Nabble - Linguistic Knowledge Builder (LKB)

Re: cannot load grammar interactively

2009-08-02T22:57:28Z

G'day,

> The file loading dialog of Load -> Complete Grammar doesn't allow me to open any
> directory or file. When I open the dialog or double-click any file, I get error
> messages like these:
>
> Warning: Xt:
> Name: Text
> Class: XmTextField
> Character '\57' not supported in font. Discarded.
>
> Double-clicks on directories/files or clicks on directories/files followed by
> "Ok" are ignored.
>
> This is on Ubuntu 9.04. Francis had the same problem which he could fix by
> uninstalling the ttf-arphic-uming package, but that didn't help in my case. Does
> anyone know how to fix this?

It's not really a fix, more of a work around, but you can load the
grammar from lisp:
(lkb::read-script-file-aux "path/to/lkb/script")
If you use the lkb-mode, then there is a helpful shortcut 'C-c g'.

--
Francis Bond <http://www3.ntu.edu.sg/home/fcbond/>
Division of Linguistics and Multilingual Studies
Nanyang Technological University

cannot load grammar interactively

2009-07-31T01:35:30Z

Hi!

The file loading dialog of Load -> Complete Grammar doesn't allow me to open any
directory or file. When I open the dialog or double-click any file, I get error
messages like these:

Warning: Xt:
Name: Text
Class: XmTextField
Character '\57' not supported in font. Discarded.

Double-clicks on directories/files or clicks on directories/files followed by
"Ok" are ignored.

This is on Ubuntu 9.04. Francis had the same problem which he could fix by
uninstalling the ttf-arphic-uming package, but that didn't help in my case. Does
anyone know how to fix this?

Thanks & Cheers,

Peter

--
Peter Adolphs
DFKI GmbH, Projektbuero Berlin, Alt-Moabit 91c, D-10559 Berlin
Telefon: 030 / 3949-1826
_______________________________________________________________

Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern

Geschaeftsfuehrung:
Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender)
Dr. Walter Olthoff

Vorsitzender des Aufsichtsrats:
Prof. Dr. h.c. Hans A. Aukes

Amtsgericht Kaiserslautern, HRB 2313

A Linguistic Platform for All Your Needs

2009-04-15T03:59:50Z

found this lately and this is a really great resource for translators to find
linguistic assignments. They are the biggest platform to find new translation
assignments daily online and they have updates on new assignments round the
clock unlike other sites. Visit them to see for yourself at www dot translatorpub dot com.

Re: Common Lisp License

2008-09-25T07:04:53Z

I'm afraid I think this will apply to all recent versions - I only
just realised that we'd had one of those licences. It's expired and I
don't currently have grant money to acquire a new one. I am trying to
find a way of acquiring one!

Ann

Re: Common Lisp License

2008-09-25T02:56:55Z

hi scott,

running Linux is, of course, an excellent work-around :-). i wanted to
ask what the value of *cvs-version* was in your Windows version of the
LKB, but i reckon it may be difficult asking it now. however, i guess
you had been using a relatively old copy, and it may just be time to do
a fresh installation. the current `latest' build on the LinGO download
site

http://lingo.stanford.edu/ftp/

should include a relatively recent Windows license. truth be told, i
cannot say for sure, because ann is providing the Windows builds, with
a license sponsored by cambridge. but give it a shot, nevertheless!

if you have CygWin installed (including wget(1), which is optional), i
suggest you try the automated install from:

http://wiki.delph-in.net/moin/LkbInstallation

by default, this will aim to install the `latest' build of everything.

all best - oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++ CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++ --- oe@...; oe@...; stephan@... ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Common Lisp License

2008-09-25T02:39:57Z

When I try to run the Windows version of the LKB (which I've had installed
for many months), I get the following error message:

Lisp has expired. Please contact sales@... for a new license
file.

Should I be able to download a new licence file from somewhere? I tried
downloading the current version of the LKB and copying over the old files,
but that didn't do the trick. The Linux version I have installed still
seems happy, though, so that's my current workaround.

- Scott Drellishak

Bottom-Up Type Hierarchy

2008-09-05T20:47:54Z

Hi,

I found that it's usually useful to have a bottom-up type hierarchy which could tell what a type's parent types are, and what types their parent types inherit, all the way up to the *root*. In this way, it's easy to tell from the hierarchy how a leaf type is formed, or in other words the family tree of a leaf type.
Any advice about this idea? Is it possible to implement it? Has anyone tried to do it?

