text
stringlengths 22
301k
|
|---|
# Fetch the first child of given concept.
L("return_con") = down(L("con")); |
@CODE
G("kb") = getconcept(findroot(),"kb");
DisplayKB(L("top con"),L("display type"))
SaveKB("mykb.kbb",G("kb"),2);
@@CODE |
@NODES _LINE
@POST
# Add attr's to parent node
X("code") = N("$text", 4);
X("type") = N("$text", 8);
# Add code to kb
# L("type") = findconcept(G("codes"), N("$text", 8));
X("con") = AddUniqueCon(L("type"), N("$text", 1));
addstrval(X("con"), "code", N("$text", 4));
excise(2,9);
noop();
@RULES
_xNIL <-
_xWILD [fail=(code)] ### (1)
_xWILD [one matches=(code)] ### (2)
\= ### (3)
_xWILD [min=1 max=2 matches=(_xALPHA _xNUM)] ### (4)
_xWHITE ### (5)
_xWILD [one matches=(type)] ### (6)
\= ### (7)
_xANY ### (8)
_xWILD [one matches=(\n \r _xEND)] ### (9)
@@
# @POST
# X("con") = AddUniqueCon(G("codes"), N("$text", 5));
# if (N(2)) {
# addstrval(X("con"), "code", N("$text", 2));
# }
# noop();
#
# @RULES
# _xNIL <-
# _xSTART ### (1)
# _xNUM [opt] ### (2)
# \, [opt] ### (3)
# \" [opt] ### (4)
# _xWILD [fails=(\n \r \")] ### (5)
# @@
|
@NODES _ROOT
@POST
if (N("verb",3) && N("noun",3)) {
pnrename(N(3),"_noun");
}
@RULES
_xNIL <-
_det ### (1)
_adj [opt] ### (2)
_ambig ### (3)
@@
@POST
if (N("verb",2) && N("noun",2)) {
pnrename(N(2),"_noun");
}
@RULES
_xNIL <-
_adj ### (1)
_ambig ### (2)
@@
@POST
if (N("prep",1) && N("noun",1)) {
pnrename(N(1),"_prep");
}
@RULES
_xNIL <-
_ambig ### (1)
_det ### (2)
@@
|
@NODES _ROOT
@POST
excise(1,1);
noop();
@RULES
_xNIL <-
_xWHITE [s] ### (1)
@@
|
@CODE
G("icd9_codes") = getconcept(findroot(), "icd9_codes");
if (!G("icd9_codes")) {
G("icd9_codes") = makeconcept(findroot(), "icd9_codes");
}
rmchildren(G("icd9_codes"));
G("icd9_terms") = getconcept(findroot(), "icd9_terms");
if (!G("icd9_terms")) {
G("icd9_terms") = makeconcept(findroot(), "icd9_terms");
}
rmchildren(G("icd9_terms"));
@@CODE |
@NODES _textZone
@POST
S("language") = N("language",1);
single();
@RULES
_languageZone <-
_language ### (1)
_xWILD [plus match=(_headerZone)] ### (2)
@@
|
@NODES _ROOT
@POST
S("con") = MakeCountCon(G("emojis"),"group");
addstrval(S("con"),"description",N("description",1));
single();
@RULES
_group <-
_groupHeader ### (1)
_xWILD [fail=(_subgroupHeader _groupHeader)] ### (2)
@@
|
@CODE
L("hello") = 0;
@@CODE
#@PATH _ROOT _TEXTZONE _sent
@NODES _sent
@CHECK
if (!N("possessive",2))
fail();
@POST
group(1,2,"_det");
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_det
_adj
@@
# det num
@POST
if (N("num",2) == 1)
N("mypos",2) = "NN";
else
N("mypos",2) = "NNS";
pnrename(N(2),"_noun"); # num -> noun
@RULES
_xNIL <-
_det
_num
_xWILD [one lookahead fail=(_xALPHA _det _quan _num _xNUM
_adj _adjc _noun _fnword)]
@@
# det noun adj
# Note: caps can serve as "adjectives".
@CHECK
if (!N("cap",2) && !N("num",2) && !N("possessive",3))
fail();
@POST
if (!N("mypos",2))
chpos(N(2),"NN");
pnrename(N(2),"_adj"); # noun -> adj
@RULES
_xNIL <-
_det
_noun
_adj
@@
@POST
dqaninfo(0,0,0,2);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_noun
_xWILD [one match=( \, )]
@@
# dqan and dqan
# When there is local disagreement but overall agreement.
#@CHECK
# if (!N(1) && !N(2) && !N(3) && !N(4))
# fail();
#@POST
# N("pos50 dqan-and-np",6) = 1;
# if (N(1))
# listadd(6,1,"true");
# else if (N(2))
# listadd(6,2,"true");
# else if (N(3))
# listadd(6,3,"true");
# else
# listadd(6,4,"true");
#@RULES
#_xNIL <-
# _xWILD [star match=(_det _pro)]
# _xWILD [star match=(_quan _num _xNUM)]
# _adj [star]
# _noun [plus] #
# _conj
# _np
# @@
# dqan alpha alpha
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3))
fail();
if (N("verb",2))
{
if (vconjq(N(2),"-ed")
|| vconjq(N(2),"-s"))
fail();
}
if (N("verb",3))
{
L("vc") = vconj(N(3));
if (L("vc") == "-ed"
|| L("vc") == "-edn"
|| L("vc") == "-en")
fail();
}
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
else
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_xWILD [one match=(_noun)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND _clausesep _fnword)]
@@
# 's alpha alpha noun verb
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
@POST
fixnpnonhead(2);
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xALPHA
_noun [lookahead]
_xWILD [one match=(_modal _verb _vg)]
@@
# noun , alpha alpha conj
@CHECK
if (!N("noun",4))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
@POST
fixnphead(4);
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_xALPHA
\, [opt]
_conj [lookahead]
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# 's alpha alpha prep
# 's alpha alpha conj
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("noun",3))
fail();
@POST
fixnpnonhead(2);
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _conj \, )]
@@
# ^ alpha det
# ^ alpha dqan
@CHECK
if (!N("verb",2))
fail();
@POST
alphatovg(2,"active","VB"); # Imperative. # Cleanup 05/25/07 AM.
N("pos50 vg1",2) = 1;
@RULES
_xNIL <-
_xSTART
_xALPHA
_det [lookahead]
@@
# alpha dqan
# ving (adj,verb)
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
alphatovg(3,"active",0); # Cleanup. # 05/25/07 AM.
N("pos50 vg2",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [opt]
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM)]
@@
# be alpha alpha dqan
# is head account manager
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("adv",2))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
if (N("verb",2))
{
S("vc") = vconj(N(2));
if (S("vc") == "-ing"
|| S("vc") == "-edn"
|| S("vc") == "-en")
fail(); # Too likely a verb.
}
@POST
# Use the hot new function...
fixnpnonhead(2);
fixnpnonhead(3);
@RULES
_xNIL <-
_be [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun _adj)]
@@
# be alpha alpha
# was just waiting
@CHECK
L("v") = vconj(N(3));
if (L("v") != "-ing"
&& L("v") != "-edn"
&& L("v") != "-en")
fail();
if (!N("adv",2))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
posacct(N(3));
posacct(N(2));
@RULES
_xNIL <- # 10
_be [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# be adv alpha
@CHECK
L("v") = vconj(N(3));
if (L("v") != "-ing"
&& L("v") != "-edn"
&& L("v") != "-en")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_be [s]
_adv
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# vg np by alpha alpha+
# vg np prep alpha alpha
@PRE
<3,3> vareq("stem","by");
<4,4> var("verb");
@CHECK
if (!vconjq(N(4),"-ing"))
fail();
if (!N("noun",5) && !N("adj",5) && !N("unknown",5))
fail();
@POST
alphatovg(4,"active","VBG");
@RULES
_xNIL <-
_vg ### (1)
_xWILD [one match=(_np _noun _pro)] ### (2)
_prep ### (3)
_xALPHA ### (4)
_xALPHA [lookahead] ### (5)
@@
# prep alpha alpha
# by alpha alpha
@CHECK
if (!N("noun",2) || !N("verb",3))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
by [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_det _quan _xNUM _num
_adj _np
_noun _pro _adv _advl)]
@@
# for adv alpha
@CHECK
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
for [s]
_adv
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# prep alpha alpha noun
# in Western populated ...
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("noun",2) || !N("adj",2)) # Need both of these.
fail();
if (!N("noun",3) && !N("adj",3))
fail();
if (N("verb",3))
{
S("vc") = vconj(N(3));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
}
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
if (N("noun",3))
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
}
else # adj
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
# if (N("verb",3))
fixadj(N(3));
}
@RULES
_xNIL <-
_prep
_xALPHA
_xALPHA
_noun
@@
# prep alpha alpha noun
@PRE
<1,1> varne("stem","to");
<2,2> varz("pos50 p-a-a-n"); # Loop guard.
@POST
N("pos50 p-a-a-n",2) = 1;
if (!N("verb",2))
{
if ((N("adj",3) || N("noun",3))
&& singular(N(2)) && !N("mass",2))
{
fixnpnonhead(2);
fixnpnonhead(3);
}
}
@RULES
_xNIL <-
_prep
_xALPHA
_xALPHA
_noun
@@
# prep alpha alpha
# by alpha alpha
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("noun",3))
fail();
if (N("verb",3))
{
S("vc") = vconj(N(3));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
}
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
by [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_np _fnword _verb _vg)]
@@
# prep adv alpha
# prep alpha
@CHECK
if (!N("adj",3))
fail();
if (plural(N(3)))
fail();
@POST
alphatoadj(3);
@RULES
_xNIL <-
_xWILD [s one match=(_prep) except=(to)]
_adv [plus]
_xALPHA
@@
# by alpha pro
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
alphatovg(2,"active","VBG"); # Cleanup. # 05/25/07 AM.
N("pos50 vg3",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(by from _conj)]
_xALPHA
_xWILD [one lookahead match=(_pro _np)]
@@
# vg advl
@PRE
<1,1> varz("voice");
<3,3> var("by-actor");
@POST
fixvg(N(1),"passive","VBN");
N("pos50 psv1",1) = 1;
@RULES
_xNIL <-
_vg
_adv [star lookahead]
_advl
@@
# vg prep np advl
@PRE
<1,1> varz("voice");
<5,5> var("by-actor");
@POST
fixvg(N(1),"passive","VBN");
N("pos50 psv2",1) = 1;
group(2,3,"_advl"); # prep np -> advl
N("pp",2) = 1;
@RULES
_xNIL <-
_vg
_prep
_np
_adv [star lookahead]
_advl
@@
# vg prep np conj alpha advl
@PRE
<5,5> var("noun");
<6,6> var("by-actor");
@POST
fixnphead(5);
@RULES
_xNIL <-
_vg
_prep
_np
_conj
_xALPHA
_advl
@@
# vg prep np conj alpha alpha
@CHECK
if(N("pos50 vpn-conj-aa",5))
fail(); # Loop guard.
@POST
N("pos50 vpn-conj-aa",5) = 1;
L("v") = pnvar(N(1),"verb node");
L("pos") = pnvar(L("v"),"mypos");
L("voice1") = N("voice",1);
# Todo: vn agreement versus nn agreement...
if (N("verb",5))
{
if (singular(N(5)) && samevconj(L("v"),N(5)))
alphatoverb(5,L("voice1"),L("pos"));
}
@RULES
_xNIL <-
_vg
_prep
_np
_conj
_xALPHA
_xALPHA
@@
# vg prep np
# passive/active computation.
@CHECK
if (N("pos50 v-p-n",2))
fail();
if (N("stem",4) == "by")
fail();
# If verb is assigned.
S("v") = N("first verb",2);
if (!S("v"))
fail();
S("p") = phrprepverbq(S("v"),N(4));
if (!S("p"))
fail();
if (pnvar(S("v"),"VBN"))
fail();
L("stem") = strtolower(pnvar(S("v"),"$text"));
if (finddictattr(L("stem"),"T3"))
succeed();
if (finddictattr(L("stem"),"D1"))
fail();
# If np sem is adverbial, eg date, could still be passive.
if (N("sem",6) == "date")
fail();
@POST
N("pos50 v-p-n",2) = 2; # Loop guard.
fixvg(N(2),"active",0);
N("pos50 vg4",2) = 1;
if (!N("mypos",4))
{
if (S("p") == 1 || S("p") == 3)
chpos(N(4),"RP"); # prep/RP
else
chpos(N(4),"IN"); # prep/IN
}
listadd(2,4,"true"); # vg prep -> vg
@RULES
_xNIL <-
_xWILD [one fail=( \, )]
_vg
_xWILD [star lookahead match=(_adv _advl)]
_prep
_xWILD [star match=(_adv _advl)]
_np
@@
# noun conj alpha prep np vg
@PRE
<3,3> var("noun");
@CHECK
if (N("verb",3))
{
if (phrprepverbq(N(3),N(4)))
fail();
}
@POST
fixnphead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_prep
_np
_vg
@@
# noun conj alpha alpha
@CHECK
if (N("verb",3) || N("mass",3))
fail();
if (!N("noun",3) || !N("adj",3))
fail();
if (!N("noun",4) && !N("adj",4))
fail();
if (!singular(N(3)))
fail();
@POST
alphatoadj(3);
@RULES
_xNIL <- # 20
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_xALPHA
@@
# conj alpha alpha .
@CHECK
if (N("pos num",3) != 2)
fail();
if (!N("noun",3) || !N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_conj _prep)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND
_conj )] # 05/13/07 AM.
@@
# to alpha alpha noun
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf"))
fail();
if (!N("noun",3) && !N("adj",3) && !N("adv",3))
fail();
@POST
alphatoverb(2,"active","VB");
N("pos50 vg5",2) = 1;
@RULES
_xNIL <-
to [s]
_xALPHA
_xALPHA
_noun [lookahead]
@@
# to alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"inf")) # New function. # 06/15/06 AM.
fail();
@POST
alphatovg(3,"active","VB");
N("pos50 vg6",3) = 1;
@RULES
_xNIL <-
to [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star match=(_adv _advl)]
_xWILD [lookahead one match=(
_det _quan _num _xNUM _adj _np _pro)]
@@
# to alpha
@PRE
<3,3> varz("noun");
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"inf")) # New function. # 06/15/06 AM.
fail();
@POST
alphatovg(3,"active","VB");
N("pos50 vg7",3) = 1;
@RULES
_xNIL <-
to [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star match=(_adv _advl)]
_xWILD [lookahead one match=(_conj _prep _fnword _whword \,)]
@@
# vg prep alpha noun
# verb prep alpha noun
@CHECK
if (!N("verb",3) && !N("adj",3))
fail();
if (N("verb",3))
{
if (!vconjq(N(3),"-ing"))
fail();
}
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_prep
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# prep alpha
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("noun",2) || !N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_xWILD [s one match=(_prep though) except=(to)]
_xALPHA
@@
# prep alpha
@PRE
<1,1> varne("stem","to");
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2))
{
S("vc2") = vconj(N(2));
if (S("vc2") != "inf"
&& S("vc2") != "-s"
&& S("vc2") != "-edn"
&& S("vc2") != "-en")
fail();
}
@POST
fixnpnonhead(2);
@RULES
_xNIL <-
_prep
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# prep alpha
@CHECK
if (!N("verb",2) || !N("adj",2))
fail();
S("vc2") = vconj(N(2));
if (S("vc2") != "-edn"
&& S("vc2") != "-en")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2)); # Probably starts np.
@RULES
_xNIL <-
_xWILD [s one match=(_prep) except=(to as)]
_xALPHA
_xWILD [one lookahead fail=(_prep)]
@@
# as noun verb
@CHECK
if (N("mypos",2) == "IN")
fail();
@POST
chpos(N(2),"IN");
@RULES
_xNIL <-
_xSTART
as [s]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [star match=(_det _quan _num _xNUM _adj)]
_xWILD [plus match=(_noun)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# prep alpha adj
# Note: a 'normal' case!? #
@PRE
<2,2> var("adj");
<2,2> varz("verb");
@POST
alphatoadj(2);
@RULES
_xNIL <-
_prep
_xALPHA
_adj [lookahead]
@@
# fnword alpha adj
# as alpha adj
@PRE
<1,1> vareq("stem","as");
<2,2> var("adj");
<2,2> varz("verb");
@POST
alphatoadj(2);
@RULES
_xNIL <-
_fnword
_xALPHA
_adj [lookahead]
@@
# as alpha adj
# prep alpha adj
@CHECK
if (N("stem",1) != "as")
fail();
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
if (N("mass",2)) # Mass noun.
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
chpos(N(1),"IN");
@RULES
_xNIL <-
_fnword
_xALPHA
_adv [star lookahead]
_adj
@@
# prep alpha adj
# prep alpha dqan
# verb vs adj.
# Assume adj-adj, idioms handled prior to here.
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <- # 30
_xWILD [one match=(_prep)]
_xALPHA
_xWILD [one lookahead match=(_adj _pro)]
@@
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_prep
even
@@
# det noun alpha det
#@CHECK
# if (!N("verb",3))
# fail();
#@POST
# L("tmp3") = N(3);
# group(3,3,"_verb");
# pncopyvars(L("tmp3"),N(3));
# fixverb(N(3),0,0);
# L("v") = N(3);
# group(3,3,"_vg");
# mhbv(N(3),L("neg"),0,0,0,0,L("v"));
# pncopyvars(L("tmp3"),N(3));
# N("verb node",3) = L("v");
# N("voice",3) = "active";
# clearpos(N(3),1,0);
#@RULES
#_xNIL <-
# _det
# _noun
# _xALPHA
# _xWILD [one lookahead match=(_det)]
# @@
# det alpha adj
@POST
L("tmp2") = N(2);
if (N("num",2)) # Eg, "first".
{
group(2,2,"_num");
chpos(N(2),"JJ");
}
else if (N("adj",2))
group(2,2,"_adj");
else if (N("noun",2))
group(2,2,"_noun");
else if (N("adv",2))
group(2,2,"_adv");
else
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
@RULES
_xNIL <-
_det
_xALPHA
_xWILD [one lookahead match=(_adj _xNUM _num _quan)]
@@
# prep dqa alpha alpha $
@PRE
<6,6> var("noun");
@CHECK
if (!N("adj",5) && !N("noun",5))
fail();
@POST
fixnphead(6);
fixnpnonhead(5);
@RULES
_xNIL <-
_prep
_xWILD [star match=(_det _adv)]
_xWILD [star match=(_quan _num _xNUM _adv)]
_adj [star]
_xALPHA
_xALPHA
_xWILD [one match=( \) _dbldash _xEND)]
@@
# conj det alpha alpha noun
@PRE
<3,3> var("noun");
<4,4> var("verb");
@POST
fixnphead(3);
alphatovg(4,"active","VBP");
@RULES
_xNIL <-
_conj
_det
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun _np _det _quan _num _xNUM)]
@@
# det alpha alpha $
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3))
fail();
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_xWILD [one match=(_prep \) _dbldash _xEND)]
@@
# det alpha alpha verb
@PRE
<2,2> var("noun");
<3,3> var("adj");
<3,3> var("adv");
<3,3> varz("verb");
<3,3> varz("noun");
@POST
alphatoadv(3);
fixnphead(2);
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _modal _vg)]
@@
# det alpha alpha
@CHECK
if (!N("noun",3))
fail();
if (N("verb",3))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
}
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _xEND)]
@@
# dqa alpha
@CHECK
if (!N("noun",2))
fail();
if (N("adj",2)) # 04/20/07 AM.
fail();
L("n") = lasteltnode(1);
# if (pnvar(L("n"),"apos-s"))
# succeed();
if (numbersagree(N(1),N(2)))
succeed();
if (N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xWILD [one lookahead match=(_qEOS _fnword _conj
_pro \, _xEND)]
@@
# adj alpha
@CHECK
if (!N("noun",3))
fail();
if (N("verb",3))
{
S("vc") = vconj(N(3));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
}
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
dqaninfo(0,0,2,3);
groupnp();
@RULES
_xNIL <- #50
_xWILD [one match=(_conj)]
_adj
_xALPHA
_xWILD [lookahead one match=(_qEOS _verb _vg _xEND)]
@@
# Too broad and old. #
# dqa alpha
# Usually will be a noun, even if verb is possible.
#@PRE
#<1,1> varne("stem","each");
#@CHECK
# if (!N("noun",2))
# fail();
# L("last") = lasteltnode(2);
# if (N(1))
# L("first") = N(1);
# else if (N(2))
# L("first") = N(2);
# if (!numbersagree(L("first"),L("last")))
# S("disagree") = 1;
#@POST
# if (S("disagree"))
# {
# group(1,1,"_np");
# N("bracket",1) = 1;
# }
# else
# {
# L("tmp2") = N(2);
# group(2,2,"_noun");
# pncopyvars(L("tmp2"),N(2));
# }
#@RULES
#_xNIL <-
# _xWILD [plus match=(_det _quan _num _xNUM)]
# _xALPHA
# _xWILD [one lookahead match=(_qEOS _fnword _conj _prep
# _pro _xEND)]
# @@
# det alpha alpha np vg
# Note: np (that) np vg
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("noun",3))
fail();
@POST
fixnphead(3);
fixnpnonhead(2);
N("glom",4) = "right";
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_np [lookahead]
_vg
@@
# det alpha
# dqa alpha noun
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
@POST
"flag.txt" << phrasetext() << "\n";
fixnpnonhead(3); # 01/17/07 AM.
@RULES
_xNIL <- # 39. 76%
_xWILD [plus match=(_det _quan _num _xNUM)]
_adj [star]
_xALPHA
_noun [plus]
@@
# det alpha
@CHECK
if (N("noun",2) && (N("adj",2) || N("adv",2)))
fail(); # Ambiguous.
if (N("adj",2) && N("adv",2))
fail();
if (!N("needs-np",1))
# if (!N("needs-np")) # BAD...
fail(); # This det can stand alone. eg, 'all are punished'.
@POST
if (N("noun",2) || N("unknown",2))
alphatonoun(2);
else if (N("adj",2))
alphatoadj(2);
else if (N("adv",2))
alphatoadv(2);
else
{
if (G("error"))
"err.txt" << "[det alpha: " << N("$text",2) << "\n";
alphatoadj(2);
N("guess",2) = 1;
}
@RULES
_xNIL <- # 40
_det
_xALPHA
@@
# det alpha
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2))
fail();
@POST
fixnpnonhead(2); # adj-noun thingy...
@RULES
_xNIL <-
_det
_xALPHA
_xWILD [one match=(_adj _noun)]
@@
# prep alpha , dqan conj alpha
@PRE
<1,1> varne("stem","to");
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",9) && !N("adj",9))
fail();
@POST
fixnphead(2);
@RULES
_xNIL <-
_xWILD [one match=(_prep)] ### (1)
_xALPHA ### (2)
\, ### (3)
_xWILD [star match=(_det _pro)] ### (4)
_xWILD [star match=(_quan _num _xNUM)] ### (5)
_adj [star] ### (6)
_noun [plus] ### (7)
_xWILD [plus match=( \, _conj)] ### (8)
_xALPHA ### (9)
@@
# prep alpha , adj
@PRE
<1,1> varne("stem","to");
<2,2> var("adj");
@POST
alphatoadj(2);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xALPHA
\,
_adj
@@
# dqa alpha
# det alpha , adj
@PRE
<2,2> var("adj");
<2,2> varz("noun");
@POST
alphatoadj(2);
@RULES
_xNIL <-
_xWILD [one match=(_det)]
_xALPHA
\,
_adj
@@
# det alpha
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2))
fail();
@POST
fixnphead(2);
@RULES
_xNIL <-
_xWILD [one match=(_det)]
_xALPHA
_xWILD [star match=( \, _conj)]
_xWILD [one fail=(_adj _noun _xALPHA)]
@@
_xNIL <-
_xWILD [one match=(those)]
_xALPHA
_xEND
@@
# det alpha alpha
# Note: Roughing out...
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2) || N("verb",3))
fail();
@POST
if (N("noun",3) || N("adj",3) || N("unknown",3))
fixnpnonhead(2);
else
fixnphead(2);
# adj-noun thingy.
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
@@
# det adj , alpha noun
@CHECK
if (!N("adj",4))
fail();
@POST
alphatoadj(4);
group(2,4,"_adj"); # adj list.
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [plus match=(_det _pro _num _quan _xNUM)]
_adj
\,
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# det adj , adj noun
# det adj , adj adj
@POST
group(2,4,"_adj"); # adj list.
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [plus match=(_det _pro _num _quan _xNUM _prep)]
_adj
\,
_adj
_xWILD [one lookahead match=(_adj _noun)]
@@
# ^ dqan alpha dqan
@PRE
<7,7> var("verb");
@POST
# todo: Check agreement.
alphatovg(7,"active","VBP");
N("pos50 vg8",7) = 1;
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_adv _advl)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_det _pro _quan _num _xNUM)]
@@
# adj alpha , adj noun $
# dqa alpha dqan
@CHECK
if (!N("adj",2))
fail();
if (N("verb",2))
if (!vconjq(N(2),"-ing")
&& !vconjq(N(2),"-en"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
group(2,4,"_adj");
N("list",2) = 1;
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
\, # 04/29/10 AM.
_adj [plus]
_noun
@@
# noun 's alpha
# Apple's (Nasdaq:AAPL) shares jumped
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
@POST
chpos(N(2),"POS");
L("tmp4") = N(4);
L("tmp1") = lasteltnode(1);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
group(1,3,"_adj");
pncopyvars(L("tmp1"),N(1));
clearpos(N(1),1,0);
@RULES
_xNIL <-
_noun [plus]
_xWILD [one match=(_aposS _aposX)]
_acro [opt]
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# conj alpha
# and unknown adj
@CHECK
if (!N("unknown",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_conj)]
_xALPHA
_xWILD [one lookahead match=(_adj)]
@@
# conj alpha adj
@PRE
<3,3> var("noun");
<3,3> var("verb");
<3,3> vareq("pos num",2);
@POST
alphatoverb(3,0,0);
@RULES
_xNIL <-
_conj
_adv [star]
_xALPHA
_adj [lookahead]
@@
# conj alpha
@CHECK
if (N("noun",2))
fail();
if (N("adv",2))
fail();
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
clearpos(N(2),1,0);
@RULES
_xNIL <- # 50
_conj
_xALPHA
_xWILD [one lookahead match=(_det _np)]
@@
# stronger than expected
# adj than alpha
#@CHECK
# if (!N("verb",3))
# fail();
# L("v") = vconj(N(3));
# if (L("v") != "-edn"
# && L("v") != "-en")
# fail();
#@POST
# single();
#@RULES
#_adj <-
# _adj
# than [s]
# _xALPHA
# @@
# on continued strong
@CHECK
if (!N("verb",2))
fail();
L("v") = vconj(N(2));
if (L("v") != "-edn"
&& L("v") != "-en")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
on [s]
_xALPHA
_adj [lookahead]
@@
# another adj
@POST
L("tmp1") = N(1);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
another
_xWILD [one lookahead match=(_quan _num _xNUM _adj _noun)]
@@
# ^ noun prep
# Also news from
@POST
L("tmp3") = lasteltnode(3);
L("fst") = N(3);
L("ne arr") = nenoderange(L("fst"),L("tmp3"));
nountonp(3,1);
if (L("ne arr"))
N("ne arr",3) = L("ne arr");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xWILD [plus match=(_adv _advl)]
_noun [one] # FIX. 04/14/10 AM.
_xWILD [one lookahead match=(_prep)]
@@
# prep noun
# conj noun prep
@PRE
<3,3> varz("cap");
@POST
if (N("ne",3))
L("ne arr") = N(3);
fixnounhead(N(3));
nountonp(3,1);
if (L("ne arr"))
N("ne arr",3) = L("ne arr");
@RULES
_xNIL <-
_xWILD [one match=(_prep _conj)]
_adv [star]
_noun
_xWILD [one lookahead match=(_qEOS _pro _conj _dbldash
_prep _xEND)]
@@
# music player posted
# noun alpha verb
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2))
{
if (vconjq(N(2),"-ed"))
fail();
if (verbfeat(N(2),"T5"))
fail();
}
@POST
# If highly ambiguous, may need to put preferences on it
# in various contexts.
L("tmp2") = N(2);
if (N("adv",2)) # Try this.
{
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
}
else
fixnphead(2);
@RULES
_xNIL <-
_noun
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# noun alpha verb
# Assume np agreement failed and so on.
# ex: This makes going to the store impossible.
@PRE
<1,1> varz("proposs");
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
N("glom",2) = "left";
N("glom",3) = "right";
# todo: _vg.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro)]
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
@POST
L("tmp2") = N(2);
L("noun") = lasteltnode(3);
L("ne arr") = nenoderange(L("tmp2"),L("noun"));
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
group(2,3,"_np");
pncopyvars(L("noun"),N(2));
clearpos(N(2),1,1);
if (L("ne arr"))
N("ne arr",2) = L("ne arr");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xALPHA
_noun [plus]
_xWILD [one lookahead match=(_verb _vg _advl _adv)]
@@
# noun alpha noun
# but: eastern industrialized countries
@PRE
<1,1> varz("cap"); # 04/20/07 AM.
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("verb",2))
fail();
if (N("noun",2))
fail();
@POST
alphatovg(2,0,0); # 05/22/07 AM.
@RULES
_xNIL <-
_noun
_xALPHA
_xWILD [one lookahead match=(_noun _np _pro)]
@@
# noun alpha dqan
# np alpha dqan
@CHECK
if (!N("verb",2))
fail();
if (N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
fixverb(N(2),"active","VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg9",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _adj _pro)]
@@
# verb dqan alpha prep
# TOO BROAD. TOO OLD. #
#@CHECK
# if (!N("noun",3))
# fail();
# if (N("verb",3))
# fail();
#@POST
# L("tmp3") = N(3);
# group(3,3,"_noun");
# pncopyvars(L("tmp3"),N(3));
# group(2,3,"_np");
# pncopyvars(L("tmp3"),N(2));
# clearpos(N(2),1,1);
# if (pnname(N(1)) == "_vg")
# if (!N("voice",1))
# N("voice",1) = "active";
#@RULES
#_xNIL <-
# _xWILD [one match=(_verb _vg)]
# _xWILD [plus match=(_det _quan _num _xNUM _adj _noun)]
# _xALPHA
# \, [opt lookahead]
# _xWILD [one match=(_prep _vg _verb _pro _fnword)]
# @@
# vg dqan alpha prep vg
# Note: Too many vgs, alpha should be a noun.
@PRE
<7,7> var("noun");
@CHECK
if (N(3))
L("s") = N(3);
else if (N(4))
L("s") = N(4);
else if (N(5))
L("s") = N(5);
else
L("s") = N(6);
if (!numbersagree(L("s"),N(7)))
fail();
@POST
fixnphead(7);
@RULES
_xNIL <-
_vg
_xWILD [star match=(_adv _advl)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_prep
_vg
@@
# vg dqan alpha np vg
# Note: Could still be a verb...
@PRE
<6,6> var("noun");
@POST
fixnphead(6);
N("ellipted-that",7) = 1;
N("glom",7) = "right";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_np [lookahead]
_vg
@@
# dqa alpha
@CHECK
if (!N("noun",3))
fail();
if (N("verb",3))
{
S("vc") = vconj(N(3));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
}
# Check agreement.
if (!numbersagree(N(2),N(3)))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
group(2,3,"_np");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <- # 60
_xSTART
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
\, [opt lookahead]
_xWILD [one match=(_prep _vg _verb _pro _fnword)]
@@
# dqan alpha dqan
@PRE
<1,1> varz("cap"); # 04/20/07 AM.
@CHECK
if (!N("verb",2))
fail();
if (vconjq(N(2),"inf"))
fail();
@POST
alphatovg(2,"active","VBP"); # Cleanup. # 05/25/07 AM.
