[developers] [ACE] [JACY] neg_x cfrom cto = -1. is this a bug?
Michael Wayne Goodman
goodmami at uw.edu
Fri Jul 28 21:17:41 CEST 2017
The coref problem was addressed in a commit on May 2nd, but that would only
fix the problem of loading with the LKB, not the CFROM/CTO issue.
But when I try and parse with the LKB, I see the expected CFROM/CTO values,
but with ACE I see <-1:-1>:
On Fri, Jul 28, 2017 at 9:52 AM, Tuấn Anh Lê <tuananh.ke at gmail.com> wrote:
> I got that from the github page https://github.com/delph-in/jacy
>
> The last commit was in May 4th.
>
> commit 7aba30accc0094c41218cdbe915ded03909e0874
> Merge: d204238 abd3425
> Author: Francis Bond <bond at ieee.org>
> Date: Thu May 4 12:44:33 2017 -0700
>
> Merge branch 'master' of https://github.com/delph-in/jacy
>
> commit d2042386c7bcb824685fc1edc1ee1c5a721b1aa1
> Author: Francis Bond <bond at ieee.org>
> Date: Wed May 3 13:47:51 2017 -0700
>
> Fixed niyotte-case-chasen
>
>
> I recompiled the grammar using that version. It's still the same.
>
> [tuananh at localhost jacy]$ echo "猫 が 好き じゃ ない 。" | ace -g jacy.dat -1Tf
> SENT: 猫 が 好き じゃ ない 。
> [ LTOP: h0
> INDEX: e2 [ e TENSE: pres MOOD: indicative PROG: - PERF: - ASPECT:
> default_aspect PASS: - SF: prop ]
> RELS: < [ udef_q_rel<0:1> LBL: h4 ARG0: x5 [ x PERS: 3 ] RSTR: h6 BODY: h7
> ]
> [ "_neko_n_rel"<0:1> LBL: h8 ARG0: x5 ]
> [ "_suki_a_1_rel"<4:6> LBL: h9 ARG0: e2 ARG1: i10 ARG2: x5 ]
> [ neg_x_rel<-1:-1> LBL: h1 ARG0: e11 [ e TENSE: tense MOOD: indicative
> PROG: - PERF: - ASPECT: default_aspect PASS: - SF: prop ] ARG1: h12 ] >
> HCONS: < h0 qeq h1 h6 qeq h8 h12 qeq h9 > ]
> NOTE: 1 readings, added 371 / 130 edges to chart (62 fully instantiated,
> 50 actives used, 28 passives used) RAM: 1835k
>
>
> NOTE: parsed 1 / 1 sentences, avg 1835k, time 0.01112s
>
>
> On 29 July 2017 at 00:44, Michael Wayne Goodman <goodmami at uw.edu> wrote:
>
>> Tuấn Anh, when did you get your version of Jacy? Francis and I noticed
>> errors like that when there was a solitary coreference #tag in an AVM (ACE
>> didn't mind as much as the LKB), but I thought we'd fixed them all. Make
>> sure your version of Jacy is up to date and try again?
>>
>> On Jul 28, 2017 09:01, "Tuấn Anh Lê" <tuananh.ke at gmail.com> wrote:
>>
>>> Hi Woodley, thanks for answering. I can't seem to load the grammar on my
>>> LKB, but logon has never really worked on my laptop for some reasons so may
>>> be it's my machine (it does load ERG though). I don't know how to use PET
>>> so I can't talk for it.
>>>
>>> When I load JACY on my laptop, there is this error message that shows up
>>> every time.
>>>
>>> Reading in type file fundamentals
>>> [23:59:36] gc-after-hook(): {L#30 N=1.2M O=0 E=97%} [S=2.2G R=150M].
>>>
>>> WARNING: Type `EXTRACTED-ADJ-PHRASE' redefined.
>>> WARNING: Type `BASIC-HEAD-FILLER-PHRASE' redefined.
>>> WARNING: Type `GAP' redefined.
>>> WARNING: Type `CONJ-REF-IND' redefined.
>>> WARNING: Type `GENERIC_ENTITY_REL' redefined.
>>> Reading in type file rule-types
>>> Reading in type file principles
>>> Reading in type file letypes
>>> [23:59:37] gc-after-hook(): {L#31 N=3.3M O=0 E=97%} [S=2.2G R=277M].