Another problem I encountered is that in a complex type hierarchy, it's hard to recognize which line leads to where. Is there a better solution?

Thanks.

Best regards,
Norman

Re: How to display thai fonts in LKB

2008-08-27T06:38:21Z

Try http://wiki.delph-in.net/moin/LkbTrollet

Best
Gautam

On Thu, Aug 21, 2008 at 11:48 PM, Porntip Chitwirat
<p_chitwirat@...> wrote:

> Dear All.
>
> Now I work for a machine translation from English to Thai by using HPSG with
> the LKB system.
> But I still problem about how to show thai's fonts in LKB. I try to set
> emacs from this link http://wiki.delph-in.net/moin/LkbEmacs but it can't
> read thai yet . If anyone have any suggestions please help me.
>
> Best regards,
>
> Porntip Chitwirat.
>
>
> ________________________________
> See what people are saying about Windows Live. Check out featured posts.
> Check It Out!

Re: How to display thai fonts in LKB

2008-08-26T18:30:23Z

G'day,

just a quick follow up, Korean (and of course Japanese) works in both
for me now.

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

ko-trees-pango.png (42K) Download Attachment
ko-trees-clim.png (17K) Download Attachment

Fwd: How to display thai fonts in LKB

2008-08-26T18:22:08Z

G'day,

> Now I work for a machine translation from English to Thai by using HPSG with
> the LKB system.
> But I still problem about how to show thai's fonts in LKB. I try to set
> emacs from this link http://wiki.delph-in.net/moin/LkbEmacs but it can't
> read thai yet . If anyone have any suggestions please help me.

I could get Thai to dsiplay with Pavel's pango LUI, see the discussion
of alternative pango lui implementations at:
http://wiki.delph-in.net/moin/LkbLui

I couldn't get the display working with CLIM (the LKB default,
although I didn't try so hard).

I attach pictures with LUI and CLIM (note it is just one Thai word in
a Japanese grammar).

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

thai-trees.png (54K) Download Attachment
thai-trees-clim.png (18K) Download Attachment

How to display thai fonts in LKB

2008-08-21T11:18:35Z

Dear All.

Now I work for a machine translation from English to Thai by using HPSG with the LKB system.
But I still problem about how to show thai's fonts in LKB. I try to set emacs from this link http://wiki.delph-in.net/moin/LkbEmacs but it can't read thai yet . If anyone have any suggestions please help me.

Best regards,

Porntip Chitwirat.

See what people are saying about Windows Live. Check out featured posts. Check It Out!

Re: single-quotes in lexicon entries

2008-08-05T03:43:28Z

hi mike, apologies for a tardy follow-up!

> lunghezza_d'onda_3 := sg-fem-common-noun-lex &
> [ STEM < "lunghezza", "d'onda" >,
> SYNSEM.LKEYS.KEYREL.PRED "_lunghezza+d'onda_n_3_rel" ].
>
> Since single-quotes are meaningful characters in Lisp, I imagine they
> must be replaced with something else.

well, yes and no. the syntax of TDL is the constraining factor here,
not Lisp, strictly speaking. but the single quote is a TDL operator,
so |lunghezza_d'onda_3| is not a valid identifier (using the pipes as
a kind of `quote mark' here). |"_lunghezza+d'onda_n_3_rel"|, on the
other hand, is a valid TDL string, hence for the PRED value you could
actually go with that. which might be nice, as we usually expect the
PRED to reflect the citation form (in `common' orthography). as for
the identifier (of the lexical entry, in this case), there are a few
more characters, besides |+|, that are available for TDL identifiers,
e.g. |-|, |_|, or |*|. these identifiers are fully grammar-internal,
so you should be free to invent a convention that makes sense here.

all best - oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++ CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++ --- oe@...; oe@...; stephan@... ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Re: single-quotes in lexicon entries

2008-07-25T07:14:12Z

G'day,

> Is there any convention for replacing single-quotes in predicate names
> and lexeme identifiers? For example, I have the Italian compound
> "lunghezza d'onda" ("wavelength"), and here's my current lexical
> entry:
>
> lunghezza_d'onda_3 := sg-fem-common-noun-lex &
> [ STEM < "lunghezza", "d'onda" >,
> SYNSEM.LKEYS.KEYREL.PRED "_lunghezza+d'onda_n_3_rel" ].
>
> Since single-quotes are meaningful characters in Lisp, I imagine they
> must be replaced with something else. + is already in use for spaces,
> so I imagine that wouldn't be a good idea.
>
> If we have an agreed-upon convention, I'll add it to the RMRS wiki
> (http://wiki.delph-in.net/moin/RmrsPos).

