[Corpora-List] Charniak parser

Wed Aug 24 17:45:43 UTC 2011

Hi all,

I just found my way to this thread.  The official Charniak parser now lives
on GitHub and BitBucket (synchronized with each other, so take your pick):

   http://github.com/BLLIP/bllip-parser
   http://bitbucket.org/bllip/bllip-parser

These versions should compile.  If they do not, please file an issue on the
GitHub page.

Take care,
David

On Wed, Aug 17, 2011 at 4:43 PM, Nitin Madnani <nmadnani at gmail.com> wrote:

> I wrote a little patch to enable charniak parser to be compiled on modern
> linux distributions. It's available from my web page.
>
> On Wednesday, August 17, 2011, Pawel Mazur <pawel.mazur at mq.edu.au> wrote:
> > Hi,
> >
> > On Mac OS X you can compile the Charniak parser (parser05Aug16) using
> gcc/g++ v. 4.2.1 (the default one in my installation of OS X 10.6.8).
> > You only need to add
> > #include "GotIter.h"
> > to parser05Aug16/PARSE/BchartSm.C
> > To compile the content of the TRAIN directory, you need to change a part
> of the line 310 in rCounts.C from
> > (int)sbrk(0)
> > to
> > (long)sbrk(0)
> >
> >
> > In the case of the Charniak-Johnson reranking parser
> (reranking-parserAug06) you need to:
> > -  change the line 128 of first-stage/PARSE/parseIn.C from
> > int id = (int)arg;
> > to
> > int id = (long)arg;
> > - run 'make clean' before 'make' to recompile
> second-stage/programs/features/best-parses
> >
> > You can also rename features.gz to features.gz.Z and
> cvlm-l1c10P1-weights.gz to cvlm-l1c10P1-weights.gz.Z and make the relevant
> changes in parse.sh;
> > without this step the parser is working, but zcat complains about missing
> gz.Z files; doing this step stopped the messages to appear.
> > Alternatively, you can use gzcat instead of zcat in Makefiles.
> >
> > I thank Brett Powley for these hints.
> >
> > regards,
> > Pawel
> >
> >
> >
> >
> > On 22/06/2010, at 8:38 PM, Leon Derczynski wrote:
> >
> >> Hi,
> >>
> >> Problem solved - using gcc/g++ 3.3 (and associated stdlib), the parser
> compiles.
> >>
> >> All the best,
> >>
> >>
> >> Leon
> >>
> >> -------- Original Message --------
> >> Subject: Re: [Corpora-List] Charniak parser
> >> Date: Tue, 22 Jun 2010 20:15:27 +0200
> >> From: Roman Klinger <roman.klinger at scai.fraunhofer.de>
> >> To: Leon Derczynski <L.Derczynski at dcs.shef.ac.uk>
> >> References: <4C20FD0B.60206 at dcs.shef.ac.uk>
> >>
> >> Hi,
> >>
> >> On 06/22/2010 08:12 PM, Leon Derczynski wrote:
> >>> Thank you for your answer! I've replied directly as I'm not sure the
> list is the right place to
> >>> discuss compilation issues, I may be wrong.
> >>
> >> Perhaps not in detail, but others may be interested. We can post the
> >> solution if we get one.
> >>
> >> Which GCC version are you using? I used 3.3. Got similar errors with
> 4.*.
> >>
> >> I am not a C expert, therefore I am not sure if this is your problem,
> >> but give it a try :-).
> >>
> >> Best,
> >> Roman
> >>
> >>
> >>
> >>>
> >>> I am also using this version (from
> http://bllip.cs.brown.edu/download/reranking-parserAug06.tar.gz);
> >>> my C++ skill ranks between very low and zero. I've had problems with
> declarations likely just
> >>> missing from my path, which I can fix, and manifest as:
> >>>
> >>> BchartSm.C: In member function ‘double Bchart::pHypgt(const
> std::string&, int)’:
> >>> BchartSm.C:554: error: ‘strpbrk’ was not declared in this scope
> >>> make[1]: *** [BchartSm.o] Error 1
> >>> make[1]: Leaving directory
> `/home/leon/charniak/reranking-parser/first-stage/PARSE'
> >>>
> >>> However, after adding include lines for stdlib.h and string.h to
> BchartSm.C, I am stumped by:
> >>>
> >>> BchartSm.C: In member function ‘double Bchart::pHypgt(const
> std::string&, int)’:
> >>> BchartSm.C:555: error: invalid conversion from ‘const char*’ to ‘char*’
> >>> make[1]: *** [BchartSm.o] Error 1
> >>>
> >>> It really seems as though I have some other misconfiguration, if you
> have compiled the same code
> >>> without issues. If you have any clue as to what it may be, I would be
> extremely grateful to hear it.
> >>>
> >>> All the best,
> >>>
> >>>
> >>> Leon
> >>>
> >>> ----------
> >>>
> >>> Hi Leon,
> >>> On 06/22/2010 07:26 PM, Leon Derczynski wrote:
> >>>> I am looking for a working version of the Charniak parser.
> >>>
> >>> You mean the reranking parser published on
> http://bllip.cs.brown.edu/resources.shtml right?
> >>>
> >>>
> >>>> I have scoured a few sites, tried and
> >>>> failed to build a few versions from the web and applied various
> patches, without much success.
> >>> [...]
> >>>
> >>>
> >>>> Can anybody help?
> >>>
> >>> I am using this parser, without any problems. Therefore, it is
> difficult to guess, what your problem
> >>> is. Compilation worked without any issues.
> >>>
> >>> Could you specify what does not work for you?
> >>>
> >>> Best,
> >>>
> >>> Roman
> >>
> >>
> >> --
> >> Roman Klinger
> >> Fraunhofer-Institute for Algorithms and Scientific Computing (SCAI)
> >> Department of Bioinformatics
> >> Schloss Birlinghoven
> >> D-53754 Sankt Augustin
> >> Tel.: +49-2241-14-2360
> >> Fax.: +49-2241-14-4-2360
> >> email: roman.klinger at scai.fhg.de
> >> http://www.scai.fraunhofer.de/klinger.html
> >>
> >> _______________________________________________
> >> Corpora mailing list
> >> Corpora at uib.no
> >> http://mailman.uib.no/listinfo/corpora
> >
> >
> > _______________________________________________
> > UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> > Corpora mailing list
> > Corpora at uib.no
> > http://mailman.uib.no/listinfo/corpora
> >
>
> --
> Linguist, Desi Linguist
> http://www.desilinguist.org
>
> _______________________________________________
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20110824/b7bdee33/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora