I made a rule that works rather well. But with a few exceptions, I think
cause by the tokenizer.
The rule:
(Intendend to catch all sentences containing just 1 word, te be expanded
to catch 2, 3 etc.), to start capturing unusual sentence constructions
structurally)
<category name="Experimenteel, vangen van zinnen">
<rule id="VANG_1" name="Vang 1 woord">
<pattern>
<token postag="SENT_START"/>
<token><exception regexp="yes">[0-9]{1,2}|[a-z]</exception></token>
<token postag="SENT_END" regexp="yes">\!|\?|\.</token>
</pattern>
<message></message>
<example type="incorrect"><marker>Hij.</marker></example>
<example type="correct">Hij kan het</example>
</rule>
</category>
This rule catches:
1°.
-3.
Is there an explanation for the rule catching 2 characters in the token,
like -3 and 1° ?
Ruud
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
Languagetool-devel mailing list
Languagetool-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/languagetool-devel