I don't think we have an agreed-upon convention. I would propose just
deleting them (and if it causes ambiguity just add a -1, -2 on the
end), but am open to cleverer suggestions.

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

single-quotes in lexicon entries

2008-07-24T22:48:36Z

Hello,

Is there any convention for replacing single-quotes in predicate names
and lexeme identifiers? For example, I have the Italian compound
"lunghezza d'onda" ("wavelength"), and here's my current lexical
entry:

lunghezza_d'onda_3 := sg-fem-common-noun-lex &
[ STEM < "lunghezza", "d'onda" >,
SYNSEM.LKEYS.KEYREL.PRED "_lunghezza+d'onda_n_3_rel" ].

Since single-quotes are meaningful characters in Lisp, I imagine they
must be replaced with something else. + is already in use for spaces,
so I imagine that wouldn't be a good idea.

If we have an agreed-upon convention, I'll add it to the RMRS wiki
(http://wiki.delph-in.net/moin/RmrsPos).

Thanks

--
-Michael Wayne Goodman

Re: Add a ranking module to LKB

2008-07-18T15:46:12Z

hi norman (and francis),

> > I want to add a ranking module to select the "best" scoped MRS
> > output. I'm not sure whether someone have done similar things
> > before.

[...]

> There is code that selects the best parse based on the same model used
> by pet (i.e. looking at the derivation tree).

there was a fair amount of MaxEnt parse selection work (using the ERG
and other DELPH-IN grammars). the code to rank LKB results is part of
the [incr tsdb()] add-on system. regrettably (unlike the LKB proper),
[incr tsdb()] is not supported in SBCL.

however, all previous parse selection work in this context (for all i
know) was disambiguating syntactic ambiguity, i.e. derivations trees
returned by the parser (or generator). hence, the problem of ranking
among the (potentially numerous) scope-resolved MRSs corresponding to
a single derivation remains unsolved at large. i would expect that a
discriminative model, using structural semantic features, might be a
suitable approach, but it would require scope-resolved training data
(which is not part of the standard Redwoods treebanks) ... i guess i
would view this as a potentially rewarding but not trivial project.

all best - oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++ CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++ --- oe@...; oe@...; stephan@... ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Re: Add a ranking module to LKB

2008-07-16T06:56:08Z

G'day,

> I want to add a ranking module to select the "best" scoped MRS output. I'm
> not sure whether someone have done similar things before.

There is code that selects the best parse based on the same model used
by pet (i.e. looking at the derivation tree).
Take a look in (tsdb::parse-item). You would have to ask Stephan
Oepen for more details.

> And, I want to use SBCL for this development, but there is little
> information about using LKB with SBCL -- I read through the mail-list
> archive and most pages in delph-in.net.
> Any advice?

Ben Waldron did some work on this --- enough so that the delphin
branch of the lkb will run using sbcl.

For example I can call the following:
sbcl --dynamic-space-size 2000 --load make-qc.lisp

and it works.

There are some errors doing this using the logon branch.

Ben also had some patches for [incr() tsdb] which allowed you to do
more interesting batch processing using sbcl (such as running the
generator server) but they have not been merged upstream.

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

P.S. make-qc.lisp
;;;
;;; call with
;;; time sbcl --dynamic-space-size 2000 --load make-qc.lisp
;;; time env DISPLAY="" /home/bond/logon/franz/linux.x86.64/alisp -L
make-qc.lisp
;;;
;;;
;;; use more memory
;(setf (sb-ext:BYTES-CONSED-BETWEEN-GCS) 200000000)

(load "/home/bond/delphin/lkb/src/general/loadup")
;(load "/home/bond/delphin/lkb.cvs.good/src/general/loadup")
;(compile-system "lkb" :force t)
(compile-system "lkb")
(defparameter grm-dir "/home/bond/logon/dfki/jacy")
;;;
;;; load jacy, with tdl lexicon
;;;
;(setf *features* (remove :psql *features*))
(read-script-file-aux
(format nil "~a/lkb/script" grm-dir))
(setf lkb::*maximum-number-of-edges* '5000)
;;;
;;; create the check-path
;;;
(lkb::with-check-path-list-collection
"/tmp/checkpaths"
(lkb::parse-sentences
(format nil "~a/testsuites/kinou1.chasen.500" grm-dir)
(format nil "/tmp/kinou1.chasen.out" grm-dir)))

(format t "~%All Done!~%")
(quit)

Re: Add a ranking module to LKB

2008-07-16T06:54:17Z

G'day,

> I want to add a ranking module to select the "best" scoped MRS output. I'm
> not sure whether someone have done similar things before.

There is code that selects the best parse based on the same model used
by pet (i.e. looking at the derivation tree).
Take a look in (tsdb::parse-item). You would have to ask Stephan
Oepen for more details.

> And, I want to use SBCL for this development, but there is little
> information about using LKB with SBCL -- I read through the mail-list
> archive and most pages in delph-in.net.
> Any advice?

Ben Waldron did some work on this --- enough so that the delphin
branch of the lkb will run using sbcl.

For example I can call the following:
sbcl --dynamic-space-size 2000 --load make-qc.lisp

and it works. There are some errors doing this using the logon branch.

Ben also had some patches for [incr() tsdb], which allowed you to do
more interesting batch processing using sbcl, (including running the
generator server) but they have not been merged upstream.

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

P.S. make-qc.lisp
;;;
;;; call with
;;; time sbcl --dynamic-space-size 2000 --load make-qc.lisp
;;; time env DISPLAY="" /home/bond/logon/franz/linux.x86.64/alisp -L
make-qc.lisp
;;;
;;;
;;; use more memory
;(setf (sb-ext:BYTES-CONSED-BETWEEN-GCS) 200000000)

(load "/home/bond/delphin/lkb/src/general/loadup")
;(load "/home/bond/delphin/lkb.cvs.good/src/general/loadup")
;(compile-system "lkb" :force t)
(compile-system "lkb")
(defparameter grm-dir "/home/bond/logon/dfki/jacy")
;;;
;;; load jacy, with tdl lexicon
;;;
;(setf *features* (remove :psql *features*))
(read-script-file-aux
(format nil "~a/lkb/script" grm-dir))
(setf lkb::*maximum-number-of-edges* '5000)
;;;
;;; create the check-path
;;;
(lkb::with-check-path-list-collection
"/tmp/checkpaths"
(lkb::parse-sentences
(format nil "~a/testsuites/kinou1.chasen.500" grm-dir)
(format nil "/tmp/kinou1.chasen.out" grm-dir)))

(format t "~%All Done!~%")
(quit)

Add a ranking module to LKB

2008-07-07T21:50:59Z

Hi ALL,

I want to add a ranking module to select the "best" scoped MRS output. I'm not sure whether someone have done similar things before.
And, I want to use SBCL for this development, but there is little information about using LKB with SBCL -- I read through the mail-list archive and most pages in delph-in.net.
Any advice?

Thanks,
Norman

Morphological preprocessors

2008-04-09T17:03:09Z

Hello,

I have a student working on an LKB grammar for Lillooet,
and now additionally working on a morphological analyzer
(initially in XFST) for that language. We'd like to know
what the state of the art is for connecting external
preprocessors to the LKB, as well as for going the other
direction and sending LKB generation output through the
FST to produce surface forms.

Thanks,
Emily

Re: Interactive unification in LUI

2008-01-14T19:56:49Z

Thanks, Francis! I've linked that page (and a few other similar) into the
LkbLui page on the wiki, which I take to be the equivalent of LuiTop.

Emily

On Jan 14, 2008 5:46 PM, Francis Bond <fcbond@...> wrote:

> G'day,
>
> > How does one access interactive unification in LUI?
> > We have been able to find "select", but not "unify".
>
> I normally drag-and-drop --- click on the thing you want to unify,
> drag it to the thing you want to unify with and then release the
> mouse.
>
> It appears to be documented here:
> http://wiki.delph-in.net/moin/LuiUnification
>
> --
> Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
> NICT Computational Linguistics Group
>

Re: Interactive unification in LUI

2008-01-14T17:46:07Z

G'day,

> How does one access interactive unification in LUI?
> We have been able to find "select", but not "unify".

I normally drag-and-drop --- click on the thing you want to unify,
drag it to the thing you want to unify with and then release the
mouse.

