ; third one."
print(myString)

nlp.has_pipe('senter')
config = {"punct_chars": [';'], "overwrite": True}
nlp.add_pipe("sentencizer", config=config)
nlp.pipe_names

doc = nlp(myString)
for sent in doc.sents:
    print(sent)
    print('')

and the result is

This is a sentrance. This is another. ;

This is ;

third one.

but still with '\n' is not working any idea why ?

[YaminoYuki]

Okay,
'\n' doesn't considered as punct that is why, but I still need to figure it out for '\n' ( I know I can replace it with other token but I prefer not)