Open MikeUnwalla opened 3 years ago
@udomai , there is also a problem with 'min' in a token:
<token postag="NNP?|NN:UN?|JJ.*" postag_regexp="yes" min="0" chunk_re="I-NP-.*">
With this token, the rule does not find the error: Many large interesting book are on the table.
But this token causes the rule to find the error:
<token postag="NNP?|NN:UN?|JJ.*" postag_regexp="yes" min="0" max="3" chunk_re="I-NP-.*">
(Aside, Premium SINGULAR_NOUN_VERB_AGREEMENT finds the error.)
I have not made an issue for this because we already have some issues about the problems of min/max not working as expected.
@udomai ,
What is special about this rule such that we must preserve the case? Does the problem occur with all rules? If only this rule, why?
From regression https://internal1.languagetool.org/regression-tests/via-http/2021-04-19/en-US/result_grammar_MANY_NN[1].html
Some suggestions have incorrect capitalization (examples 1, 65, 67). Try adding case_conversion="preserve"? The fact that, before the suggestion, there is a match no="3" in the message (instead of just \3) might cause some unexpected behaviour, also.
False positives: