Bugs item #3605206, was opened at 2013-02-18 09:23
Message generated for change (Tracker Item Submitted) made by mikeunwalla
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=655717&aid=3605206&group_id=110216
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Problem with English
Group: 2.0
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: MikeUnwalla (mikeunwalla)
Assigned to: Marcin Miłkowski (milek_pl)
Summary: Multiple instances of the same postag for 1 token
Initial Comment:
With a particular rule, I get 5 instances of postag 'VB' for 1 token:
<S> You[you/PRP] must[must/MD] check[check/VB, check/VB, check/VB, check/VB,
check/VB] this[this/DT] text[text/NN:UN].[./., </S>]
This is the rule that gives multiple postags. I have simplified the rule as
much as possible so that there are no distractions:
<rule id="test" name="test">
<pattern>
<token/>
<marker>
<token>check</token>
</marker>
</pattern>
<disambig><match no="1" postag="VB"/></disambig>
</rule>
With an 'ambiguous' example in the rule
(http://wiki.languagetool.org/developing-a-disambiguator#toc11), testrules
gives this error (line breaks removed):
Exception in thread "main" junit.framework.AssertionFailedError: The output
form for the rule test in the example: You must <marker>check</marker> the
sentence.: check[check/JJ,check/NN:UN,check/UH,check/VB,check/VBP] ->
check[check/VB] is different than expected (expected check[check/VB] but got
check[check/VB,check/VB,check/VB,check/VB,check/VB]). The token has been
changed by the disambiguator:
A similar problem occurs with other postags and with other words. The problem
appears to be only with words that are both nouns and verbs, such as ‘check’,
‘test’, and ‘oil’.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=655717&aid=3605206&group_id=110216
------------------------------------------------------------------------------
The Go Parallel Website, sponsored by Intel - in partnership with Geeknet,
is your hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials, tech docs,
whitepapers, evaluation guides, and opinion stories. Check out the most
recent posts - join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Languagetool-commits mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/languagetool-commits