It appears to be documented here:
http://wiki.delph-in.net/moin/LuiUnification

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Computational Linguistics Group

Interactive unification in LUI

2008-01-14T17:14:57Z

Hello,

How does one access interactive unification in LUI?
We have been able to find "select", but not "unify".

Thanks,
Emily

Defeasible constraints

2007-12-13T15:21:44Z

Hi folks,

The lexical rule hierarchy in the Matrix is getting
sufficiently unwieldy to the point where I'm tempted
to try defeasible constraints, much as I have been
hesitant to in the past. Is there any current work
exercising that part of the LKB that I could look at
for reference?

Thanks,
Emily

Re: Redundancy detection

2007-12-12T08:37:15Z

Much better, thanks :) And it does still give me the "redundancy
involving <TYPE>" message, which is all I really need.

Emily

On Tue, Dec 11, 2007 at 04:18:22PM +0000, Ann Copestake wrote:

>
> ebender@... said:
> > Is it possible to turn off the functionality that discovers where exactly the
> > redundancy is after an error message like "Redundancy involving IMP-AUX" is
> > printed? With my current grammar at least, there seems to be a lot of
> > gargbage collecting involved in figuring out what exactly the redundancy is;
> > I can usually spot it way faster than the machine can.
>
> hmm - I just looked at the code that does that - the inefficiency is clear.
> If it's annoying you, the easiest thing to do would just be to put in a patch
> e.g., in your user-fns file that redefines the function e.g.,
>
> (defun find-all-redundancies nil
> (format t "Redundancy somewhere but I'm not going to tell you where!"))
>
> I don't really want to put in a global variable that switches the check off
> because it's just one more thing to (not) document and I believe that
> find-all-redundancies can be rewritten. I will have a look at it sometime.
>
> Ann
>
>

Re: Redundancy detection

2007-12-11T08:18:22Z

ebender@... said:
> Is it possible to turn off the functionality that discovers where exactly the
> redundancy is after an error message like "Redundancy involving IMP-AUX" is
> printed? With my current grammar at least, there seems to be a lot of
> gargbage collecting involved in figuring out what exactly the redundancy is;
> I can usually spot it way faster than the machine can.

hmm - I just looked at the code that does that - the inefficiency is clear.
If it's annoying you, the easiest thing to do would just be to put in a patch
e.g., in your user-fns file that redefines the function e.g.,

(defun find-all-redundancies nil
(format t "Redundancy somewhere but I'm not going to tell you where!"))

I don't really want to put in a global variable that switches the check off
because it's just one more thing to (not) document and I believe that
find-all-redundancies can be rewritten. I will have a look at it sometime.

Ann

Redundancy detection

2007-12-10T16:40:58Z

Hi folks,

Is it possible to turn off the functionality that
discovers where exactly the redundancy is after
an error message like "Redundancy involving IMP-AUX"
is printed? With my current grammar at least, there
seems to be a lot of gargbage collecting involved
in figuring out what exactly the redundancy is; I
can usually spot it way faster than the machine can.

Emily

Re: Morphology question

2007-12-05T09:44:30Z

Ah, indeed they aren't. (And I could have *sworn* I'd seen it go
interactively.) I've fixed the grammar bug and now things are going
as expected.

--- I hadn't realized how efficient that rule filter is!

Thanks,
Emily

On Wed, Dec 05, 2007 at 12:39:18PM +0000, Ann Copestake wrote:
>
> can you confirm that the rules are allowed to feed eachother as far as the
> morphosyntax goes? try (print-nospfeeding)
>
>
>

Re: Morphology question

2007-12-05T04:39:18Z

can you confirm that the rules are allowed to feed eachother as far as the
morphosyntax goes? try (print-nospfeeding)

Morphology question

2007-12-04T17:39:32Z

Hi,

I'm looking for some help on a puzzle that's come up
with my Wambaya grammar. I have a word form that
as near as I can tell ought to be analyzed by the
orthographic subrules, but is failing to be. I know
that each of the rules fires independently (see below),
and the form in question should be using just the simplest
subrule for each.

Thanks,
Emily

The word form is wurlungguji. Here are the relevant
rules and the relevant lexical entry:

3du-s-aux := third-du-aux &
[ STEM < "wurlu" > ].

rr :=
%suffix (* ngg)
refl-recip-lex-rule.

non-sg-subj-nact-pres :=
%suffix (* uji) (!v uji) (giny gunyuji) (gininy gununyuji) (ngiyiny nguyunyuji) (mirndiny murndunyuji) (nying nyunguji) (ging gunguji) (gining gununguji) (ngiying nguyunguji) (mirnding murndunguji)
non-sg-subj-nact-present-lex-rule.

Here's what I get from the LKB:

TSNLP(11): (lkb::do-parse-tty "wurlungguji")

No analysis found corresponding to token 0-1 WURLUNGGUJI
No parses found
0
0
-1
0
46
TSNLP(12): lkb::*tchart*
#2A((NIL ([#[Token edge # 1: "WURLUNGGUJI" 0 1 (0 c 11) ] NIL]))
(([#[Token edge # 1: "WURLUNGGUJI" 0 1 (0 c 11) ] NIL]) NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
...)
TSNLP(13):

It can parse the root with each suffix separately:

TSNLP(13): (lkb::do-parse-tty "wurlungga")

No parses found
29
3
-1
41
33
TSNLP(14): lkb::*tchart*
#2A((NIL
([#[Morph edge # 5: "WURLUNGGA" "WURLU" 0 1 ((IV-GEN-NON-ABS WURLU)
(NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 4: "WURLUNGGA" "WURLU" 0 1 ((NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 3: "WURLUNGGA" "WURLU" 0 1 ((ADJ-NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 2: "WURLUNGGA" "WURLU" 0 1 ((RR WURLUNGG)
(NON-FUT-WITH-OBJ
WURLUNGGA)) (0 c 9) ] NIL]
[#[Token edge # 1: "WURLUNGGA" 0 1 (0 c 9) ] NIL]))
(([#[Morph edge # 5: "WURLUNGGA" "WURLU" 0 1 ((IV-GEN-NON-ABS WURLU)
(NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 4: "WURLUNGGA" "WURLU" 0 1 ((NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 3: "WURLUNGGA" "WURLU" 0 1 ((ADJ-NON-DU-DAT
WURLUNGGA)) (0 c 9) ] NIL]
[#[Morph edge # 2: "WURLUNGGA" "WURLU" 0 1 ((RR WURLUNGG)
(NON-FUT-WITH-OBJ
WURLUNGGA)) (0 c 9) ] NIL]
[#[Token edge # 1: "WURLUNGGA" 0 1 (0 c 9) ] NIL])
NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
...)
TSNLP(15): (lkb::do-parse-tty "wurluji")

No parses found
59
6
-1
32
46
TSNLP(16): lkb::*tchart*
#2A((NIL
([#[Morph edge # 8: "WURLUJI" "WURLU" 0 1 ((IV-GEN-NON-ABS WURLU)
(DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 7: "WURLUJI" "WURLU" 0 1 ((NON-SG-SUBJ-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 6: "WURLUJI" "WURLU" 0 1 ((3SM-AGT-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 5: "WURLUJI" "WURLU" 0 1 ((DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 4: "WURLUJI" "WURLU" 0 1 ((ADJ-DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 3: "WURLUJI" "WURLA" 0 1 ((NON-SG-SUBJ-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 2: "WURLUJI" "WURLA" 0 1 ((3SM-AGT-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Token edge # 1: "WURLUJI" 0 1 (0 c 7) ] NIL]))
(([#[Morph edge # 8: "WURLUJI" "WURLU" 0 1 ((IV-GEN-NON-ABS WURLU)
(DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 7: "WURLUJI" "WURLU" 0 1 ((NON-SG-SUBJ-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 6: "WURLUJI" "WURLU" 0 1 ((3SM-AGT-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 5: "WURLUJI" "WURLU" 0 1 ((DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 4: "WURLUJI" "WURLU" 0 1 ((ADJ-DU-ERG
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 3: "WURLUJI" "WURLA" 0 1 ((NON-SG-SUBJ-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Morph edge # 2: "WURLUJI" "WURLA" 0 1 ((3SM-AGT-NACT-PRES
WURLUJI)) (0 c 7) ] NIL]
[#[Token edge # 1: "WURLUJI" 0 1 (0 c 7) ] NIL])
NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
(NIL NIL)
...)
TSNLP(17):

Re: MRS v. syntactic discriminants

2007-09-16T16:15:32Z

G'day,

ones back. Is there a switch for this?
>
> The global variable you want to reset is as follows:
>
> (setf lkb::*tree-discriminants-mode* :classic)

Would it be possible to change at this early stage to a naming scheme
that doesn't require new users to be knowledgable about the history of
the function in question, and call it something like (setf
lkb::*tree-discriminants-mode* :syntactic)? Or if we planning to
replace it soon, then :syntactic-clim or something?