N("pos50 vg10",2) = 1;
@RULES
_xNIL <-
_noun
_xALPHA
_noun [lookahead]
@@
_xNIL <-
_noun
_xALPHA
_proObj [s lookahead]
@@
# ^ noun alpha adv alpha
@CHECK
if (!N("noun",3))
fail();
if (!N("verb",5))
fail();
if (!singular(N(5)))
fail();
if (!plural(N(3)))
fail();
@POST
fixnphead(3);
alphatoverb(5,"active","VBP");
N("pos50 vg11",5) = 1;
@RULES
_xNIL <-
_xSTART
_noun
_xALPHA
_adv
_xALPHA
_xWILD [one lookahead match=( \, _clausesep _xEND)]
@@
# ^ noun alpha
# ^ noun noun alpha
@CHECK
if (N("adj",3))
fail();
if (!N("noun",3) && !N("verb",3))
fail();
if (vconjq(N(3),"-edn"))
fail();
@POST
L("last") = lasteltnode(2);
if (plural(L("last")) || pnvar(L("last"),"mass"))
{
if (vconjq(N(3),"-s"))
alphatonoun(3);
else
alphatoverb(3,0,"VBP");
}
else if (singular(L("last")))
alphatonoun(3);
@RULES
_xNIL <-
_xSTART
_noun [plus]
_xALPHA
_xWILD [lookahead one fail=(_xALPHA _num _xNUM _quan _np _det _adj)]
@@
# Note: Enabling this rule reduced accuracy of tagger.
# # ^ dqan alpha [NIBBLE]
#@CHECK
# if (N("pos num",6) != 2)
# fail();
# if (!N("noun",6) || !N("verb",6))
# fail();
#@POST
# # Todo: A "need plural" function.
# # Todo: An "np complete" function.
# L("last") = lasteltnode(5);
# if (!N(2) && !N(3) && plural(N(6))
# && singular(L("last")) )
# if (numbersagree(N(1),N(6)))
# fixnphead(6);
# else
# alphatoverb(6,0,0);
#@RULES
#_xNIL <-
# _xWILD [one match=(_xSTART _qEOS)]
# _xWILD [star match=(_det _pro)]
# _xWILD [star match=(_quan _num _xNUM)]
# _adj [star]
# _noun [plus]
# _xALPHA
# @@
# Too old. Too broad. #
# verb dqan alpha [commented out]
# If a verb, vg must be a helping verb etc.
#@CHECK
# if (!N("noun",4))
# fail();
#@POST
# L("tmp4") = N(4);
# group(4,4,"_noun");
# pncopyvars(L("tmp4"),N(4));
# L("ne arr") = nenoderange(N(3),N(4));
# group(2,4,"_np");
# pncopyvars(L("tmp4"),N(2));
# clearpos(N(2),1,1);
# if (L("ne arr"))
# N("ne arr",2) = L("ne arr");
# if (pnname(N(1)) == "_vg")
# if (!N("voice",1))
# N("voice",1) = "active";
#@RULES
#_xNIL <-
# _xWILD [one match=(_verb _vg)]
# _det [star]
# _xWILD [plus match=(_quan _num _xNUM _adj _noun)]
# _xALPHA
# _xWILD [one match=(_qEOS _xEND)]
# @@
#_xNIL <-
# _xWILD [one match=(_verb _vg)]
# _proPoss [s star]
# _xWILD [plus match=(_quan _num _xNUM _adj _noun)]
# _xALPHA
# _xWILD [one match=(_qEOS _xEND)]
# @@
# conj noun .
@POST
L("tmp2") = lasteltnode(2);
L("ne arr") = nenoderange(N(2),L("tmp2"));
fixnouns(N(2),L("tmp2"));
nountonp(2,1);
N("ne",2) = 0;
if (L("ne arr"))
N("ne arr",2) = L("ne arr");
@RULES
_xNIL <-
_xWILD [one match=(_conj)]
_noun [plus]
_xWILD [one lookahead match=(_qEOS _xEND)]
@@
# prep noun ,
#@CHECK
# L("n") = lasteltnode(2);
# if (pnvar(L("n"),"num"))
# fail();
#@POST
# L("tmp2") = lasteltnode(2);
# group(2,2,"_np");
# pncopyvars(L("tmp2"),N(2));
# clearpos(N(2),1,1); # Zero out token info.
# N("ne",2) = 0;
#@RULES
#_xNIL <-
# _prep
# _noun [plus]
# \, [opt lookahead]
# _xWILD [one match=(_det _quan _num _xNUM _adj _noun _np _adv)]
# @@
# he 's held it
# pro 's alpha
# 's => HAD, IS
@CHECK
if (!N("verb",3))
fail();
# Check verb agreement "has held"...
@POST
# Don't know if 's is be or have.
if (!N("mypos",2))
{
chpos(N(2),"VBZ");
N("apos-s to is/has",2) = 1;
}
L("b") = N(2);
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
L("v") = N(3);
fixverb(N(3),"active","VBP");
group(2,3,"_vg");
mhbv(N(2),L("neg"),0,0,L("b"),0,L("v"));
pncopyvars(L("tmp3"),N(2));
N("voice",2) = "active"; # Not absolute...
clearpos(N(2),1,0); # Zero out token info.
N("pos50 vg12",2) = 1;
@RULES
_xNIL <-
_proSubj [s]
_aposS
_xALPHA
_xWILD [one lookahead match=(_np _noun _det _quan _num _xNUM
_adj)]
@@
# pro 's alpha
# he's playing
@CHECK
if (!N("verb",3))
fail();
L("v") = vconj(N(3));
if (L("v") != "-ing"
&& L("v") != "-edn"
&& L("v") != "-ed")
fail();
@POST
L("b") = N(2);
L("tmp3") = N(3);
group(3,3,"_verb");
L("v") = N(3);
pncopyvars(L("tmp3"),N(3));
group(2,3,"_vg");
mhbv(N(2),L("neg"),0,0,L("b"),0,L("v"));
pncopyvars(L("tmp3"),N(3));
# N("voice") = "active";
clearpos(N(3),1,0); # Zero out token info.
N("pos50 vg13",3) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(he she)]
_aposS
_xALPHA
@@
# prep alpha 's
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_prep _conj _det)]
_xALPHA
_xWILD [one match=(_aposS _aposX)]
@@
# noun prep
@POST
nountonp(2,1);
@RULES
_xNIL <- # 70
_xWILD [one match=(_dbldash)]
_noun
_xWILD [one lookahead match=(_prep)]
@@
# vg np alpha
# Note: ditrans verb pattern.
@CHECK
if (!verbfeat(N(1),"D1"))
fail();
@POST
fixnpnonhead(3);
@RULES
_xNIL <-
_vg
_np
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# vg np alpha
@CHECK
if (!verbfeat(N(1),"X7"))
fail();
@POST
alphatoadj(3);
L("tmp3") = N(3);
group(3,3,"_adjc");
pncopyvars(L("tmp3"),N(3));
clearpos(N(3),1,0);
@RULES
_xNIL <-
_vg
_np
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# verb noun alpha
# vg np alpha
# verb patterns...
@PRE
<3,3> var("prosubj");
<5,5> var("verb");
@CHECK
if (!vconjq(N(5),"inf"))
fail();
if (!verbfeat(N(1),"V2")
&& !verbfeat(N(1),"T5"))
fail();
@POST
alphatovg(5,"active","VB");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_pro _np _noun)]
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# pro alpha
# our book
@PRE
<1,1> varz("proobj");
<1,1> varz("prosubj");
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("noun",2) && N("adj",2))
fail();
if (N("noun",2) && N("verb",2))
{
if (vconjq(N(2),"-ing"))
fail(); # noun/vbg
}
@POST
L("tmp") = N(2);
if (N("adj",2))
group(2,2,"_adj");
else
group(2,2,"_noun");
pncopyvars(L("tmp"),N(2));
@RULES
_xNIL <-
_proPoss [s]
_xALPHA
@@
# pro alpha
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_proPoss [s]
_xALPHA
_xWILD [one lookahead match=(_verb _vg _qEOS _xEND)]
@@
# pro alpha verb
@CHECK
if (!N("adv",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_proSubj [s]
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# pro alpha alpha
@CHECK
if (!N("verb",2))
fail();
if (vconjq(N(2),"-ed"))
{
S("nd") = "VBD";
succeed();
}
S("nd") = N("need-vconj",1);
if (!S("nd")) # "it".
S("nd") = "-s";
if (!vconjq(N(2),S("nd")))
fail();
@POST
if (L("nd") == "inf")
L("pos") = "VBP";
else
L("pos") = "VBZ";
alphatoverb(2,"active",L("pos"));
N("pos50 vg14",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(_proSubj it)]
_xALPHA
_xALPHA
@@
# pro alpha
@PRE
<1,1> var("prosubj");
<1,1> varz("proobj");
<1,1> varz("proposs");
@CHECK
if (!N("verb",2) && !N("adv",2))
fail();
@POST
L("tmp2") = N(2);
if (N("verb",2))
{
alphatoverb(2,"active","VBP");
N("pos50 vg15",2) = 1;
}
else
{
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
}
@RULES
_xNIL <-
_xWILD [one match=(_pro)]
_xALPHA
_xWILD [one lookahead fail=(_verb _vg _xALPHA)]
@@
# who verb
@PRE
<1,1> varz("pos50 who-v");
@POST
N("pos50 who-v",1) = 1;
fixverb(N(2),"active","VBP");
@RULES
_xNIL <-
who [s]
_verb
@@
# who alpha
@CHECK
if (!N("verb",2))
fail();
@POST
alphatovg(2,"active","VBP");
N("pos50 vg15",2) = 1;
@RULES
_xNIL <-
_proSubj [s]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_det _quan _xNUM _num _adj _qEOS
_np _xEND)]
@@
# adj and alpha
@CHECK
if (!N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
L("txt") = nodetext(N(3));
L("myp") = adjconj(L("txt")); # Don't want VBG here.
if (L("myp"))
chpos(N(3),L("myp"));
else
N("mypos",3) = "JJ";
L("nm") = pnname(N(1));
group(1,3,L("nm"));
@RULES
_xNIL <-
_xWILD [one match=(_adj _adjc)]
_conj
_xALPHA
_xWILD [one fail=(_det _quan _num _xNUM _adj _noun _np)]
@@
# verb alpha adj
# note: Watch for vnj pattern "help verb" patterns etc.
# Need vnn watch also.
@POST
L("stem") = nodestem(N(1));
L("tmp2") = N(2);
if (!L("done") && N("verb",2))
{
if (finddictattr(L("stem"),"T2") && vconjq(N(2),"inf"))
{
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
else if (finddictattr(L("stem"),"T4") && vconjq(N(2),"-ing"))
{
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
}
if (!L("done") && N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
if (!L("done") && N("noun",2))
{
if (finddictattr(L("stem"),"X7") # keep it green ...
&& (N("mass",2) || plural(N(2)) ) )
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
}
if (!L("done") && N("adv",2))
{
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
# Default.
if (!L("done"))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
L("done") = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_adj [lookahead]
@@
# verb alpha alpha noun
# Todo: Look for verb-verb patterns etc...
@CHECK
if (N("pos50 v-a-a-n1",4))
fail();
@POST
N("pos50 v-a-a-n1",4) = 1;
if (N("adj",3) && N("noun",3))
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_xALPHA
_noun [lookahead]
@@
# verb alpha alpha verb
# Verb patterns....
# Ex: send people tumbling.
# Ex: make people sing.
@CHECK
# Todo: could handle adverbs too.
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <- # 80
_xWILD [one match=(_verb _vg)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# verb alpha alpha , verb
# vg alpha alpha , vg
@PRE
<3,3> var("noun");
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!plural(N(3)))
fail();
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <- #100
_xWILD [one match=(_verb _vg)]
_xALPHA
_xALPHA
\, [opt lookahead]
_vg
@@
# verb alpha alpha
@CHECK
if (N("stem",1) != "be"
&& N("$text",1) != "being")
fail();
if (!N("verb",3))
fail();
if (!N("adv",2) && !N("noun",2))
fail();
S("vc3") = vconj(N(3));
if (S("vc3") != "-ing"
&& S("vc3") != "-edn"
&& S("vc3") != "-en")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_verb
_xALPHA
_xALPHA
@@
# noun like alpha alpha verb
@CHECK
if (pnname(N(2)) != "_fnword")
fail();
if (!N("noun",4))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
if (vconjq(N(5),"-ing"))
fail();
@POST
fixnphead(4);
fixnpnonhead(3);
chpos(N(2),"IN"); # like/IN.
pnrename(N(2),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
like [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _vg _modal)]
@@
# expect further declines in
# verb alpha alpha prep
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("noun",3))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnounhead(N(3));
if (N("adj",2))
group(2,2,"_adj");
else
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
group(2,3,"_np");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [s one match=(_vg _verb) except=(_have _be _modal)]
_xALPHA
_xALPHA
_xWILD [lookahead one match=(_prep _fnword _xEND _dbldash)]
@@
# pro alpha verb
@CHECK
if (N("adv",2))
fail();
if (!N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "-s"
&& S("vc") != "inf")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_proSubj [s]
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# Looks like a weak rule. Need bounding...
# alpha verb
#@CHECK
# if (!N("verb",1))
# fail();
# if (N("pos num",1) != 2)
# fail();
#@POST
# L("tmp1") = N(1);
# if (N("noun",1))
# group(1,1,"_noun");
# else if (N("adj",1))
# group(1,1,"_adj");
# else if (N("adv",1))
# group(1,1,"_adv");
# else # Error?
# group(1,1,"_noun");
# pncopyvars(L("tmp1"),N(1));
# if (pnname(N(1)) == "_adj")
# fixadj(N(1));
#@RULES
#_xNIL <-
# _xALPHA
# _xWILD [one lookahead match=(_verb _vg)]
# @@
# Looks old and rare.
# which noun alpha verb
# which dqan alpha verb
# that noun alpha verb
@CHECK
if (!N("verb",5))
fail();
if (!vconjq(N(5),"-ed")
&& !vconjq(N(5),"-s"))
fail();
@POST
alphatovg(5,"active","VBP"); # Cleanup. # 05/25/07 AM.
N("pos50 vg16",5) = 1;
N("glom",5) = "left";
if (N(6))
N("glom",6) = "right";
else
N("glom",7) = "right";
@RULES
_xNIL <-
_xWILD [s one match=(_whword that)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_np _pro)]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# since early last june
# since alpha alpha noun
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (N("verb",2))
fail();
if (!N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
chpos(N(1),"IN");
L("tmp2") = N(2);
if (N("adj",2))
group(2,2,"_adj");
else # if (N("noun",2))
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
@RULES
_xNIL <-
since [s]
_xALPHA
_xALPHA
_noun [lookahead]
@@
# what np verb alpha dqan
# Note: Looking for a clause break here.
# Watch for diagreeing verbs and verb patterns.
# NOTE: Assume the mhbv pattern has tried and failed.
# (NOT a good assumption. Would rather do the current rule first.)
# Note: Other patterns like "what we looked at sent shivers..."
@PRE
<1,1> vareq("stem","what");
<2,2> var("prosubj");
<4,4> var("verb");
<4,4> varz("pos50 what-n-v-a");
@POST
N("pos50 what-n-v-a",4) = 1;
# Should also watch for highly ambig verbs (set etc.)
if (vconjq(N(4),"-ed"))
{
verbtovg(3,3,"active",0);
alphatovg(4,"active","VBD");
}
else if (vconjq(N(4),"inf"))
{
if (verbfeat(N(3),"T2"))
{
alphatoverb(4,"active","VB");
vgvgcompound(3,4,"vg-vg");
}
# Here, either a pattern like "did fight" or not a verb.
}
else if (vconjq(N(4),"-ing"))
{
if (verbfeat(N(3),"T4"))
{
alphatovg(4,"active","VBG");
vgvgcompound(3,4,"vg-vg");
}
else
{
# Here, maybe need to add a "likes eating" verb to "T4".
# Or not a verb.
}
}
else if (vconjq(N(4),"-en"))
{
# If not a pattern like "has eaten", then probably adj.
alphatoadj(4);
}
else
{
alphatovg(4,"active",0);
verbtovg(3,3,"active",0);
}
@RULES
_xNIL <-
_fnword
_xWILD [one match=(_noun _np _pro)]
_verb
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM _adj _noun)]
@@
# what alpha alpha verb
@CHECK
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
if (!plural(N(4)))
fail();
if (!singular(N(3)) && !vconjq(N(3),"-ing"))
fail();
if (!singular(N(2)) && !vconjq(N(2),"-ing"))
fail();
@POST
fixnpnonhead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_whword [s]
_xALPHA
_xALPHA
_noun [lookahead]
_xWILD [one match=(_modal _verb _vg)]
@@
# fnword alpha alpha $
# than alpha alpha $
@CHECK
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
if (!plural(N(3)) && !vconjq(N(3),"-ing"))
fail();
if (!singular(N(2)))
fail();
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_fnword
_xALPHA
_xALPHA
_xWILD [one match=(_xEND)]
@@
# since early june
# since alpha noun
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (N("verb",2))
fail();
@POST
chpos(N(1),"IN");
L("tmp2") = N(2);
if (N("adj",2))
group(2,2,"_adj");
else # if (N("noun",2))
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(3));
@RULES
_xNIL <-
since [s]
_xALPHA
_noun [lookahead]
@@
# back noun
# Should be locative.
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
back
_xWILD [one lookahead match=(_noun)]
@@
# alpha noun
# longer boats
@CHECK
if (N("pos num",1) != 2)
fail();
if (!N("adj",1) || !N("adv",1))
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
fixadj(N(1));
@RULES
_xNIL <- # 90
_xALPHA
_noun
@@
# ^ alpha noun
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (N("verb",2))
{
# S("num") = number(N(3));
# if (S("num") == "plural")
# fail();
if (!vconjq(N(2),"inf")) # 04/23/07 AM.
fail();
}
@POST
fixnpnonhead(2); # 01/17/07 AM.
@RULES
_xNIL <-
_xSTART
_xALPHA
_noun
_xWILD [one lookahead fail=(_noun _xALPHA)]
@@
# alpha noun [NIBBLE]
@CHECK
if (!N("adj",1) && !N("noun",1))
fail();
if (N("verb",1))
fail();
@POST
fixnpnonhead(1);
@RULES
_xNIL <-
# _xWILD [one match=(_xSTART _conj \, )]
_xALPHA
_noun [lookahead]
@@
# verb adj
@CHECK
if (!copulaq(N(1)))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adjc");
pncopyvars(L("tmp3"),N(3));
clearpos(N(3),1,0);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_adj
_xWILD [one lookahead match=(_prep _conj _fnword _dbldash
_qEOS _xEND)]
@@
# verb alpha
# Trying contextless...
@CHECK
if (N("stem",1) != "be"
&& N("$text",1) != "being")
fail();
if (!N("verb",3))
fail();
S("vc3") = vconj(N(3));
if (S("vc3") != "-ing"
&& S("vc3") != "-edn"
&& S("vc3") != "-en")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_verb
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# Note: What to do if it's a "to be" copula?...this doesn't
# seem to handle it.
# verb alpha prep
@CHECK
if (N("pos50 copula-alpha",3))
fail(); # Loop guard.
if (N("adv",3)) # 01/18/07 AM.
fail(); # Don't know which.
if ((N("stem",1) == "be"
|| N("$text",1) == "being"
|| N("stem",1) == "have")
&& N("verb",3))
fail();
if (!N("adj",3))
fail();
if (N("verb",3) &&
(N("sem",1) == "be" || N("sem",1) == "being"))
{
S("vc3") = vconj(N(3));
if (S("vc3") == "-edn"
|| S("vc3") == "-en"
|| S("vc3") == "-ing")
fail();
}
@POST
"74-95-adj.txt" << phrasetext() << "\n";
L("txt") = nodesem(N(1));
L("tmp3") = N(3);
if (copulaq(N(1)))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
group(3,3,"_adjc");
pncopyvars(L("tmp3"),N(3));
clearpos(N(3),1,0);
}
N("pos50 copula-alpha",3) = 1;
@RULES
_xNIL <- # 95
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_conj)]
_xWILD [one fail=(_adj _adjc _xALPHA)]
@@
@CHECK
if (!N("verb",3))
fail();
if (!verbfeat(N(1),"T5"))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
alphatoverb(3,"active","VBG");
N("pos50 vg17",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# verb much prep
@POST
fixnphead(2);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
much
_xWILD [one lookahead match=(_prep _xEND)]
@@
# verb alpha prep
@CHECK
if (!N("adv",3))
fail();
S("txt1") = nodesem(N(1));
if (copulaq(N(1)))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adv");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv _premod)]
_xALPHA
_xWILD [one lookahead match=(_xEND _prep _fnword)]
@@
# verb alpha that
@CHECK
if (!N("advl",2)) # 04/15/05 AM.
fail();
if (!N("adv",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"), N(2));
if (plural(N(2))) # Todo: or mass noun. # 06/21/06 AM.
chpos(N(2),"NN"); # Conformity.
else if (copulaq(N(1)) && N("adj",2))
chpos(N(2),"JJ");
else
chpos(N(2),"RB");
if (!N("mypos",3))
chpos(N(3),"IN"); # that/IN
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
that [s]
@@
# vg alpha adv
# verb alpha adv
@PRE
<2,2> var("noun");
@CHECK
if (N("verb",2))
{
if (vconjq(N(2),"-ing") && verbfeat(N(1),"T4"))
fail();
if (vconjq(N(2),"inf") && verbfeat(N(1),"T2"))
fail();
}
@POST
fixnphead(2);
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_xALPHA
_adv [lookahead]
@@
# vg alpha
@CHECK
if (N("voice",1) != "passive")
fail();
if (!N("adv",2))
fail();
@POST
alphatoadv(2);
@RULES
_xNIL <-
_vg
_xALPHA
_xWILD [s one lookahead match=(if)]
@@
# verb alpha conj
# vg alpha conj
# Watch verb patterns...
#@CHECK
# if (!N("noun",3))
# fail();
#@POST
# alphatonoun(3);
#@RULES
#_xNIL <- #
# _xWILD [one match=(_verb _vg)]
# _xWILD [star match=(_adv _advl)]
# _xALPHA
# _xWILD [one lookahead match=(_conj)]
# @@
# verb alpha
# foreshadow rising
@PRE
<2,2> varz("noun");
@CHECK
if (!N("adj",2))
fail();
if (!N("adv",2) && !N("verb",2))
fail();
if (N("verb",2))
{
if (!vconjq(N(2),"-ing") && !vconjq(N(2),"-en"))
fail();
if (verbfeat(N(1),"T4") && vconjq(N(2),"-ing"))
fail();
}
@POST
alphatoadj(2);
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
N("pos50 vg18",1) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(_verb _vg) except=(_modal _have _be
being)]
_xALPHA
@@
# verb alpha
# mhbv alpha
@CHECK
if (N("verb",2))
{
if (!vconjq(N(2),"inf"))
fail();
}
@POST
L("tmp2") = N(2);
if (verbfeat(N(1),"T2")) # help eat ...
{
alphatoverb(2,"active","VB");
N("pos50 vg19",2) = 1;
}
else if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
else if (N("adv",2))
{
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
}
else
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# Idiomatic.
# verb fine
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [s one match=(_verb _vg) except=(_modal _have)]
_xWILD [one match=(fine swell super great well good)]
@@
# to adv verb
@PRE
<1,1> vareq("stem","to");
<3,3> varz("mypos");
@CHECK
if (!vconjq(N(3),"inf"))
fail();
@POST
fixverb(N(3),"active","VB");
@RULES
_xNIL <-
_prep
_adv [plus]
_verb
@@
# help alpha
# verb patterns....
@CHECK
if (!N("verb",2))
fail();
if (verbfeat(N(1),"T2") && vconjq(N(2),"inf"))
succeed();
if (verbfeat(N(1),"T4") && vconjq(N(2),"-ing"))
succeed();
fail();
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
alphatovg(2,"active","VB");
vgvgcompound(1,2,"vg-vg");
N("vg node",1) = L("tmp1");
# todo: semantics from nonhelping verb...
N("stem",1) = nodestem(L("tmp2"));
N("sem",1) = nodesem(L("tmp2"));
clearpos(N(1),1,0);
N("pos50 vg20",1) = 1;
@RULES
_xNIL <-
_vg
_xALPHA
_xWILD [one lookahead fail=(_aposS _aposX _verb _vg)]
@@
# verb verb
# help verb
@CHECK
if (!verbfeat(N(2),"T2"))
fail();
if (N("mypos",3) == "VB")
succeed();
if (!vconjq(N(3),"inf"))
fail();
@POST
verbtovg(3,3,"active","VB");
vgvgcompound(2,3,"vg-vg");
@RULES
_xNIL <-
_xWILD [one fail=(_verb _modal)]
_vg
_verb
@@
# help alpha
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf"))
fail();
@POST
alphatoverb(2,"active","VB");
alphatoverb(1,"active",0);
# L("tmp1") = N(1);
# L("tmp2") = N(2);
# group(2,2,"_verb");
# pncopyvars(L("tmp2"),N(2));
# fixverb(N(2),"active","VB");
# L("v") = N(2);
#
# group(1,1,"_verb");
# pncopyvars(L("tmp1"),N(1));
# fixverb(N(1),"active",0);
#
# group(1,2,"_vg");
# pncopyvars(L("tmp1"),N(1)); # Copy up "syntax" stuff.
# N("compoun-vg",1) = 1;
# N("vg node",1) = L("tmp1");
# # todo: semantics from nonhelping verb...
# N("stem",1) = nodestem(L("tmp2"));
# N("sem",1) = nodesem(L("tmp2"));
# clearpos(N(1),1,0);
N("pos50 vg21",1) = 1;
@RULES
_xNIL <-
help
_xALPHA
_xWILD [one lookahead fail=(_noun _xALPHA)]
@@
# noun alpha alpha
# verb patterns...
@CHECK
if (!N("verb",2) || !N("verb",3))
fail();
if (vconjq(N(2),"inf"))
S("pos2") = "VBP";
else if (vconjq(N(2),"-s"))
S("pos2") = "VBZ";
else if (vconjq(N(2),"-ed"))
S("pos2") = "VBD";
else
fail();
if (!vconjq(N(3),"-ing"))
fail();
S("stem2") = nodestem(N(2));
S("sem3") = nodesem(N(3));
if (!finddictattr(S("stem2"),"T4"))
fail();
@POST
L("v3") = alphatovg(3,"active",S("pos2"));
L("v2") = alphatovg(2,"active","VBP");
group(2,3,"_vg");
pncopyvars(L("v2"),N(2)); # Copy up "syntax" stuff.
N("compound-vg",2) = 1;
N("vg node",2) = L("v2");
# todo: semantics from nonhelping verb...
N("stem",2) = nodestem(L("stem2"));
N("sem",2) = nodesem(L("sem3"));
clearpos(N(2),1,0);
N("pos50 vg22",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun _pro _np _det
_quan _num _xNUM
_adj _fnword _prep)]
@@
# verb alpha dqan
# begin ving ...
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
if (!verbfeat(N(1),"T4"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBG");
if (pnname(N(1)) == "_vg")
{
N("complex-verb",1) = 1;
listadd(1,2,"false");
}
N("pos50 vg23",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_adj _noun _np _adv _advl)]
@@
# take place
@CHECK
if (!N("noun",2) || !N("verb",2))
fail();
if (N("adj",2))
fail(); # 10/20/04 AM.
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [s one match=(_verb _vg) except=(_modal _have _be _do)]
_xALPHA
@@
# alpha np verb
@PRE
<1,1> var("verb");
@CHECK
if (!verbfeat(N(1),"V8"))
fail();
if (!vconjq(N(5),"-en"))
fail();
@POST
alphatoverb(1,"active",0);
N("pos50 vg24",1) = 1;
@RULES
_xNIL <-
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# Too weak. #
# of working
# of alpha
#@CHECK
# if (!N("verb",2))
# fail();
# if (!N("noun",2) && !N("adj",2))
# fail();
#@POST
# L("tmp2") = N(2);
# L("vc") = vconj(N(2));
# if (N("adj",2))
# {
# group(2,2,"_adj");
# pncopyvars(L("tmp2"),N(2));
# fixadj(N(2));
# }
# else if (L("vc") == "-ing")
# group(2,2,"_verb");
# else if (N("noun",2))
# group(2,2,"_noun");
# pncopyvars(L("tmp2"),N(2));
#@RULES
#_xNIL <-
# _xWILD [s one match=(_prep) except=(to)]
# _xALPHA
# @@
# said ,
@POST
alphatovg(1,"active","VBD"); # Cleanup. # 05/25/07 AM.
N("pos50 vg25",1) = 1;
N("glom",1) = "left";
if (N(2))
N("glom",2) = "right";
@RULES
_xNIL <-
said
_xWILD [one lookahead match=(\, _dblquote _qEOS _xEND
_pro _np)]
@@
# noun alpha alpha prep
@CHECK
if (N("verb",2))
fail();
if (N("adj",2) || !N("noun",2))
fail();
if (N("noun",3))
fail();
if (!N("verb",3) && !N("adj",3))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("verb",3))
{
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
}
else if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# noun alpha alpha prep np vg
@CHECK
if (N("pos50 n-a-a-p-n-v",2))
fail();
@POST
N("pos50 n-a-a-p-n-v",2) = 1;
if (vconjq(N(3),"-ing"))
{
alphatoverb(3,"active","VBG");
N("pos50 vg26",3) = 1;
}
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
_np
_vg
@@
# noun alpha alpha prep verb
# noun alpha alpha to verb
@CHECK
if (!N("verb",3))
fail();
if (!N("noun",2))
fail();
if (!verbfeat(N(3),"T3"))
fail();
@POST
alphatoverb(3,0,0);
fixnphead(2);
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
to [s lookahead]
_xWILD [one match=(_modal _verb _vg)]
@@
# alpha alpha to verb
@PRE
<1,1> var("noun");
<1,1> var("verb");
<2,2> var("adj");
<2,2> varz("verb");
<2,2> varz("noun");
@CHECK
if (N("pos num",2) != 2)
fail();
@POST
alphatoadj(2);
@RULES
_xNIL <-
_xALPHA
_xALPHA
to [s lookahead]
_xWILD [one match=(_verb _vg)]
@@
# noun adv alpha prep
@CHECK
if (!N("verb",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [plus match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# adv alpha prep
@PRE
<2,2> var("verb");
@POST
alphatoverb(2,0,0);
@RULES
_xNIL <-
_adv
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# verb worth
@POST
alphatoadj(3);
group(3,3,"_adjc");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_adv [star]
worth
@@
# adj alpha noun
@PRE
<2,2> var("verb");
@CHECK
if (!vconjq(N(2),"-ing"))
fail();
@POST
alphatoverb(2,"active","VBG");
N("pos50 vg27",2) = 1;
@RULES
_xNIL <-
_adjc
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _np)]
@@
# verb adv alpha [NIBBLE]
@CHECK
if (!N("adj",3))
fail();
@POST
alphatoadj(3);
@RULES
_xNIL <-
_xWILD [s one match=(_verb _vg) except=(_be _have _do)]
_adv
_xALPHA
@@
# dqan alpha prep
# noun alpha prep
@PRE
<2,2> var("verb");
@CHECK
if (N("cap",1))
succeed(); # 05/21/07 AM.
if (!plural(N(1)))
fail();
if (vconjq(N(2),"-s"))
fail();
@POST
# Could be "people expected to", ie ambig active/passive.
if (vconjq(N(2),"-edn"))
alphatoverb(2,0,0);
else
alphatoverb(2,"active",0);
if (N("cap",1))
nountonp(1,1);
N("pos50 vg28",2) = 1;
@RULES
_xNIL <-
_noun
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# noun alpha prep
# Seems too loosy goosy #
#@CHECK
# if (N("noun",2) && N("verb",2))
# fail();
# if (N("eos",1))
# fail();
#@POST
# L("tmp2") = N(2);
# if (N("verb",2))
# {
# if (vconjq(N(2),"-ed") && phrprepverbq(N(2),N(4)))
# alphatovg(2,"active","VBD");
# else
# alphatovg(2,0,"VBP");
# }
# else if (N("adv",2))
# {
# group(2,2,"_adv");
# pncopyvars(L("tmp2"),N(2));
# }
# else if (N("adj",2))
# {
# group(2,2,"_adj");
# pncopyvars(L("tmp2"),N(2));
# fixadj(N(2));
# }
# else
# {
# group(2,2,"_noun");
# pncopyvars(L("tmp2"),N(2));
# fixnoun(N(2));
# }
#@RULES
#_xNIL <-
# _xWILD [one match=(_noun _np)]
# _xALPHA
# _xWILD [star match=(_advl _adv)]
# _xWILD [lookahead one match=(_prep)]
# @@
# noun alpha prep
# noun alpha adv prep
# Too broad.
#@CHECK
# if (!N("verb",2))
# fail();
#@POST
# alphatoverb(2,0,0);
#@RULES
#_xNIL <-
# _xWILD [one match=(_noun _np)]
# _xALPHA
# _xWILD [plus match=(_advl _adv)]
# _xWILD [lookahead one match=(_prep _adj)]
# @@
# noun alpha [IDIOM]
# locatives get adv.
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_noun
_xWILD [one match=(
overseas
)]
@@
# noun alpha [IDIOM]
# noun away
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [one match=(away)]
@@
# noun alpha
# week earlier
# (becomes an adverbial with its prior np)
# Tough to decide adjective vs adverb....
# Too broad. #
#@CHECK
# if (N("pos num",2) != 2)
# fail();
# if (!N("adj",2) || !N("adv",2))
# fail();
#@POST
# L("tmp2") = N(2);
# group(2,2,"_adv");
# pncopyvars(L("tmp2"),N(2));
# # Treat as adverb. #
# if (pnvar(L("tmp2"),"adj"))
# fixadj(N(2));
#@RULES
#_xNIL <-
# _noun
# _xALPHA
# @@
# noun alpha
# noun ved
@CHECK
if (!N("verb",3))
fail();
S("vc") = vconj(N(3));
if (S("vc") != "-ed"
&& S("vc") != "-edn")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
L("v") = N(3);
pncopyvars(L("tmp3"),N(3));
fixverb(N(3),"active","VBP");
group(3,3,"_vg");
mhbv(N(3),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp3"),N(3));
N("voice",3) = "active";
clearpos(N(3),1,0);
N("pos50 vg29",3) = 1;
@RULES
_xNIL <-
_xSTART
_noun
_xALPHA
@@
# np conj noun alpha
@CHECK
if (!N("noun",4))
fail();
if (!singular(N(3)) || !singular(N(4)))
fail();
@POST
if (N("verb",4) && vconjq(N(4),"-ing"))
{
alphatovg(4,"active","VBG");
N("pos50 vg30",4) = 1;
}
else
alphatonoun(4);
@RULES
_xNIL <-
_np
_conj
_noun
_xALPHA
@@
# conj noun alpha
@PRE
<3,3> lowercase();
@CHECK
if (!singular(N(3)))
fail();
if (!N("noun",4) || !N("verb",4))
fail();
if (N("pos num",4) > 2)
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <- #150
_xWILD [one match=(_xSTART)]
_conj
_noun
_xALPHA
@@
# noun which helps its
# noun which alpha pro
# Todo: agreement checks.
# that
@CHECK
if (!N("verb",3))
fail();
@POST
L("xx") = pnparent(X()); # 07/10/12 AM.
# Some temporal context.
if (pnvar(L("xx"),"date=present")) # 07/10/12 AM.
alphatovg(3,"active","VBP");
else
alphatovg(3,"active","VBD");
# Get what was actually assigned.
L("pos") = vgpos(N(3));
if (L("pos") == "VBG")
chpos(N(2),"IN"); # that/IN
else
chpos(N(2),"WDT");
N("pos50 vg31",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [s one match=(which that)]
_xALPHA
_xWILD [one lookahead match=(_pro _det _quan _num _xNUM
_prep)]
@@
# which verb
# that
@POST
L("tmp2") = N(2);
group(2,2,"_vg");
L("voice") = mhbv(N(2),L("neg"),0,0,0,0,L("tmp2"));
pncopyvars(L("tmp2"),N(2));
N("first verb",2) = N("verb node",2) = L("tmp2");
clearpos(N(2),1,0);
L("xx") = pnparent(X()); # 07/10/12 AM.
if (pnvar(L("xx"),"date=present")) # Some temporal context.
fixvg(N(2),"active","VBP");
else
fixvg(N(2),"active","VBD");
N("pos50 vg32",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xWILD [s one match=(_verb) except=(_vg _have _be _do)]
@@
# which alpha noun
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf")
&& !vconjq(N(2),"-ed")
&& !vconjq(N(2),"-s"))
fail();
if (!plural(N(3)))
fail();
@POST
alphatovg(2,"active","VBP");
N("pos50 vg33",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xALPHA
_xWILD [one match=(_noun _np _pro)]
_xWILD [one match=(_prep)]
@@
# which alpha pro
@PRE
<2,2> var("verb");
@POST
alphatovg(2,"active","VBP");
N("pos50 vg34",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xALPHA
_xWILD [one lookahead match=(_pro)]
@@
# which helps its
# Todo: agreement checks.
# that
@CHECK
if (!N("verb",2))
fail();
if (N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(L("v"),"active","VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg35",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_prep)]
@@
# that alpha noun verb
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
@POST
fixnpnonhead(2);
@RULES
_xNIL <-
that [s]
_xALPHA
_noun [lookahead]
_xWILD [one match=(_modal _verb _vg)]
@@
# which alpha dqan
# that alpha dqan
@CHECK
if (N("pos50 which-a-dqan",2))
fail();
@POST
N("pos50 which-a-dqan",2) = 1;
if (N("verb",2))
{
alphatoverb(2,0,0);
}
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xALPHA
_xWILD [one lookahead match=(_np
_det _pro _quan _num _xNUM _adj)]
@@
# that alpha noun
@PRE
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"-s"))
fail();
@POST
alphatoverb(3,"active","VBZ");
N("pos50 vg36",3) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [one lookahead match=(_noun _np)]
@@
# that alpha prep
@PRE
<3,3> varz("verb");
<3,3> var("noun");
@POST
fixnphead(3);
@RULES
_xNIL <-
that [s]
_adv [star]
_xALPHA
_prep [lookahead]
@@
# which alpha alpha
# that alpha alpha
@CHECK
if (N("pos50 which-a-a",3))
fail();
if (N("noun",3) && N("verb",4))
fail(); # Another interp.
@POST
N("pos50 which-a-a",3) = 1;
if (N("verb",3) && !vconjq(N(3),"inf")
&& (N("adj",4) || N("noun",4)) )
{
alphatoverb(3,"active","VBP");
}
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_adv [star]
_xALPHA
_xALPHA
_xWILD [one lookahead fail=(_xALPHA _verb _vg _modal)]
@@
# det alpha alpha alpha prep
# Note: should check on last alpha verb patterns...
@PRE
<2,2> varz("pos50 daaap");
@POST
N("pos50 daaap",2) = 1;
if (!N("noun",2) && !N("adj",2) && !N("unknown",2))
{
if (N("verb",2)
&& (vconjq(N(2),"-en") || vconjq(N(2),"-ing")) )
{
alphatoadj(2); # Fix to vbg/vbn/adj thingy.
}
else if (N("adv",2))
alphatoadv(2);
# Else don't know what to do here.
}
else if (!N("noun",2))
alphatoadj(2);
else if (!N("adj",2))
alphatonoun(2);
else
{
if (N("noun",4) && singular(N(4))) # Uninflected.
{
if (N("verb",3) && !vconjq(N(3),"inf")) # Inflected.
{
fixnphead(4);
alphatovg(3,"active",0);
fixnphead(2);
}
}
}
@RULES
_xNIL <-
_xWILD [s one match=(a an the)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# Biased rule.
# dqa alpha alpha alpha dqan
# Some multi alpha handlers.
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
if (!N("verb",4))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
L("tmp4") = N(4);
group(4,4,"_verb");
L("v") = N(4);
pncopyvars(L("tmp4"),N(4));
group(4,4,"_vg");
mhbv(N(4),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp4"),N(4));
N("verb node",4) = L("v");
# N("voice",4) = "active";
clearpos(N(4),1,0);
if (N("noun",3))
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
}
else
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
}
if (pnname(N(5)) != "_prep")
fixvg(N(4),"active","VBP");
N("pos50 vg37",4) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_det _np)]
@@
# adj alpha alpha alpha noun
# dqa alpha alpha alpha noun
@CHECK
if (!N("adj",4))
fail();
if (!N("adj",5) && !N("noun",5))
fail();
if (!N("noun",6) && !N("adj",6))
fail();
if (plural(N(4)) || plural(N(5)) || plural(N(6)))
fail();
@POST
fixnpnonhead(6);
fixnpnonhead(5);
fixnpnonhead(4);
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# noun alpha alpha alpha prep
@CHECK
if (N("pos50 naaap",2))
fail(); # Loop guard.
@POST
N("pos50 naaap",2) = 1; # Loop guard.
if (!N("verb",2) && !N("verb",3) && !N("verb",4))
{
# Rough it out for now.
if (N("noun",4))
fixnphead(4);
}
else if (!N("verb",3) && !N("verb",4))
alphatoverb(2,0,0);
else if (!N("verb",2) && !N("verb",4))
alphatoverb(3,0,0);
else if (!N("verb",2) && !N("verb",3))
alphatoverb(4,0,0);
else if (N("noun",3) && plural(N(3)))
{
# Need trans/intr logic too.
if (vconjq(N(4),"inf"))
{
L("voice") = "active";
L("pos") = "VBP";
}
alphatovg(4,L("voice"),L("pos"));
N("pos50 vg38",4) = 1;
fixnphead(3);
}
# if (!N("noun",3) && !N("adj",3))
# fail();
# if (singular(N(4)) && !N("mass",4))
# fail();
# if (N("verb",2) && N("noun",4))
# {
# group(2,2,"_verb");
# pncopyvars(L("tmp2"),N(2));
# L("v") = N(2);
# group(2,2,"_vg");
# pncopyvars(L("tmp2"),N(2));
# clearpos(N(2),1,0);
# N("verb node",2) = L("v");
# fixvg(N(2),"active","VBP");
# }
# else # Loop guard. #
# noop(); # FAIL. #
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
_xALPHA
_prep [lookahead]
@@
# noun alpha alpha alpha $
# Favor first as verb.
@PRE
<2,2> varz("pos50 faaae");
@CHECK
if (!N("verb",2) || !N("noun",4))
fail(); # 06/08/06 AM.
if (!N("noun",3) && !N("adj",3))
fail();
if (singular(N(4)) && !N("mass",4))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
L("tmp4") = N(4);
if (N("verb",2) && N("noun",4))
{
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
pncopyvars(L("tmp2"),N(2));
clearpos(N(2),1,0);
N("verb node",2) = L("v");
fixvg(N(2),"active","VBP");
}
N("pos50 faaae",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_fnword _noun)]
_xALPHA
_xALPHA
_xALPHA
_xEND
@@
# 's alpha alpha alpha verb
@CHECK
if (!N("noun",4))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
if (!N("adj",2) && !N("noun",2))
fail();
# Todo: agreement. nn and nv.
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
L("tmp4") = N(4);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
else
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
}
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
@RULES
_xNIL <-
_xWILD [one match=( _aposS _det _pro _adj _aposX )]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# 's alpha alpha alpha dqan
@CHECK
if (!N("verb",4))
fail();
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
if (plural(N(2)))
fail();
if (!vconjq(N(4),"-ed"))
fail();
@POST
alphatoverb(4,"active","VBP");
N("pos50 vg39",4) = 1;
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_xWILD [one match=( _aposS _aposX
# _det _pro _adj #
)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_adj _noun _np)]
@@
# that alpha alpha alpha verb
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf")
&& !vconjq(N(2),"-s"))
fail();
if (!N("noun",4))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnpnonhead(3);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
N("pos50 vg40",2) = 1;
@RULES
_xNIL <-
that [s]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# , alpha alpha alpha verb
@CHECK
if (!N("verb",2))
fail();
L("stem") = nodestem(N(2));
if (!finddictattr(L("stem"),"V2"))
fail();
if (!vconjq(N(5),"inf"))
fail();
@POST
alphatoverb(2,"active","VBD");
if (pnname(N(5)) == "_vg")
L("v") = N("verb node",5);
else
L("v") = N(5);
fixverb(L("v"),"active","VB");
N("pos50 vg41",5) = 1;
@RULES
_xNIL <-
_xWILD [one match=( \, _clausesep)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# verb alpha alpha alpha
# Check out single np hypothesis.
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
if (!N("noun",4))
fail();
if (N("noun",2))
{
L("n2") = number(N(2));
if (L("n2") != "singular")
fail();
}
if (N("noun",3))
{
L("n3") = number(N(3));
if (L("n3") != "singular")
fail();
}
L("n4") = number(N(4));
if (L("n4") != "plural")
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
L("tmp4") = N(4);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
else
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
}
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
group(2,4,"_np");
pncopyvars(L("tmp4"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_qEOS \, _dbldash _xEND
_prep)]
@@
# prep alpha alpha alpha prep
# 64 combinations. 4x4x4.
# NEED A NEW APPROACH TO THIS STUFF.
# to alpha alpha alpha prep
# to verb:
# to eat people quickly
# to eat green people
# to eat only people
# to paint houses green
# to keep eating people
# to keep people eating
# to help people eat
# to help eat people
# to adv:
# to quickly eat people
# to quickly help eat
# to quickly help only
#@CHECK
# if (N("pos50 to-a-a-a1",1))
# fail();
#@POST
# N("pos50 to-a-a-a1",1) = 1; # Loop guard.
#@RULES
#_xNIL <-
# to [s]
# _xALPHA
# _xALPHA
# _xALPHA
# _xWILD [one lookahead match=(_prep)]
# @@
# ^ noun prep alpha
# ^ np prep alpha
@PRE
<5,5> varne("stem","to");
@CHECK
if (!N("noun",7) && !N("adj",7))
fail();
@POST
fixnpnonhead(7); # adj-noun thingy...
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_xWILD [one match=(_noun _np)] ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_prep ### (5)
_xWILD [star match=(_adv _advl)] ### (6)
_xALPHA ### (7)
_noun [lookahead]
@@
# ^ np conj alpha [NIBBLE]
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
@POST
if (N("adj",4))
alphatoadj(4);
else
alphatonoun(4);
@RULES
_xNIL <-
_xSTART
_np
_conj
_xALPHA
@@
# prep alpha alpha alpha
@CHECK
if (!N("noun",4))
fail();
if (N("verb",4))
if (!vconjq(N(4),"-ing"))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
if (plural(N(3)))
fail();
@POST
L("tmp3") = N(3);
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep \, _fnword)]
@@
# adj alpha
# hyphenated adj
@CHECK
if (!N("hyphenated",1))
fail();
if (N("pos num",2) != 2)
fail();
if (!N("verb",2))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
@POST
L("tmp2") = N(2);
if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
}
else
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
@RULES
_xNIL <-
_adj
_xALPHA
@@
# 's alpha alpha verb
@CHECK
if (!N("noun",3))
fail();
if (!N("adj",2) && !N("noun",2))
fail();
# Todo: agreement. nn and nv.
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_xWILD [one match=( _aposS _aposX _det _pro )]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# dqan prep dqan alpha alpha
# noun prep noun alpha alpha
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
if (!N("noun",5))
fail();
if (!singular(N(4)))
fail();
if (!N("eventive",1) && !N("verb",1))
fail(); # requiring "eventive" noun.
@POST
fixnpnonhead(4);
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [s one match=(_prep) except=(to)]
_xWILD [one match=(_noun)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _clausesep \, )]
@@
# verb dqan conj dqan alpha
@CHECK
if (!N("noun",8) && !N("adj",8) && !N("adv",8))
fail();
@POST
if (N("noun",8))
alphatonoun(8);
else if (N("adv",8))
alphatoadv(8);
else
alphatoadj(8);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_conj
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_fnword _prep _conj
_modal _verb _vg _xEND _clausesep)]
@@
# verb noun conj dqa alpha alpha
# vg np conj dqa alpha alpha
@CHECK
if (!N("noun",5) && !N("adj",5))
fail();
if (!N("noun",6))
fail();
@POST
fixnphead(6);
fixnpnonhead(5);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_np
_conj
_adj [star]
_xALPHA
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_xEND _clausesep \, _prep)]
@@
# verb noun conj dqan alpha
# vg np conj dqan alpha
@PRE
<8,8> var("noun");
@POST
fixnphead(8);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_np
_conj
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_num _quan _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_xEND _clausesep \, _prep)]
@@
# dqa conj dqan alpha
@CHECK
if (!N("noun",6))
fail();
if (!singular(N(5)))
fail();
@POST
alphatonoun(6);
@RULES
_xNIL <-
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [plus]
_conj
_noun
_xALPHA
@@
# conj dqa alpha
@CHECK
if (N("pos50 conj-n-a",5))
fail();
@POST
N("pos50 conj-n-a",5) = 1;
if (N("noun",5) || N("adj",5))
{
if (N("verb",5))
{
--N("pos num",5);
N("verb",5) = 0;
}
if (N("adv",5))
{
--N("pos num",5);
N("adv",5) = 0;
}
}
alphaunambigred(5);
# adj-noun thingy.
@RULES
_xNIL <-
_conj
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [plus]
_xALPHA
@@
# prep dqan conj alpha
@CHECK
if (!N("noun",5))
fail();
S("lastn") = lasteltnode(3);
if (!singular(S("lastn")))
fail();
if (!singular(N(5)))
fail();
if (!plural(N(6)))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
group(3,5,"_noun"); # Noun list.
N("list",3) = 1;
N("ignorepos",3) = 1;
@RULES
_xNIL <-
_prep
_adj [star]
_noun [plus]
_conj
_xALPHA
_noun [lookahead]
@@
# ^ prep dqa alpha alpha
@CHECK
if (!N("noun",5))
fail();
if (!N("noun",4) && !N("adj",4))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
fixnoun(N(5));
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_prep)]
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _adv _clausesep _fnword)]
@@
# prep dqa alpha alpha
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
if (!N("noun",4))
fail();
@POST
L("tmp4") = N(4);
fixnphead(4);
fixnpnonhead(3);
group(2,4,"_np");
pncopyvars(L("tmp4"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND _clausesep _fnword
_prep)] # 05/31/07 AM.
@@
# prep dqan alpha alpha adv
@PRE
<6,7> var("noun"); # Have to be possible nouns.
@CHECK
L("last5") = lasteltnode(5);
if (!singular(L("last5"))
|| !singular(N(6))
|| !singular(N(7))
)
fail();
@POST
L("tmp6") = N(6);
L("tmp7") = N(7);
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
group(7,7,"_noun");
pncopyvars(L("tmp7"),N(7));
@RULES
_xNIL <-
_xWILD [one match=(_prep)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_xALPHA # (6)
_xALPHA # (7)
_xWILD [one lookahead match=(_adv _prep \, _qEOS _xEND
_fnword _clausesep)] ### (8)
@@
# prep dqa alpha alpha
# Note: Nibble off the edges.
@CHECK
if (N("pos50 to-j-a-a1",1))
fail();
@POST
N("pos50 to-j-a-a1",1) = 1;
if (N("noun",3) && N("adj",3))
{
if (plural(N(3)))
fixnphead(3);
else if (N("noun",4) && plural(N(4)))
{
fixnphead(4);
fixnpnonhead(3);
}
else
alphatoadj(3);
}
else if (N("adj",3))
alphatoadj(3);
else if (N("noun",3))
alphatonoun(3);
else if (N("verb",3))
; # Ignore for now.
else if (N("adv",3))
alphatoadv(3);
@RULES
_xNIL <-
to [s]
_adj
_xALPHA
_xALPHA
@@
# dqan alpha
# NIBBLE one alpha.
@PRE
<2,2> varz("adj");
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("verb",2) || !N("adj",2))
fail();
if (!vconjq(N(2),"-ed") && !vconjq(N(2),"-en"))
fail();
@POST
alphatoverb(2,0,0);
@RULES
_xNIL <-
_noun [plus]
_xALPHA
@@
# adj alpha adj
# dqa alpha dqan
@CHECK
if (!N("apos-s",1))
fail();
if (!N("adj",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_adj
_xALPHA
_adj [lookahead]
@@
# adj alpha adj noun
# dqan
@CHECK
if (!N("adj",2))
fail();
if (!singular(N(4)))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_adj
_xALPHA
_adj
_noun
@@
# prep dqa alpha
@CHECK
if (!N("noun",5) && !N("unknown",5))
fail();
if (N(2))
L("f") = N(2);
else if (N(3))
L("f") = N(3);
else if (N(4))
L("f") = N(4);
else
fail();
if (!numbersagree(L("f"),N(5)))
fail();
@POST
fixnphead(5);
@RULES
_xNIL <-
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# Don't like this rule. #
# prep dqa alpha
#@CHECK
# if (N("disagree",1))
# fail(); # Loop guard.
# if (!N("noun",4))
# fail();
# L("last") = lasteltnode(4);
# if (N(3))
# L("first") = N(3);
# else if (N(4))
# L("first") = N(4);
# else
# L("first") = N(5);
# if (!numbersagree(L("first"),L("last")))
# S("disagree") = 1;
#@POST
# if (S("disagree"))
# N("disagree",1) = 1;
# else
# {
# L("tmp4") = N(4);
# L("next") = pnnext(N(4));
# if (L("next"))
# {
# L("nm") = pnname(L("next"));
# if (L("nm") == "_noun" || L("nm") == "adj")
# L("dqan") = 1; # Inside a noun phrase.
# }
#
# if (N("verb",4))
# {
# if (vconjq(N(4),"-ing"))
# L("vadj") = 1;
# }
#
# if ((L("vadj") && L("dqan")) || (L("dqan") && N("adj",4)))
# {
# group(4,4,"_adj");
# pncopyvars(L("tmp4"),N(4));
# fixadj(N(4));
# }
# else
# {
# fixnphead(4);
# }
# }
#@RULES
#_xNIL <-
# _xWILD [one match=(_prep)]
# _proPoss [s opt]
# _xWILD [plus match=(_det _quan _num _xNUM _adj)]
# _xALPHA
# \, [opt lookahead]
# _xWILD [one fail=(_adj _noun _xALPHA _quan _num _xNUM)]
# @@
# people reported .
@CHECK
if (!N("verb",2))
fail();
if (N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg42",2) = 1;
@RULES
_xNIL <-
_noun
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND)]
@@
# more than alpha prep
#@CHECK
# if (!N("verb",3))
# fail();
# S("vc") = vconj(N(3));
# if (S("vc") != "-edn"
# && S("vc") != "-en")
# fail();
#@POST
# L("tmp3") = N(3);
# group(3,3,"_verb");
# pncopyvars(L("tmp3"),N(3));
# L("v") = N(3);
# group(3,3,"_vg");
# pncopyvars(L("tmp3"),N(3));
# N("verb node",3) = L("v");
# fixvg(N(3),"passive","VBN");
# clearpos(N(3),1,0);
#@RULES
#_xNIL <-
# _xWILD [one match=(_fnword)]
# _xWILD [star match=(_advl _adv)]
# _xALPHA
# _xWILD [one lookahead match=(_prep)]
# @@
# ^ that verb
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"DT"); # that/DT.
