Received-SPF: pass (google.com: domain of durka42@gmail.com designates 2607:f8b0:400d:c04::22f as permitted sender) client-ip=2607:f8b0:400d:c04::22f;
Date: Wed, 8 Apr 2015 11:15:03 -0400
From: Alex Burka <durka42@gmail.com>
To: bpfk-list@googlegroups.com
Message-ID: <166E1503C5E24261B88E4F7B41741A53@gmail.com>
In-Reply-To: <CAO7bV+g3dr6p-1foeJWNHMmOnP+mYb5XVqwpFWGqjhEejfBo-Q@mail.gmail.com>
References: <CAO7bV+g3dr6p-1foeJWNHMmOnP+mYb5XVqwpFWGqjhEejfBo-Q@mail.gmail.com>
Subject: Re: [bpfk] te sumti detection using PEG
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="552545f7_431bd7b7_bb09"
Reply-To: bpfk-list@googlegroups.com
Precedence: list
Mailing-list: list bpfk-list@googlegroups.com; contact bpfk-list+owners@googlegroups.com
Sender: bpfk-list@googlegroups.com
X-Spam_score: -1.7
X-Spam_score_int: -16
X-Spam_bar: -

--552545f7_431bd7b7_bb09
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

I'm impressed that you got it this far, but as I've said before I really do=
n't see the PEG as the place for this. First of all, it's mixing two separa=
te steps in the interpretation of a sentence (namely parsing and sumti plac=
e resolution). And as you said, this basic functionality requires 60 new ru=
les ... and it has severe limitations as a te sumti detector, since it give=
s up after the first explicit FA, doesn't interact with SE/GIhE/BE/JAI, etc=
. Like I said, good proof of concept, but I'd be surprised if this is the r=
oute to a general te sumti detector. =20

mu'o mi'e durkavore =20


On Wednesday, April 8, 2015 at 8:50 AM, Gleki Arxokuna wrote:

> Terminology:
> *FAM - a term taking a FA-position with FA explicitely filled with {faxiv=
eimo'eko'a} where mo'eko'a is a precise number (e.g. not {xo'e}
> * ZAM - a bare term taking a FA-position with FA omitted. Positioning rul=
es can restore exact value of ko'a in {faxiveimo'eko'a}
> * BAM - all other terms, e.g. prefixed with BAI or PU etc.
> =20
> So the issue of te sumti detection is to turn all ZAMs into FAMs in the s=
yntax tree.
> Can we do that using PEG? I'm not that sure because
> some brivla have infinite number of places like e.g. {jutsi}. {du} is a s=
pecial case since every te sumti of it can just take {faxixo'e} position.
> =20
> Currently I'm unaware of any possibilities for remembering values of vari=
ables (te sumti numbers) in PEG.js thus we cant increment to any given numb=
er of te sumti without first hardcoding all of them in PEG itself.
> =20
> However, if we limit ourselves to just 5 places and basic cases of omitti=
ng FA then we can do that using PEG.
> The current version of my fork of camxes.js produces these outputs:
> =20
> 1. ([FAXIPA mi] [CU {prami <FAXIRE do> VAU}]) =20
> 2. ([FAXIPA mi] [CU {djuno <fi do> VAU}]) =20
> 3. ([FAXIPA mi] [CU {djica <FAXIRE (=C2=B9lo [nu {<FAXIPA (=C2=B2lo plise=
 KU=C2=B2)> <cu (=C2=B2farlu [FAXIRE mi] [FAXICI {lo tricu KU}] VAU=C2=B2)>=
} KEI] KU=C2=B9)> VAU}]) =20
> =20
> FAXIPA, FAXIRE, FAXICI are restored FA.
> =20
> =20
> This is how sentence looks now in my PEG:
> sentence =3D expr:(
> &(terms_1ZAM CU_elidible selbri terms_1ZAM terms_1ZAM !terms_1ZAM) (terms=
fa bridi_tail_t1fefi) / /* mi klama do ti*/
> &(terms_1ZAM CU_elidible selbri terms_1ZAM) (termsfa bridi_tail_t1fe) / /=
* mi klama do*/
> &(terms_1ZAM terms_1ZAM CU_elidible selbri !terms_1ZAM) (termsfa termsfe =
bridi_tail_t1) / /* mi do klama*/
> &(terms_1ZAM terms_1ZAM CU_elidible selbri terms_1ZAM !terms_1ZAM) (terms=
fa termsfe bridi_tail_t1fi) / /* mi do klama ti*/
> &(terms_1ZAM terms_1ZAM terms_1ZAM CU_elidible selbri !terms_1ZAM) (terms=
fa termsfe termsfi bridi_tail_t1) / /* mi do ti klama*/
> &(terms_1ZAM CU_elidible selbri) (termsfa bridi_tail_t1) / /* mi klama*/
> terms? bridi_tail_t1 (joik_jek bridi_tail / joik_jek stag? KE_clause free=
* bridi_tail KEhE_elidible free*)*) {return _node("sentence", expr);}
> =20
> =20
> Examples for each case is shown in comments.
> This addition to PEG required hardcoding terms for fa,fe,fi separately an=
d every case of bridi tail like "selbri x2 x3", "selbri x2", "selbri x3". T=
o hardcode them one needs to also hardcode all inner rules until you meet s=
elbri at bridi_tail_3 level and till tense_modal for sumti (a modification =
of tense_modal where {fa} or {fe} or {fi} are hardcoded correspondingly).
> =20
> This required adding 60 new lines to PEG.
> Even for supporting these basic cases the work isn't done yet, optimizati=
ons of those copy-pasted strings might be possible.
> =20
> Anyway this is just a proof of concept.
> =20
> To test the current state of alta parser say "alta: mi prami do" on La Na=
xle page (http://vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/ilment=
ufa/ircbot/naxle.html). =20
> =20
> -- =20
> You received this message because you are subscribed to the Google Groups=
 "BPFK" group.
> To unsubscribe from this group and stop receiving emails from it, send an=
 email to bpfk-list+unsubscribe@googlegroups.com (mailto:bpfk-list+unsubscr=
ibe@googlegroups.com).
> To post to this group, send email to bpfk-list@googlegroups.com (mailto:b=
pfk-list@googlegroups.com).
> Visit this group at http://groups.google.com/group/bpfk-list.
> For more options, visit https://groups.google.com/d/optout.