>>>
>>> Syntax error at position 437529: Coreference NEG_X_REL only used once
>>> Reading in type file mtr
>>> Reading in type file tmt
>>> Error: Syntax error(s) in type file
>>>
>>> Restart actions (select using :continue):
>>> 0: Cancel load
>>> 1: retry the load of /home/tuananh/workspace/jacy/lkb/script.common
>>> 2: skip loading /home/tuananh/workspace/jacy/lkb/script.common
>>> 3: retry the load of ~/wk/jacy/lkb/script
>>> 4: skip loading ~/wk/jacy/lkb/script
>>> 5: Return to Top Level (an "abort" restart).
>>> 6: Abort entirely from this (lisp) process.
>>>
>>> [changing package from "TSDB" to "LKB"]
>>> [1c] LKB(3):
>>>
>>>
>>>
>>> On 28 July 2017 at 23:27, Woodley Packard <sweaglesw at sweaglesw.org>
>>> wrote:
>>>
>>>> Hi Tuấn Anh,
>>>>
>>>> This can happen if the CFROM/CTO attributes in the AVM are never
>>>> specialized to specific values like "4" or "13". In this case, it appears
>>>> to me that the lexeme "janai-cop-stem" does not do this. Do you observe
>>>> different behavior from LKB or PET?
>>>>
>>>> Best of luck,
>>>> Woodley
>>>>
>>>>
>>>> On Jul 27, 2017, at 11:38 PM, Tuấn Anh Lê <tuananh.ke at gmail.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Is there anyone working with JACY and ACE. Sometimes cfrom and cto of
>>>> predicates are missing. For example, when I parse these two sentences:
>>>>
>>>>
>>>> - 猫 が 好き じゃ ない 。
>>>> - 猫 が あり ませ ん 。
>>>>
>>>> Both will produce a neg_x relation, but the neg_x one in the first
>>>> sentence doesn't have cfrom and cto.
>>>>
>>>>
>>>> -
>>>>
>>>> [ TOP: h0 RELS: < [ udef_q_rel<0:1> LBL: h1 ARG0: x5 [ x PERS: 3 ] RSTR: h8 ]
>>>> [ _neko_n_rel<0:1> LBL: h2 ARG0: x5 ]
>>>> [ _suki_a_1_rel<4:6> LBL: h3 ARG0: e6 [ e TENSE: pres PROG: - PASS: - PERF: - ASPECT: default_aspect MOOD: indicative SF: prop ] ARG2: x5 ]
>>>> [ *neg_x_rel<-1:-1>* LBL: h4 ARG0: e7 [ e TENSE: tense PROG: - PASS: - PERF: - ASPECT: default_aspect MOOD: indicative SF: prop ] ARG1: h9 ] > HCONS: < h0 qeq h4 h8 qeq h2 h9 qeq h3 > ]
>>>>
>>>> -
>>>>
>>>> [ TOP: h0 RELS: < [ udef_q_rel<0:1> LBL: h1 ARG0: x5 [ x PERS: 3 ] RSTR: h8 ]
>>>> [ _neko_n_rel<0:1> LBL: h2 ARG0: x5 ]
>>>> [ _exist_v_rel<4:6> LBL: h3 ARG0: e6 [ e TENSE: pres PROG: - PASS: - PERF: - ASPECT: default_aspect MOOD: indicative SF: prop ] ARG1: x5 ]
>>>> [ neg_x_rel<7:11> LBL: h4 ARG0: e7 [ e TENSE: tense PROG: - PASS: - PERF: - ASPECT: default_aspect MOOD: indicative SF: prop ] ARG1: h9 ] > HCONS: < h0 qeq h4 h8 qeq h2 h9 qeq h3 > ]
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>> --
>>> Yours,
>>> --
>>> Tuan Anh Le
>>>
>>
>
>
> --
> Yours,
> --
> Tuan Anh Le
>
--
Michael Wayne Goodman
Ph.D. Candidate, UW Linguistics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.delph-in.net/archives/developers/attachments/20170728/c34b8f1e/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: neko-ga-suki-jya-nai.png
Type: image/png
Size: 25912 bytes
Desc: not available
URL: <http://lists.delph-in.net/archives/developers/attachments/20170728/c34b8f1e/attachment-0001.png>
More information about the developers
mailing list