[developers] Getting MRS elementary triples from batch parsing?

Stephan Oepen oe at ifi.uio.no
Mon Aug 29 13:44:03 CEST 2011

hi lea,

when you say 'smrs', 'omrs', and 'pmrs', what exactly do you have in mind?
are these MRSs associated to different stages in an MT setup, or something
rather different?

in general, the conversion of an MRS to ED triples is probably best done in
a post-processing step, e.g. by exporting (or possibly 'thinning') a
profile.  the
'redwoods' script in the LOGON tree has support for exporting into quite a few
different formats, including ED triples.  note that 'redwoods' (in $LOGONROOT)
supersedes the older 'export' or 'itsdb' scripts ...

best, oe

On Mon, Aug 29, 2011 at 12:56, Lea Frermann
<frermann at coli.uni-saarland.de> wrote:
> Hi all,
> I am currently parsing and translating the Tanaka corpus with the
> DelphinTools in order to obtain the smrs, omrs and pmrs. It would be
> very convenient to obtain the MRS elementary triples for the pmrs and
> omrs directly during the parsing process. Does anybody know if this is
> possible?
> Thanks,
> Lea.

+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++    --- oe at ifi.uio.no; stephan at oepen.net; http://www.emmtee.net/oe/ ---

More information about the developers mailing list