pnrename(N(2),"_np");
@RULES
_xNIL <-
_xSTART
that [s except=(_det _np)]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg)]
@@
# that alpha conj verb
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-s")
&& !vconjq(N(2),"-ed"))
fail();
@POST
alphatoverb(2,"active","VBP");
N("pos50 vg43",2) = 1;
@RULES
_xNIL <-
that [s]
_xALPHA
_conj
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# that alpha
@CHECK
if (N("mypos",1))
fail();
if (N("verb",2))
fail();
if (!N("noun",2))
fail();
if (!singular(N(2)))
fail();
@POST
pnrename(N(1),"_det"); # that -> det
chpos(N(1),"DT"); # that/DT
@RULES
_xNIL <-
that [s]
_xALPHA
_xWILD [one lookahead match=( \, _xEND)]
@@
# that alpha noun , dqan
@CHECK
if (!N("verb",2))
fail();
S("ed") = vconjq(N(2),"-ed");
if (!S("ed")
&& !vconjq(N(2),"-s"))
fail();
@POST
if (!N("mypos",1))
chpos(N(1),"IN"); # that/IN
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
if (S("ed"))
chpos(N(2),"VBD");
else
chpos(N(2),"VBZ");
@RULES
_xNIL <-
that [s]
_xALPHA
_noun [lookahead]
\,
_xWILD [one match=(_det _pro _quan _num _xNUM _adj _noun)]
@@
# that alpha dqan
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-s")
&& !vconjq(N(2),"inf"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg44",2) = 1;
@RULES
_xNIL <-
that [s]
_xALPHA
_xWILD [star lookahead match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one match=(_prep)]
@@
_xNIL <-
that [s]
_xALPHA
_pro [lookahead]
@@
# that books reservations
# that alpha dqan
@CHECK
if (!N("verb",2))
fail();
S("v") = vconj(N(2));
if (S("v") != "-s" && S("v") != "inf")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
# if (S("v") == "-s")
# chpos(N(2),"VBZ");
# else if (S("v") == "inf")
# chpos(N(2),"VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
# chpos(N(1),"IN"); # Not enough context. #
# N("bracket",1) = 1;
N("pos50 vg45",2) = 1;
@RULES
_xNIL <-
that [s]
_xALPHA
_xWILD [one lookahead match=(_det _quan _xNUM _num _np
# _adj _noun
)]
@@
# that alpha prep
@CHECK
if (!N("verb",3))
fail();
if (vconjq(N(3),"inf"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
fixverb(N(3),"active","VBP");
# if (!N("mypos",1))
# chpos(N(1),"IN"); # that/IN
N("pos50 vg46",3) = 1;
@RULES
_xNIL <-
that [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star match=(_adv _advl)]
_xWILD [one lookahead match=(_prep)]
@@
# that verb
@CHECK
if (N("mypos",2))
fail();
@POST
if (copulaq(N(3))
|| vconjq(N(3),"-s")
|| vconjq(N(3),"-ed")
|| vconjq(N(3),"-en")
|| pnname(N(3)) == "_modal")
chpos(N(2),"DT"); # that/DT
else
chpos(N(2),"IN"); # that/IN
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _conj _fnword)]
_xWILD [s one match=(that) except=(_np)]
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# , alpha alpha noun verb
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
@POST
fixnpnonhead(3);
fixnpnonhead(2);
@RULES
_xNIL <- #200
_xWILD [one match=( \, _clausesep _xSTART)]
_xALPHA
_xALPHA
_noun [lookahead]
_xWILD [one match=(_modal _verb _vg)]
@@
# , alpha alpha noun
# began eating
@CHECK
if (!N("verb",2) || !N("verb",3))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
# Want a composite vg...
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
chpos(N(3),"VBG");
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
group(2,3,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp3"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg47",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=( \, )]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun _np)]
@@
# , alpha alpha prep
@CHECK
if (N("pos50 comma-aa-prep",2))
fail();
@POST
N("pos50 comma-aa-prep",2) = 1;
if (!N("verb",2))
{
if (N("adj",2))
alphatoadj(2);
}
@RULES
_xNIL <-
_xWILD [one match=( \, _clausesep)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_xEND _prep)]
@@
# noun that alpha alpha noun
@CHECK
if (!N("verb",3))
fail();
if (!N("adj",4) && !N("noun",4))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
chpos(N(3),"VBG");
# chpos(N(2),"WDT"); # that/WDT #
chpos(N(2),"IN"); # that/IN # 04/23/07 AM.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
that [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun _np)]
@@
# alpha alpha noun
@CHECK
if (N("verb",2) || N("verb",3))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
# Todo: fail on noun-adj also.
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
else if (N("noun",3))
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
}
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
}
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# alpha alpha ,
@CHECK
if (N("verb",1) || N("verb",2))
fail();
if (!N("adj",1) && !N("noun",1) && !N("unknown",1))
fail();
if (!N("noun",2) && !N("unknown",2))
fail();
@POST
fixnpnonhead(1);
L("tmp2") = N(2);
# Todo: fixnphead()
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xALPHA
_xALPHA
_xWILD [one lookahead match=( \, _modal _verb _vg
_adv _advl)]
@@
# fnword alpha noun
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("verb",2) || N("adv",2))
fail();
@POST
L("tmp2") = N(2);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
_xWILD [one match=(_fnword _conj _prep)]
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# two times the
# num times det
# should be premod
@POST
if (literal(N(2)))
alphatonoun(2);
group(1,2,"_np");
group(1,1,"_advl");
N("ellipted-which-is",1) = 1;
N("n-times",1) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_xNUM _num _quan)]
_xWILD [s one match=(time times)]
_det [lookahead]
@@
# more quickly
# more adv
@POST
if (G("error"))
"err.txt" << "more quickly" << "\n";
L("tmp1") = N(1);
group(1,1,"_adv");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"RBR");
@RULES
_xNIL <-
_xWILD [s one match=(more less)
except=(_adj _adv _noun)] # Infinite loop guard.
_adv [lookahead]
@@
# most quickly
# most adv
@POST
L("tmp1") = N(1);
group(1,1,"_adv");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"RBS");
@RULES
_xNIL <-
_xWILD [s one match=(most)
except=(_adj _adv _noun)] # Infinite loop guard.
_adv [lookahead]
@@
# , alpha by noun to
# , alpha by some to
@CHECK
if (!N("verb",3))
fail();
S("vc") = vconj(N(3));
if (S("vc") != "-en"
&& S("vc") != "-edn")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
fixverb(N(3),"passive",0);
N("pos50 psv3",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=( \, )]
_xWILD [star match=(_adv _advl)]
_xALPHA
by [s lookahead]
_xWILD [one match=(_det _noun _np)]
_xWILD [s one match=(to _xEND)]
@@
# more than alpha by
@CHECK
if (!N("verb",3))
fail();
S("vc") = vconj(N(3));
if (S("vc") != "-en"
&& S("vc") != "-edn")
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
fixverb(N(3),"passive",0);
N("pos50 psv4",3) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(_morethan )]
_xWILD [star match=(_advl _adv)]
_xALPHA
by [s lookahead]
_xWILD [one match=(_det _noun _np)]
@@
# once np
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"IN"); # coordinating conjunction.
@RULES
_xNIL <-
once [s]
_xWILD [one lookahead match=(_np _noun _det
_quan _num _xNUM _adj)]
@@
# ... to them causing changes
# np alpha noun
@CHECK
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
fixverb(N(2),"active","VBP");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
clearpos(N(2),1,0);
N("pos50 vg48",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_np)]
_xALPHA
_xWILD [one lookahead match=(_noun _conj)]
@@
# this verb
# fnword verb
@CHECK
if (N("id",1) == "pos50 this-verb")
fail();
@POST
L("tmp1") = N(1);
chpos(N(1),"DT");
nountonp(1,1);
N("id",1) = "pos50 this-verb";
@RULES
_xNIL <-
this [s]
_xWILD [one lookahead match=(_verb _vg)]
@@
# pro verb
@POST
nountonp(2,1);
if (pnname(N(3)) == "_verb")
L("v") = N(3);
else if (pnname(N(3)) == "_vg")
L("v") = N("verb node",3);
if (L("v"))
fixverb(L("v"),"active","VBP");
N("pos50 vg49",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS _dblquote \,)]
_pro
_xWILD [one match=(_verb _vg)]
@@
# pro verb
@CHECK
if (N("pos50 pro-verb",2))
fail(); # Loop guard.
@POST
if (N("prosubj",1) && !N("proobj",1))
fixverb(N(2),"active","VBP");
else
fixverb(N(2),"active",0);
++N("pos50 pro-verb",2); # Loop guard.
@RULES
_xNIL <-
_pro
_verb
@@
# premod adj
@POST
pncopyvars(2);
sclearpos(1,0);
single();
@RULES
_adj <-
_premod
_adj
@@
# prep verb noun
# prep ven noun
# Changing verb to an adj
@PRE
<3,3> varz("done"); # Don't muck with this verb.
@CHECK
if (N("paap",1))
fail(); # Already reduced to verb...
S("vc") = vconj(N(3));
if (S("vc") != "-en"
&& S("vc") != "-edn"
&& S("vc") != "-ing")
fail();
if (S("vc") == "-ing" && !N("adj",3))
fail(); # 06/28/06 AM.
@POST
N("pos50 prep-verb-noun rename",3) = 1;
pnrename(N(3),"_adj"); # verb -> adj
if (S("vc") == "-en"
|| S("vc") == "-edn")
chpos(N(3),"VBN");
else
chpos(N(3),"VBG");
@RULES
_xNIL <-
_prep
_xWILD [star match=(_advl _adv _premod)]
_verb
_xWILD [one lookahead match=(_noun)]
@@
# alpha adj
# far greater
@CHECK
if (N("pos num",1) != 2)
fail();
if (!N("adj",1) || !N("adv",1))
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_adv");
pncopyvars(L("tmp1"),N(1));
pncopyvars(2);
sclearpos(1,0);
single();
@RULES
_adj <-
_xALPHA
_adj
@@
# dqan alpha
@CHECK
if (!N("noun",5))
fail();
if (N("verb",5))
fail();
L("num1") = number(N(1));
L("last") = lasteltnode(4);
L("num4") = number(L("last"));
if (L("num1") == L("num2"))
fail();
L("num5") = number(N(5));
if (L("num1") != L("num5"))
fail();
@POST
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
@@
# dqa alpha
# det adj alpha
@CHECK
if (!N("noun",3))
fail();
# Agreement...
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
L("fj") = N(2);
L("lj") = lasteltnode(2);
fixadjs(L("fj"),L("lj"));
group(1,3,"_np");
pncopyvars(L("tmp3"),N(1));
clearpos(N(1),1,1);
@RULES
_xNIL <-
_xWILD [s one match=(_det _proPoss)]
_adj
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND
_modal _vg _verb _pro
\, # 06/29/06 AM.
)]
@@
# dqan adj
# det adj
@POST
L("tmp2") = N(2);
if (N("cap",2))
{
if (pnname(N(3)) == "_verb")
{
L("m") = numberv(N(3));
if (N("sem",2) == "nationality" && L("m") == "plural")
N("mypos",2) = "NPS";
}
else
cappos(N(2),0);
}
group(1,2,"_np");
pncopyvars(L("tmp2"),N(1));
clearpos(N(1),1,1);
@RULES
_xNIL <-
_det
_adj
_xWILD [one lookahead match=(_prep _adv _verb)]
@@
# rather than
# X rather than Y
# X more than Y
# More than 3 people did X.
# *Rather than 3 people did X.
# "rather" is special, and is not a true
# comparative, though it looks like one.
# num more than
@POST
chpos(N(2),"JJR");
# Fuse the phrase at least.
group(2,3,"_morethan");
group(2,2,"_fnword");
N("morethan",2) = 1;
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [s one match=(_quan _num _xNUM _money)]
_xWILD [s one match=(more less fewer)]
_xWILD [s one match=(than)]
@@
# more than tripled
@POST
chpos(N(1),"JJR"); # conform treebank. 75% more/jjr than/
group(1,2,"_morethan");
group(1,1,"_adv");
clearpos(N(1),1,0);
if (pnname(N(2)) == "_vg")
fixvg(N(2),"active","VBP");
else
fixverb(N(2),"active","VBP");
@RULES
_xNIL <-
more [s]
than [s]
_xWILD [one lookahead match=(_verb _vg)]
@@
# more than a
@POST
chpos(N(1),"JJR");
group(1,2,"_det"); # 04/21/07 AM. Better than _prep.
clearpos(N(1),1,0);
@RULES
_xNIL <-
_xWILD [s one match=(more less fewer)]
than [s]
a [s]
@@
# Treating as default. #
# more than num
# Not certain, but usually parse this way.
@POST
chpos(N(2),"JJR");
if (N(1))
{
group(1,3,"_det"); # 05/08/07 AM.
clearpos(N(1),1,0);
}
else
{
group(2,3,"_advl"); # 06/02/06 AM.
clearpos(N(2),1,0);
}
@RULES
_xNIL <-
no [s opt]
_xWILD [s one match=(more less fewer)]
than [s]
# _xWILD [one match=(_xNUM _num _quan)]
@@
# verb more
@CHECK
if (N("mypos",2))
fail(); # Loop guard.
@POST
pnrename(N(2),"_adj");
chpos(N(2),"JJR"); # conform treebank. # 06/17/06 AM.
@RULES
_xNIL <-
_xWILD [s one match=(_verb _vg)]
_xWILD [s one match=(more less) except=(_adj)]
_xWILD [one fail=(_adj _xALPHA _adv)]
@@
# more adv
# more adj
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"RBR");
pnrename(N(1),"_adv");
@RULES
_xNIL <-
_xWILD [s one match=(more less)]
_xWILD [one lookahead match=(_adv _advl _adj)]
@@
# most adv
@POST
if (!N("mypos",1))
chpos(N(1),"RBS");
if (pnname(N(1)) != "_adv")
pnrename(N(1),"_adv");
L("nm") = pnname(N(2));
group(1,2,L("nm"));
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(most least)]
_xWILD [one match=(_adv _advl
# _adj # Still ambig. #
)]
@@
# more and more
@CHECK
if (!N("mypos",3))
fail();
if (N("mypos",1))
fail();
@POST
chpos(N(1),N("mypos",3));
if (N("mypos",3) == "JJR")
{
group(1,3,"_adj");
clearpos(N(1),1,0);
}
@RULES
_xNIL <-
more [s]
_conj
more [s]
@@
# more noun
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"JJR");
pnrename(N(1),"_adj");
@RULES
_xNIL <-
_xWILD [s one match=(more less)]
_xWILD [one lookahead match=(_noun _np)]
@@
# Probably not enough context. #
# pattern...
# noun that alpha prep
# that
@PRE
<3,3> var("verb");
@CHECK
if (vconjq(N(3),"-ing"))
fail();
if (vconjq(N(3),"-en") && !vconjq(N(3),"-ed"))
fail();
# Todo: noun-verb agreement.
@POST
alphatovg(3,"active","VBP");
L("pos") = vgpos(N(3));
if (!L("pos") || N("mypos",2))
;
else if (L("pos") == "VBG")
chpos(N(2),"IN");
else
chpos(N(2),"WDT");
N("bracket",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _advl)]
that [s]
_xALPHA
_xWILD [lookahead one match=(_prep _qEOS _xEND)]
@@
# that vg
# np that vg
# Broad, but could be pretty good. #
# Advl should be a prepositional phrase.
@CHECK
if (N("mypos",2))
fail();
if (pnname(N(1)) == "_advl")
if (!N("pp",1))
fail();
if (!vconjq(N(4),"-s")
&& !vconjq(N(4),"-ed"))
fail();
@POST
chpos(N(2),"WDT"); # that/WDT.
