[developers] Problems parsing with ACE and gg
ned at nedned.net
Thu Sep 17 15:03:39 CEST 2015
Thanks for the response and suggestion for the workaround and sorry for my
slow response. The context I was trying to get gg working in is a web-app
that depends only upon ACE and doesn't invoke any other DELPH-IN machinery,
so I think I'm stuck with using ACE's repp. I was mainly trying to get gg
working in an effort to plug in as many grammars as possible to my app for
a demo at the recent summit, however standalone ACE support would be good
to have more generally as it seems that embedding ACE within an
application/library is an increasingly popular way of using DELPH-IN
grammars (eg pyDelphin <https://github.com/delph-in/pydelphin>, demophin
<https://github.com/goodmami/demophin> in addition to my Typediff
On Sat, 8 Aug 2015 at 09:15 Berthold Crysmann <berthold.crysmann at gmail.com>
> Dear Ned,
> thanks for your report and sorry for the week's delay. I usually still use
> pet for parsing with gg, so that's why this probably went unnoticed. Should
> make the switch, though.
> After a bit of looking around, I figured out that tokenisation appeared
> to be the culprit. For now, ace parses in yy mode (just like pet).
> If you are using the fine system ([incr tsdb()]), just add "-y" to the ace
> parameters, and copy over the :preprocessor statement from the gg at pet
> I do know that ace natively includes repp, and I did have it working at
> least back in 2012/2013, and some calls are in config.tdl, but they do not
> seem to work right at the mo. I shall look into that next week.
> On 31/07/15 11:07, Ned Letcher wrote:
> Hi all,
> I've encountered some problems parsing with ACE and gg. I cant get it to
> parse even simple German sentences:
> echo "Am Anfang war das Wort." | ace -g ~/data/grammar-data/gg.dat
>> NOTE: 0 readings, added 1237 / 542 edges to chart (254 fully
>> instantiated, 48 actives used, 177 passives used) RAM: 6822k
>> SKIP: Am Anfang war das Wort.
> I'm using the version of gg in the logon repo along with the supplied
> Anyone have any leads? Any extra config I need perhaps?
> Berthold Crysmann <crysmann at linguist.jussieu.fr> <crysmann at linguist.jussieu.fr>
> CNRS, Laboratoire de linguistique formelle (UMR 7110), U Paris Diderot
> Case 7031, 5 rue Thomas Mann, 75205 Paris cedex 13
> Bureau 545, bâtiment Olympe de Gouges, rue Albert Einstein, 75013 Paris
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the developers