Cannot extract entity sentences

[bharath8847]

text=""" @ ! 01/02/21 10:13:46 10:21:02
GROSS VOL Litres START 0.
GROSS VOL Litres FINISH 1000.
TICKET NUMBER 31328.
SALE NUMBER 30046.
DRIVER NUMBER 1.
PRODUCT: KEROSENE DISTILLATE
** MULTIPLE DELIVERIES AT ONE SITE **
"""
doc = nlp(text)
for ent in doc.ents:
print(ent.text,ent.label_)
print('Sentence extracted from : ', ent.sent)

Give me this output
Litres volume
Sentence extracted from : None
Litres volume
Sentence extracted from : None

I have trained my custom NER model for this for this domain specific problem

[adrianeboyd]

What version of spacy are you using (spacy info)? What components are in your pipeline (nlp.pipe_names)? Is there a component that sets sentence boundaries (sentencizer, senter, parser, custom)?

The most recent versions of spacy v2 and all versions of spacy v3 should raise an error instead of returning None, but I think that releases that are over a year old would return None if the doc doesn't have any sentence boundaries.

It's possible the rule-based sentencizer would work well enough for this example text because there are periods at the ends of the lines, but you may need a custom component for this kind of text, which doesn't contain the kinds of full (newspaper-ish) sentences that the parser has learned to segment.