L("xx") = pnparent(X()); # 07/10/12 AM.
if (pnvar(L("xx"),"date=present")) # 07/10/12 AM.
fixverb(N(4),"active","VBP");
else
fixverb(N(4),"active","VBD");
N("pos50 fixv1",4) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _advl)]
that [s]
_xWILD [star match=(_advl _adv)] # 04/24/10 AM.
_xWILD [one match=(_modal _verb _vg)]
@@
# vg np that alpha alpha+ prep
@PRE
<5,5> var("verb");
@CHECK
S("n") = lasteltnode(6);
if (!pnvar(S("n"),"noun"))
fail();
if (!plural(S("n")))
fail();
@POST
alphatovg(5,"active","VBP");
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
_xWILD [star match=(_adv _advl)] ### (3)
that [s] ### (4)
_xALPHA ### (5)
_xALPHA [plus] ### (6)
_prep ### (7)
_xWILD [one lookahead match=(_noun _np _adj _det _pro _quan
_num _xNUM)] ### (8)
@@
# vg np that alpha
@PRE
<4,4> var("verb");
@CHECK
if (!vconjq(N(4),"-s") && !vconjq(N(4),"inf"))
fail();
@POST
alphatovg(4,"active","VBP");
if (!N("mypos",3))
{
L("pos") = vgpos(N(4)); # See what was assigned.
if (L("pos") == "VBG")
chpos(N(3),"IN"); # that/IN
else
chpos(N(3),"WDT"); # that/WDT.
}
@RULES
_xNIL <-
_vg
_np
that [s]
_xALPHA
_xWILD [one lookahead match=(_adv _advl)]
@@
# that alpha alpha $
@PRE
<2,2> varz("pos50 that-aa-p");
@POST
N("pos50 that-aa-p",2) = 1;
# Leave pos of "that" for something with more context.
if (!N("verb",2) && !N("verb",3))
{
}
else if (!N("verb",2))
alphatovg(3,"active","VBP");
else if (!N("verb",3))
alphatovg(2,"active","VBP");
# Todo: Else can elaborate further.
@RULES
_xNIL <-
that [s]
_xALPHA
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_qEOS _xEND)]
@@
# that alpha alpha verb
# Note: Not worrying about that/det vs others.
@CHECK
if (!N("noun",4))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
@POST
fixnphead(4);
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_fnword \, _clause )]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# adv alpha
@CHECK
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_adv
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM)]
@@
# pro alpha and adj
@CHECK
if (!N("adj",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS _pro _det _quan _num _xNUM
# _adj # Too chancy. #
_adv \, )]
_xALPHA
_conj [lookahead]
_adj
@@
# adj alpha conj adj
# Ambiguous, will choose noun for simplicity.
@CHECK
if (!N("noun",2) || !N("adj",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_adj
_xALPHA
_conj [lookahead]
_adj
@@
# alpha conj alpha
@CHECK
if (!N("adj",3) || !N("adj",5))
fail();
@POST
L("tmp3") = N(3);
L("tmp5") = N(5);
group(5,5,"_adj");
pncopyvars(L("tmp5"),N(5));
fixadj(N(5));
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
group(3,5,"_adj");
@RULES
_xNIL <-
_xWILD [plus match=(_pro _det _quan _num _xNUM)]
_adj [star]
_xALPHA
_conj
_xALPHA
_xWILD [one lookahead match=(_noun _adj)]
@@
# Sentinel to keep "pro alpha any" from matching
# if there's a conjunction possibility.
# Don't know if alpha is a noun or adj yet.
@PRE
<1,1> varz("pos50 pro-a-c-a");
@POST
N("pos50 pro-a-c-a",1) = 1;
@RULES
_xNIL <-
_proPoss [s]
_xALPHA
_conj
_xALPHA
@@
# pro alpha
@CHECK
if (!N("adj",2) && !N("noun",2) && !N("unknown",2))
fail();
@POST
L("tmp2") = N(2);
L("nm") = pnname(N(3));
if (L("nm") == "_adj")
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (L("nm") == "_noun" && N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
_proPoss [s]
_xALPHA
_xWILD [one lookahead fail=(_xALPHA)]
@@
# prep this adj
@CHECK
if (N("mypos",2))
fail();
@POST
pnrename(N(2),"_det");
chpos(N(2),"DT"); # that/DT
@RULES
_xNIL <-
_prep
this [s]
_xWILD [one lookahead match=(_adj _quan _xNUM _num _noun)]
@@
# noun , dqan 's alpha alpha alpha noun ,
# Apposition.
@CHECK
if (!N("adj",8) && !N("noun",8))
fail();
if (N(9))
if (!N("adj",9) && !N("noun",9))
fail();
if (N(10))
if (!N("adj",10) && !N("noun",10))
fail();
@POST
if (N(10))
fixnpnonhead(10);
if (N(9))
fixnpnonhead(9);
fixnpnonhead(8);
chpos(N(7),"POS"); # 's/POS
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [s]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_aposS
_xALPHA
_xALPHA [opt]
_xALPHA [opt]
_noun [plus]
_xWILD [one lookahead match=(_qEOS _xEND \, _clausesep
_fnword)]
@@
# noun , dqa alpha alpha
# apposition
@CHECK
# if (N("mypos",1) != "NP") # check proper noun.
# fail();
if (!N("adj",4) && !N("verb",4) && !N("noun",4))
fail();
if (!N("noun",5))
fail();
S("vc4") = vconj(N(4));
if (S("vc4") == "-s"
|| S("vc4") == "-ed"
|| S("vc4") == "inf")
fail();
@POST
L("tmp4") = N(4);
L("tmp5") = N(5);
if (N("noun",5))
{
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
}
else
{
group(5,5,"_adj");
pncopyvars(L("tmp5"),N(5));
fixadj(N(5));
}
if (N("adj",4))
{
group(4,4,"_adj");
pncopyvars(L("tmp4"),N(4));
fixadj(N(4));
}
else if (N("verb",4))
{
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
fixverb(N(4),"passive","VBP");
N("pos50 psv5",4) = 1;
}
else
{
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
}
group(3,5,"_np");
pncopyvars(L("tmp5"),N(3));
clearpos(N(3),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [s]
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND \, _clausesep
_fnword)]
@@
# noun , dqan alpha alpha ,
# apposition.
@CHECK
if (!N("noun",5) || !N("noun",6))
fail();
if (!singular(N(5)) || !singular(N(6)))
fail();
@POST
fixnphead(6);
fixnpnonhead(5);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xWILD [one match=(_det _pro)]
_noun [plus]
_xALPHA
_xALPHA
\, [lookahead]
@@
# vg dqan alpha alpha noun
@PRE
<1,1> varz("pos50 v-dqan-aa-n");
@POST
N("pos50 v-dqan-aa-n",1) = 1;
if (N("noun",6) && N("verb",7))
{
if (vconjq(N(7),"-ing"))
{
fixnphead(6);
alphatovg(7,"active","VBG");
}
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xALPHA
_noun [lookahead]
@@
# vg dqa alpha alpha
@PRE
<2,2> var("proposs");
<5,5> varz("pos50 v-dqa-aa*");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail(); # for now, ignore...
@POST
N("pos50 v-dqa-aa*",5) = 1; # Loop guard.
if (!N("noun",5) && !N("adj",5))
{
# Handle ving, ven, verb patterns, etc.
}
else if (N("adj",5) && !N("noun",5))
alphatoadj(5);
else if (N("noun",5) && !N("adj",5))
alphatonoun(5);
else if (plural(N(5)))
fixnphead(5);
else if (N("noun",6))
{
if (plural(N(6)))
{
fixnphead(6);
fixnpnonhead(5);
}
else
fixnpnonhead(5);
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_xALPHA ### (5)
_xALPHA ### (6)
_xALPHA [star] ### (7)
\, [opt lookahead] ### (8)
_adv [star] ### (9)
_xWILD [one match=(_prep)] ### (10)
@@
# noun , dqa alpha
@PRE
<6,6> var("noun");
@POST
fixnphead(6);
@RULES
_xNIL <- #250
_xWILD [one match=(_np _noun)]
\,
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [plus]
_xALPHA
_xWILD [one lookahead match=(_adv _advl \( \, )]
@@
# conj that the
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"IN"); # that/IN
@RULES
_xNIL <-
_xWILD [one match=(_conj)]
that [s]
_xWILD [one lookahead match=(_det)]
@@
# dqan alpha
@CHECK
if (!N(2) && !N(3) && !N(4) && !N(5))
fail();
if (!N("noun",7))
fail();
if (N("verb",7))
{
if (!vconjq(N(7),"inf"))
fail();
if (vconjq(N(7),"-ed")) # 04/21/07 AM.
fail();
}
@POST
L("tmp7") = N(7);
group(7,7,"_noun");
pncopyvars(L("tmp7"),N(7));
fixnoun(N(7));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [opt]
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
@@
# TRYING TO ACCEPT NEW PARTS OF SPEECH NOT IN DICT.
# Mangle preassigned pos sometimes.
# may noun up
# may noun down
# he nouns up
# he nouns up
@CHECK
if (number(N(2)) == "plural")
fail();
@POST
pnrename(N(2),"_verb"); # noun -> verb.
chpos(N(2),"VB");
chpos(N(3),"RP");
@RULES
_xNIL <-
_modal
_noun
_xWILD [s one lookahead match=(up down)]
@@
# In this case, check agreement etc.
@CHECK
# if (!numbersagree(N(1),N(2)))
# fail();
if (N("mypos",2))
fail();
@POST
if (number(N(2)) == "plural")
chpos(N(2),"VBZ");
else
chpos(N(2),"VBP");
chpos(N(3),"RP");
pnrename(N(2),"_verb"); # noun -> verb
@RULES
_xNIL <-
_proSubj [s]
_noun
_xWILD [s one lookahead match=(up down)]
@@
# , said
@POST
if (vconjq(N(3),"-ed"))
alphatovg(3,"active","VBD");
else
alphatovg(3,"active","VBP");
N("glom",3) = "left";
clearpos(N(3),1,0);
if (L("next") = pnnext(N(3)))
pnreplaceval(L("next"),"glom","right");
@RULES
_xNIL <-
_xWILD [one match=( \, _dblquote _qEOS _xSTART _conj _clause
_noun _np)]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(say says said quip quips quipped
recalls recalled reads read)]
@@
# noun , alpha to
@PRE
<4,4> var("verb");
<4,4> var("adj");
<4,4> vareq("pos num",2);
@CHECK
if (!vconjq(N(4),"-en"))
fail();
@POST
alphatovg(4,"passive","VBN");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [s]
_adv [star]
_xALPHA
_xWILD [s one lookahead match=(to)]
@@
# noun , alpha prep
@CHECK
if (!N("verb",4))
fail();
if (N("adv",4))
fail();
if (!phrprepverbq(N(4),N(5)))
fail(); # 05/22/07 AM.
@POST
alphatovg(4,0,0); # Cleanup. # 05/27/07 AM.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\, [s]
_adv [star]
_xALPHA
_xWILD [s one lookahead match=(_prep as)
except=(of)]
@@
# Initial adverbial.
# adv ,
@POST
S("no subj") = 1;
S("pattern") = "advl";
single();
@RULES
_clause [unsealed] <-
_xSTART
_adv
\,
@@
# down from
# up from
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"RB");
@RULES
_xNIL <-
down [s]
_prep [lookahead]
@@
# Complex apposition
# np prep np , noun ,
# Assume intervening adverbial.
@POST
L("tmp3") = N(3);
group(3,6,"_np"); # Appositive np-np.
if (N("stem",2) != "by")
{
group(2,3,"_advl");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,0);
N("pp",2) = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(_np _noun)]
_prep
_np
\, # 04/24/10 AM.
_xWILD [one match=(_noun _np)]
\,
_xWILD [one lookahead fail=(_noun _np _xALPHA _det _pro _quan _num
_xNUM )]
@@
# including np vg
@PRE
<4,4> varz("voice");
@CHECK
if (!vconjq(N(4),"-en"))
fail();
@POST
fixvg(N(4),"passive","VBN");
@RULES
_xNIL <-
_xWILD [one match=(\, )]
including [s]
_np
_vg
@@
# prep np
@POST
L("year") = N("year",3);
group(2,3,"_advl");
N("pp",2) = 1;
if (L("year")) N("year",2) = L("year");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART \, _clausesep)]
_prep
_np
_xWILD [one fail=(_verb # want verbs to go to vg first.
_conj _prep)]
@@
# , prep noun ,
@POST
L("year") = N("year",3);
nountonp(3,1);
group(2,3,"_advl");
N("pp",2) = 1;
if (L("year")) N("year",2) = L("year"); # 02/13/14 AM.
@RULES
_xNIL <-
\,
_prep
_noun
\, [lookahead]
_xWILD [one fail=(_noun _xALPHA)]
@@
# watch for lists...
# noun , noun ,
# Apposition.
@CHECK
L("prev") = pnprev(N(1));
if (L("prev"))
{
L("nm") = pnname(L("prev"));
if (L("nm") == ",")
fail();
}
if (N("cap",1) && N("cap",4))
succeed();
if (N("mypos",1) != "NP"
&& N("mypos",4) != "NP")
fail();
@POST
N("apposition",1) = 1;
N("pos50 n-n",1) = 1;
# Don't know about that last comma yet.
group(1,4,"_np");
clearpos(N(1),1,0);
N("compound-np",1) = 1;
@RULES
_xNIL <-
_np # 06/07/05 AM.
_xWILD [star match=(_advl _adv)]
\,
_xWILD [one match=(_noun _np)]
\,
_xWILD [one lookahead fail=(_noun _np)]
@@
# Apposition.
# ^ dqan , dqan ,
@POST
dqaninfo(7,8,9,10);
L("np2") = groupnp();
dqaninfo(2,3,4,5);
L("np1") = groupnp();
group(2,4,"_np");
clearpos(N(2),1,0);
N("apposition",2) = 1;
N("compound-np",2) = 1;
N("commas-in-here",2) = 1;
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_det)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
\, ### (6)
_xWILD [star match=(_det)] ### (7)
_xWILD [star match=(_quan _num _xNUM)] ### (8)
_adj [star] ### (9)
_noun [plus] ### (10)
\, ### (11)
@@
# ^ dqa alpha ,
# Apposition
@PRE
<5,5> var("noun");
@POST
fixnphead(5);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_xALPHA
\,
@@
# Apposition. #
# noun , num ,
@CHECK
if (!N("ne",1) && !N("ne arr",1))
fail();
@POST
N("apposition",1) = 1;
listadd(1,3,"true");
@RULES
_xNIL <-
_np
\,
_xNUM
\, [lookahead]
@@
# Possible apposition. #
# noun , dqan ,
@CHECK
if (!N("noun",8) && !N("adj",8))
succeed();
L("n") = lasteltnode(6);
if (pnvar(L("n"),"cap"))
fail(); # Could be part of bigger np.
@POST
# Todo: reduce to np first.
dqaninfo(3,4,5,6);
groupnp();
@RULES
_xNIL <-
_np
\,
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
\, [lookahead]
_xANY
@@
# Apposition. #
# noun , alpha ,
@CHECK
if (!N("noun",3))
fail();
@POST
N("apposition",1) = 1;
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
listadd(1,3,"true");
@RULES
_xNIL <-
_np
\,
_xALPHA
_xWILD [one lookahead match=( \, _xEND)]
@@
_xNIL <-
_xWILD [one match=(_np)]
\,
_xALPHA
_xWILD [s one lookahead match=(of)]
@@
@POST
nountonp(2,1);
L("tmp2a") = N(2);
group(2,5,"_np");
pncopyvars(L("tmp2a"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_noun
\,
_np
\,
@@
# noun verb
# noun vg
@POST
nountonp(2,1);
if (pnname(N(3)) == "_vg")
if (!N("voice",3))
N("voice",3) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_noun
_xWILD [one lookahead match=(_vg _verb)]
@@
# clause-like pattern.
# noun , verb
# np , vg
@PRE
<4,4> varz("pos50 n-comma-v");
@POST
if (N("voice",4))
; # Done.
else if (pnname(N(4)) == "_vg")
{
if (N("commas-in-here",2))
fixvg(N(4),"active","VBP");
else
fixvg(N(4),"passive","VBN");
}
else
{
if (N("commas-in-here",2))
fixverb(N(4),"active","VBP");
else
fixverb(N(4),"passive","VBN");
}
N("pos50 n-comma-v",4) = 1;
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_noun _np)]
\, [s]
_xWILD [one match=(_verb _vg)]
@@
# np np (date)
@CHECK
if (!N("advl",2)) # 04/15/05 AM.
fail();
@POST
L("tmp2") = N(2);
if (pnname(N(2)) == "_noun")
{
nountonp(2,1);
}
group(2,2,"_advl");
pncopyvars(L("tmp2"),N(2));
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _adv _advl)]
_xWILD [one match=(_noun _np)]
@@
# mr. alpha
@CHECK
if (N("mypos",2) != "NP")
fail();
if (!N("unknown",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
chpos(N(3),"NP");
L("tmp3n") = N(3);
group(2,3,"_np");
pncopyvars(L("tmp3n"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _verb _vg)]
_noun
_xALPHA
_xWILD [one lookahead match=(
_prep _verb _vg _adv _advl _np \, _qEOS _xEND)]
@@
# alpha inc.
@CHECK
if (N("mypos",3) != "NP")
fail();
if (!N("unknown",2))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"NP");
L("tmp2n") = N(2);
group(2,3,"_np");
pncopyvars(L("tmp2n"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep _verb _vg _adv _advl _noun _np \, )]
_xALPHA
_noun
_xWILD [one lookahead match=( \, _qEOS _xEND _prep _verb _vg
_np)]
@@
# those noun
# these noun
# Slough fnword.
@CHECK
if (N("id",1) == "pos50 those-n")
fail(); # Infinite loop guard.
