Matcher does not support attribute IS_SENT_START

[alessandrobokan]

How to reproduce the behaviour

When I tried to add a pattern to the spacy Matcher with the attribute IS_SENT_START gives me an error

from spacy.matcher import Matcher
matcher = Matcher(nlp.vocab)
matcher.add("rule1", None, [{"IS_SENT_START": True, "LOWER": "hello"}])

ValueError: [E152] The attribute None is not supported for token patterns. Please use the option validate=True with Matcher, PhraseMatcher, or EntityRuler for more details.

I analysed the source-code of matcher.pyx and I see that attr IS_SENT_START is in TOKEN_PATTERN_SCHEMA but in the file attrs.pdx appears as SENT_START, causing the error.

This issue could be solved by adding following condition after this line 674

if attr == "IS_SENT_START":
    attr = "SENT_START"

Your Environment

• Operating System: Ubuntu 18.04
• Python Version Used: 3.6
• spaCy Version Used: 2.2.3
• Environment Information: virtualenv

[adrianeboyd]

Thanks for the report! This should be fixed by #5080.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.