--=20
You received this message because you are subscribed to the Google Groups "=
BPFK" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to bpfk-list+unsubscribe@googlegroups.com.
To post to this group, send email to bpfk-list@googlegroups.com.
Visit this group at http://groups.google.com/group/bpfk-list.
For more options, visit https://groups.google.com/d/optout.

--552545f7_431bd7b7_bb09
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


                <div>I'm impressed that you got it this far, but as I've sa=
id before I really don't see the PEG as the place for this. First of all, i=
t's mixing two separate steps in the interpretation of a sentence (namely p=
arsing and sumti place resolution). And as you said, this basic functionali=
ty requires 60 new rules ... and it has severe limitations as a te sumti de=
tector, since it gives up after the first explicit FA, doesn't interact wit=
h SE/GIhE/BE/JAI, etc. Like I said, good proof of concept, but I'd be surpr=
ised if this is the route to a general te sumti detector.
                </div><div><br></div><div>mu'o mi'e durkavore</div>
                <div></div>
                =20
                <p style=3D"color: #A0A0A8;">On Wednesday, April 8, 2015 at=
 8:50 AM, Gleki Arxokuna wrote:</p>
                <blockquote type=3D"cite" style=3D"border-left-style:solid;=
border-width:1px;margin-left:0px;padding-left:10px;">
                    <span><div><div><div dir=3D"ltr">Terminology:<div>*FAM =
- a term taking a FA-position with FA explicitely filled with {faxiveimo'ek=
o'a} where mo'eko'a is a precise number (e.g. not {xo'e}</div><div>* ZAM - =
a bare term taking a FA-position with FA omitted. Positioning rules can res=
tore exact value of ko'a in {faxiveimo'eko'a}</div><div>* BAM - all other t=
erms, e.g. prefixed with BAI or PU etc.<br></div><div><br></div><div>So the=
 issue of te sumti detection is to turn all ZAMs into FAMs in the syntax tr=
ee.</div><div>Can we do that using PEG? I'm not that sure because</div><div=
><b>some brivla have infinite number of places</b> like e.g. {jutsi}. {du} =
is a special case since every te sumti of it can just&nbsp;take&nbsp;{faxix=
o'e} position.</div><div><br></div><div>Currently I'm unaware of any possib=
ilities for remembering values of variables (te sumti numbers) in PEG.js th=
us we cant increment to any given number of te sumti without first hardcodi=
ng all of them in PEG itself.</div><div><br></div><div>However, if we limit=
 ourselves to just 5 places and basic cases of omitting FA then we can do t=
hat using PEG.</div><div>The current version of my fork of camxes.js produc=
es these outputs:</div><div><br></div><div>1. ([FAXIPA mi] [CU {prami &lt;F=
AXIRE do&gt; VAU}])&nbsp;<br>2. ([FAXIPA mi] [CU {djuno &lt;fi do&gt; VAU}]=
)&nbsp;<br>3. ([FAXIPA mi] [CU {djica &lt;FAXIRE (=C2=B9lo [nu {&lt;FAXIPA =
(=C2=B2lo plise KU=C2=B2)&gt; &lt;cu (=C2=B2farlu [FAXIRE mi] [FAXICI {lo t=
ricu KU}] VAU=C2=B2)&gt;} KEI] KU=C2=B9)&gt; VAU}])&nbsp;</div><div><br></d=
iv><div>FAXIPA,&nbsp;FAXIRE,&nbsp;FAXICI are restored FA.</div><div><br></d=
iv><div><br></div><div>This is how sentence looks now in my PEG:</div><div>=
<div>sentence =3D expr:(</div><div>&amp;(terms_1ZAM CU_elidible selbri term=
s_1ZAM terms_1ZAM !terms_1ZAM) (termsfa bridi_tail_t1fefi) / /* mi klama do=