--
Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Computational Linguistics Group

Re: MRS v. syntactic discriminants

2007-09-16T08:30:21Z

> When I invoke "compare" from the LKB in the logon
> distribution (*cvs-version*: "$Date: 2007/04/02 20:43:45 $"),
> it seems to be doing semantic discriminants. That's nifty,
> but for my present purposes, I'd rather have the syntactic
> ones back. Is there a switch for this?

The global variable you want to reset is as follows:

(setf lkb::*tree-discriminants-mode* :classic)
Cheers,

Dan

MRS v. syntactic discriminants

2007-09-15T17:40:03Z

Hi,

When I invoke "compare" from the LKB in the logon
distribution (*cvs-version*: "$Date: 2007/04/02 20:43:45 $"),
it seems to be doing semantic discriminants. That's nifty,
but for my present purposes, I'd rather have the syntactic
ones back. Is there a switch for this?

Thanks,
Emily

Arboretum

2007-09-15T04:28:53Z

Hi,
I'd like to implement arboretum (robust parsing) in my Lisp parsing
program. When I type (lkb::grammar-check "dogs barks"), the message
shows "The sentence is grammatical". Why not error message such as
"There is a verb which does not agree with its subject." Hopefully
someone may help me.

Kindly regards,
Halizah
Univ. of Otago
New Zealand

Re: Debugging trigger rules

2007-08-17T10:30:24Z

Thanks, Dan. The EQUAL trick worked.

Emily

On Fri, Aug 17, 2007 at 12:33:53AM -0700, Dan Flickinger wrote:

>
> > Here is an example trigger rule:
> >
> > ngurlu-trigger-rule := generator-rule &
> > [ CONTEXT.RELS <! [ ARG1.PNG.PN [ PER excl,
> > NUM du ]] !>,
> > FLAGS [ TRIGGER "1du-excl-s-aux" ]].
>
> While I'm not certain of the properties of the MRS-matching step regarding
> access to type constraints which would ensure that ARG1 is incompatible
> with any of your EPs (like maybe the verb's, if you didn't say that its
> EP is an arg1-only-relation), you can avoid this indeterminacy by further
> constraining your rule to insist that the ARG1 be actually instantiated
> in the input MRS. This requirement can be expressed by adding the ARG1
> variable to the FLAGS.EQUAL list of the rule, as follows:
>
> ngurlu-trigger-rule := generator-rule &
> [ CONTEXT.RELS <! [ ARG1 #x1 &
> [ PNG.PN [ PER excl,
> NUM du ]] !>,
> FLAGS [ TRIGGER "1du-excl-s-aux",
> EQUAL < #x1 > ]].
>
> You might for safety also declare the type of that ARG1 to be a referential
> index, in order to avoid another source of spurious matching in an input
> MRS where you have an EP with an ARG1 which is a handle or an event.
>
> Dan

Re: Debugging trigger rules

2007-08-17T01:33:53Z

> Here is an example trigger rule:
>
> ngurlu-trigger-rule := generator-rule &
> [ CONTEXT.RELS <! [ ARG1.PNG.PN [ PER excl,
> NUM du ]] !>,
> FLAGS [ TRIGGER "1du-excl-s-aux" ]].

While I'm not certain of the properties of the MRS-matching step regarding
access to type constraints which would ensure that ARG1 is incompatible
with any of your EPs (like maybe the verb's, if you didn't say that its
EP is an arg1-only-relation), you can avoid this indeterminacy by further
constraining your rule to insist that the ARG1 be actually instantiated
in the input MRS. This requirement can be expressed by adding the ARG1
variable to the FLAGS.EQUAL list of the rule, as follows:

ngurlu-trigger-rule := generator-rule &
[ CONTEXT.RELS <! [ ARG1 #x1 &
[ PNG.PN [ PER excl,
NUM du ]] !>,
FLAGS [ TRIGGER "1du-excl-s-aux",
EQUAL < #x1 > ]].

You might for safety also declare the type of that ARG1 to be a referential
index, in order to avoid another source of spurious matching in an input
MRS where you have an EP with an ARG1 which is a handle or an event.

Dan