 ti*/</div><div>&amp;(terms_1ZAM CU_elidible selbri terms_1ZAM) (termsfa br=
idi_tail_t1fe) / /* mi klama do*/</div><div>&amp;(terms_1ZAM terms_1ZAM CU_=
elidible selbri !terms_1ZAM) (termsfa termsfe bridi_tail_t1) / /* mi do kla=
ma*/</div><div>&amp;(terms_1ZAM terms_1ZAM CU_elidible selbri terms_1ZAM !t=
erms_1ZAM) (termsfa termsfe bridi_tail_t1fi) / /* mi do klama ti*/</div><di=
v>&amp;(terms_1ZAM terms_1ZAM terms_1ZAM CU_elidible selbri !terms_1ZAM) (t=
ermsfa termsfe termsfi bridi_tail_t1) / /* mi do ti klama*/</div><div>&amp;=
(terms_1ZAM CU_elidible selbri) (termsfa bridi_tail_t1) / /* mi klama*/</di=
v><div>terms? bridi_tail_t1 (joik_jek bridi_tail / joik_jek stag? KE_clause=
 free* bridi_tail KEhE_elidible free*)*) {return _node("sentence", expr);}<=
/div></div><div><br></div><div>Examples for each case is shown in comments.=
</div><div>This addition to PEG required hardcoding terms for fa,fe,fi sepa=
rately and every case of bridi tail like "selbri x2 x3", "selbri x2", "selb=
ri x3". To hardcode them one needs to also hardcode all inner rules until y=
ou meet selbri at&nbsp;bridi_tail_3 level and till&nbsp;tense_modal for sum=
ti (a modification of tense_modal where {fa} or {fe} or {fi} are hardcoded =
correspondingly).</div><div><br></div><div>This required adding 60 new line=
s to PEG.</div><div>Even for supporting these basic cases the work isn't do=
ne yet, optimizations of those copy-pasted strings might be possible.</div>=
<div><br></div><div>Anyway this is just a proof of concept.</div><div><br><=
/div><div>To test the current state of alta parser say "alta: mi prami do" =
on <a href=3D"http://vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/il=
mentufa/ircbot/naxle.html">La Naxle page</a>.</div></div>

<p></p>

-- <br>
You received this message because you are subscribed to the Google Groups "=
BPFK" group.<br>
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to <a href=3D"mailto:bpfk-list+unsubscribe@googlegroups.com">bpfk-list=
+unsubscribe@googlegroups.com</a>.<br>
To post to this group, send email to <a href=3D"mailto:bpfk-list@googlegrou=
ps.com">bpfk-list@googlegroups.com</a>.<br>
Visit this group at <a href=3D"http://groups.google.com/group/bpfk-list">ht=
tp://groups.google.com/group/bpfk-list</a>.<br>
For more options, visit <a href=3D"https://groups.google.com/d/optout">http=
s://groups.google.com/d/optout</a>.<br>
</div></div></span>
                =20
                =20
                =20
                =20
                </blockquote>
                =20
                <div>
                    <br>
                </div>
           =20

<p></p>

-- <br />
You received this message because you are subscribed to the Google Groups &=
quot;BPFK&quot; group.<br />
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to <a href=3D"mailto:bpfk-list+unsubscribe@googlegroups.com">bpfk-list=
+unsubscribe@googlegroups.com</a>.<br />
To post to this group, send email to <a href=3D"mailto:bpfk-list@googlegrou=
ps.com">bpfk-list@googlegroups.com</a>.<br />
Visit this group at <a href=3D"http://groups.google.com/group/bpfk-list">ht=
tp://groups.google.com/group/bpfk-list</a>.<br />
For more options, visit <a href=3D"https://groups.google.com/d/optout">http=
s://groups.google.com/d/optout</a>.<br />

--552545f7_431bd7b7_bb09--