@POST
if (pnname(N(1)) == "_fnword")
pnrename(N(1),"_det");
N("id",1) = "pos50 those-n";
@RULES
_xNIL <-
_xWILD [s one match=(these those)]
_noun [lookahead]
@@
# those verb
@CHECK
if (N("mypos",2))
fail();
if (N("pos50 those-verb",1))
fail(); # Loop guard.
@POST
N("pos50 those-verb",1) = 1;
fixverb(N(2),"passive","VBN");
N("pos50 psv6",2) = 1;
@RULES
_xNIL <-
those [s]
_xWILD [one match=(_verb) except=(_have _be)]
@@
# prep num
@POST
if (!N("mypos",1))
{
if (N("stem",1) == "up"
|| N("stem",1) == "down")
chpos(N(1),"RB");
}
nountonp(2,1);
group(1,2,"_advl");
N("pp",1) = 1;
@RULES
_xNIL <-
_prep
_xWILD [one match=(_xNUM _num _quan)]
_xWILD [s one lookahead match=(_fnword _qEOS _xEND)
except=(more less)]
@@
# np of np
@POST
L("det stem") = N("det stem",1);
if (N("mypos",1) == "NP" && N("mypos",3) == "NPS")
setpos(N(3),"NP");
if (N("cap",1) && N("cap",3))
L("cap") = 1;
L("stem") = N("stem",3);
if (!L("stem"))
L("stem") = N("stem",1);
L("arr") = semnpnp(N(1),N(3));
L("sem") = L("arr")[0];
L("ne text") = L("arr")[1];
L("tmp1") = N(1);
L("tmp3") = N(3);
group(1,3,"_np");
domnpofnp(N(1),L("tmp1"),L("tmp3"));
# semnpofnp(N(1),L("tmp1"),L("tmp3"));
if (L("cap"))
N("cap",1) = 1;
if (L("sem"))
N("sem",1) = L("sem");
if (L("ne text"))
N("ne text",1) = L("ne text");
if (L("stem"))
N("stem",1) = L("stem");
if (L("det stem"))
N("det stem",1) = L("det stem");
@RULES
_xNIL <-
_np
of [s]
_np
_xWILD [one fail=(_conj)]
@@
# prep np of np
@POST
group(2,4,"_np");
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xWILD [one match=(_noun _np)]
of [s]
_xWILD [one match=(_noun _np)]
_xWILD [one lookahead match=(_adv _advl _clausesep \, )]
@@
# as np vg
# as noun verb
# Note: as - conjunction.
@PRE
<3,3> varz("renamed"); # Already renamed.
@POST
pnrename(N(3),"_fnword"); # as -> fnword
@RULES
_xNIL <-
_xWILD [one match=(_pro _np)]
_xWILD [star match=(_adv _advl)]
as [s except=_fnword] # Loop guard.
_np
_xWILD [one match=(_modal _verb _vg)]
@@
# np prep np verb prep
# Note: Check for verb+prep idiom...
@CHECK
if (pnvar(N(5),"mypos"))
fail();
if (phrprepverbq(N(5),N(6)))
fail();
@POST
L("v") = N(5);
group(5,5,"_vg");
pncopyvars(L("tmp5"),N(5));
clearpos(N(5),1,0);
N("verb node",5) = N("first verb",5) = L("v");
fixvg(N(5),0,0);
@RULES
_xNIL <-
_xWILD [one match=(_pro _np)]
_xWILD [star match=(_adv _advl)]
_prep
_np
_verb
_prep [lookahead]
@@
# prep np $
@POST
L("tmp3") = N(3);
group(2,3,"_advl");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,0);
N("pp",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_adjc)]
_prep
_np
_xEND
@@
# NOTE: Trying to defer this. #
# by np
#@POST
# L("tmp2") = N(2);
# group(1,2,"_advl");
# pncopyvars(L("tmp2"),N(1));
# clearpos(N(1),1,0);
# N("pp",1) = 1;
# if (N("sem",1) != "date" && N("sem",1) != "geoloc"
# && N("sem",1) != "quantity")
# N("by-actor",1) = 1; #
# N("by-np") = 1; #
#@RULES
#_xNIL <-
# by [s]
# _np
# _xWILD [one fail=(_conj)]
# @@
# A clausal look.
# vg np be
@CHECK
if (N("id",1) == "pos50 v-n-be")
fail(); # Infinite loop guard.
if (N("q-q remote",1) == "right")
fail();
@POST
N("glom",1) = "left";
N("glom",2) = "right";
N("id",1) = "pos50 v-n-be";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one lookahead match=(_noun _np _pro)]
_be [s]
@@
# np pro vg
@POST
nountonp(2,1);
N("glom",2) = "right";
@RULES
_xNIL <-
_xWILD [one match=(_np)]
_pro
_xWILD [one lookahead match=(_vg)]
@@
# np pro alpha
@PRE
<2,2> var("prosubj");
<3,3> var("verb");
@POST
N("ellipted-that",2) = 1;
alphatovg(3,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_pro
_xALPHA
_xWILD [one lookahead match=(_verb _np _det _pro _quan _num _xNUM
_adj _noun)]
@@
# money
@POST
S("sem") = "money";
single();
@RULES
_np <-
_xWILD [one match=(\$ \#)]
_xWILD [plus match=(_num _quan _xNUM)]
_noun [plus]
@@
#@POST
# S("sem") = "money";
# singler(1,2);
#@RULES
#_np <-
# \$
# _xWILD [plus match=(_num _quan _xNUM)]
# _xWILD [one lookahead fail=(_xALPHA)]
# @@
# noun & noun
@RULES
_noun <-
_noun
\& [s]
_noun
@@
# $5 a share
@CHECK
if (N("pos50 $-a-share",2) = 1)
fail();
@POST
pnrename(N(2),"_prep"); # as in "$5 per share".
N("pos50 $-a-share",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
a [s]
_noun
@@
# to verb
# to infinitive...
@CHECK
if (N("mypos",3))
fail();
if (N("pos50 to-verb-inf",3))
fail();
@POST
if (vconjq(N(3),"inf"))
chpos(N(3),"VB");
N("pos50 to-verb-inf",3) = 1;
@RULES
_xNIL <-
to [s]
_adv [star]
_verb [lookahead]
@@
# vg np prep adj
# vg np as adj
# Note: if a sample such as "viewed John as happy",
# then flag no renaming of the adj.
@PRE
<3,3> vareq("stem","as");
<4,4> varz("no-rename"); # Loop guard.
@CHECK
if (!verbfeat(N(1),"C1pr"))
fail();
@POST
N("no-rename",4) = 1; # Don't rename adj.
@RULES
_xNIL <-
_vg
_np
_prep
_adj
@@
# prep adj
@PRE
<2,2> varz("possessive");
<2,2> varz("no-rename");
@POST
pnrename(N(2),"_noun"); # adj -> noun
# chpos(N(2),0); # Clear out adj etc.
fixnoun(N(2)); # adj -> noun
N("rename-adj-noun",2) = 1;
@RULES
_xNIL <-
_prep
_adj
_xWILD [star match=( \, \" _dblquote)]
_xWILD [one lookahead fail=(_np _adj _noun
_quan _num _xNUM _xALPHA _dblquote)]
@@
# Some transitive verbs.
# Avoiding hardwiring such rules in general.
# Want the analyzer to identify these automatically....
# verb prep
@CHECK
if (N("pos50 verb-prep",1))
fail();
@POST
L("nm") = pnname(N(1));
if (strpiece(L("nm"),0,0) == "_")
{
# nonliteral.
if (L("nm") == "_vg")
{
N("voice",1) = "passive";
L("v") = N("verb node",1);
}
else
L("v") = N(1);
if (L("v"))
{
chpos(L("v"),"VBN");
pnreplaceval(L("v"),"voice","passive");
}
}
else
{
L("tmp1") = N(1);
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"VBN");
}
N("pos50 verb-prep",1) = 1; # Loop guard.
@RULES
_xNIL <-
_xWILD [s one match=(found)]
_xWILD [star lookahead match=(_adv _advl)]
_prep
@@
# prep dqan alpha
# unless something happens...
@CHECK
if (!N("verb",5))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_verb");
pncopyvars(L("tmp5"),N(5));
L("v") = N(5);
group(5,5,"_vg");
mhbv(N(5),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp5"),N(5));
N("verb node",5) = L("v");
clearpos(N(5),1,0);
fixvg(N(5),"active","VBP");
@RULES
_xNIL <-
_xWILD [s one match=(until unless _fnword)]
_xWILD [star match=(_det _quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=( \, _qEOS _xEND _conj)]
@@
# prep dqan alpha
# for the boy scout
@CHECK
if (!N("noun",5))
fail();
if (N(2))
S("first") = N(2);
else if (N(3))
S("first") = N(3);
else
S("first") = N(4);
S("head") = lasteltnode(4);
if (N("verb",5))
{
if (!vconjq(N(5),"inf"))
fail();
}
if (!numbersagree(S("first"),S("head"))
|| numbersagree(S("first"),N(5)))
succeed();
if (N("advl",5)) # 04/15/05 AM.
fail();
# If last noun is plural, no good. #
L("txt") = nodetext(S("head"));
if (finddictattr(L("txt"),"soft-plural"))
succeed();
if (plural(S("head")))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
# Fix the range of nouns.
# fixnouns(N(4),N(5));
# group(2,5,"_np");
# pncopyvars(L("tmp5"),N(2));
# clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [s one match=(for with of at in on under by)]
_xWILD [star match=(_det _quan _num _adv _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [star match=(_adv _advl)]
_xWILD [one match=( \, _qEOS _xEND
_modal _verb _vg # 06/29/06 AM.
_prep _fnword # 07/10/06 AM.
_conj
)]
@@
# prep dqan alpha
# for the boy scout
@CHECK
if (!N("noun",5))
fail();
if (N("advl",5)) # 04/15/05 AM.
fail();
if (N("adj",5)) # 06/04/06 AM.
fail();
# Not enough context...
if (N("verb",5))
{
if (!vconjq(N(5),"inf")) # 07/20/07 AM.
fail(); # 08/24/06 AM.
}
# If last noun is plural, no good. #
S("head") = lasteltnode(4);
if (plural(S("head")))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
# Fix the range of nouns.
# fixnouns(N(4),N(5));
# group(2,5,"_np");
# pncopyvars(L("tmp5"),N(2));
# clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [s one match=(for with of at in under by)]
_xWILD [star match=(_det _quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=( _conj _noun _np)]
# _xWILD [s one match=(which that)]
@@
# prep alpha noun
# by alpha noun
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
alphatoverb(2,"active","VBG");
@RULES
_xNIL <-
by [s]
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# prep dqa alpha noun
@CHECK
if (!N(2) && !N(3)
&& N("verb",4))
{
if (vconjq(N(4),"-ing"))
fail();
}
if (!N("noun",4) && !N("adj",4))
fail();
@POST
fixnpnonhead(4);
@RULES
_xNIL <-
_xWILD [s one match=(for with of at in under by)]
_xWILD [star match=(_det _quan _num _xNUM)]
_adj [star]
_xALPHA
_noun [lookahead]
@@
# prep adj alpha alpha
# of
@CHECK
if (!N("adj",3) && !N("noun",3))
fail();
if (!N("noun",4))
fail();
if (!singular(N(3)))
fail();
if (!plural(N(4)))
fail();
@POST
fixnphead(4);
fixnpnonhead(3);
@RULES
_xNIL <- #300
_prep
_adj
_xALPHA
_xALPHA
_xWILD [one match=( \, )]
@@
# prep dqa alpha
# prep adj alpha
@CHECK
if (!N("noun",4) && !N("unknown",4))
fail();
@POST
fixnphead(4);
@RULES
_xNIL <-
_prep
_adv [plus]
_adj
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# that noun alpha
@CHECK
if (!N("verb",3) || N("noun",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [s one match=(which that)]
_xWILD [one match=(_noun _np)]
_xALPHA
@@
# prep noun alpha
# Handle mass nouns...
@CHECK
if (!N("noun",3))
fail();
if (N("adj",3))
fail(); # 06/04/06 AM.
S("number") = number(N(2));
if (S("number") != "singular")
fail();
if (N("verb",3))
{
# Need some context. #
if (N("cap",2) || N("ne",2) || N("ne arr",2))
fail();
if (!N("verbal",1) # Not a "verbal" preposition. # 06/08/06 AM.
&& N("mass",2)) # And a mass noun.
fail(); # 06/08/06 AM.
S("vc") = vconj(N(3));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
}
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
# fixnoun(N(3));
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_noun
_xALPHA
# _xWILD [one lookahead fail=(_xALPHA _np
# _pro _adv)] #
@@
# prep noun alpha
@CHECK
if (!singular(N(2)))
fail();
if (!N("noun",3))
fail();
if (N("verb",3))
{
# Todo: could broaden this.
if (!vconjq(N(3),"-ing"))
fail();
}
@POST
fixnphead(3);
@RULES
_xNIL <-
_prep
_noun
_xALPHA
_xWILD [one fail=(_noun _xALPHA)]
@@
# dqa alpha alpha noun
# adj alpha alpha noun
@CHECK
if (N("verb",2) || N("verb",3))
fail();
if (!N("adj",2) && !N("noun",2) && !N("unknown",2))
fail();
if (!N("adj",3) && !N("noun",3) && !N("unknown",3))
fail();
@POST
fixnpnonhead(2);
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_det _pro _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# dqa alpha alpha
@CHECK
if (!N("noun",3) && !N("unknown",3))
fail();
if (!N("verb",4))
fail();
@POST
L("tmp3") = N(3);
L("tmp4") = N(4);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
L("v") = N(4);
group(4,4,"_vg");
mhbv(N(4),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp4"),N(4));
N("verb node",4) = L("v");
clearpos(N(4),1,0);
fixvg(N(4),"active","VBP");
group(2,3,"_np");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM)]
@@
# dqan alpha alpha dqan
@CHECK
if (N("pos50 dqan-aa-dqan",5))
fail();
@POST
N("pos50 dqan-aa-dqan",5) = 1;
if (N("noun",5) && N("verb",6))
{
if (plural(N(5)) && vconjq(N(6),"inf"))
{
fixnphead(5);
alphatoverb(6,"active","VBD");
}
}
@RULES
_xNIL <-
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM)]
@@
# verb dqan alpha [NIBBLE]
# verb adj noun alpha [NIBBLE]
@CHECK
if (!N("noun",4))
fail();
if (!singular(N(3)) && !N("mass",3))
fail();
@POST
alphatonoun(4);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_adj
_noun
_xALPHA
@@
# verb dqa alpha alpha [NIBBLE]
@PRE
<1,1> varne("stem","do");
<3,3> varz("prosubj");
<3,3> varz("proobj");
@CHECK
if (!N("adj",6) && !N("noun",6))
fail();
if (N("adv",6) && !N(4) && !N(5))
fail();
if (N("verb",7) && N("noun",6) && !N("adj",6))
{
# Could be noun verb case.
if (vconjq(N(7),"-ing"))
fail();
}
if (!N("noun",7))
fail();
if (N(3))
L("first") = eltnode(3);
else if (N(4))
L("first") = eltnode(4);
else if (N(5))
L("first") = eltnode(5);
else
L("first") = N(6);
if (!numbersagree(L("first"),N(7)))
fail();
@POST
fixnphead(7);
fixnpnonhead(6);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA
_xWILD [one match=(_xEND _prep _modal _verb _vg
_fnword _conj _clausesep)]
@@
# verb dqan alpha
# NIBBLE.
@PRE
<7,7> varne("sem","date");
<7,7> varz("adv");
@CHECK
if (!N("noun",7))
fail();
if (!singular(N(7)) || !singular(N(6)))
fail();
if (N("mass",6))
fail();
if (N("verb",7) && copulaq(N(1)))
fail();
if (N("verb",7))
{
if (vconjq(N(7),"inf") && verbfeat(N(1),"V2"))
fail(); # let someone have ...
}
@POST
alphatonoun(7);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun
_xALPHA # 7
@@
# verb dqan alpha alpha
@CHECK
if (!N("noun",4) || !N("noun",5))
fail();
@POST
if (N(2))
{
# Fix up any adjs in here...
L("lastj") = lasteltnode(2);
fixadjs(N(2),L("lastj"));
}
L("tmp4") = N(4);
L("tmp5") = N(5);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
fixnoun(N(5));
group(2,5,"_np");
pncopyvars(L("tmp5"),N(5));
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro _quan _num _xNUM _adj)]
_noun [plus]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _conj _fnword _verb _vg
_qEOS _xEND \, _clausesep )]
@@
# dqan 's alpha noun verb
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
@POST
chpos(N(2),"POS");
L("tmp3") = N(3);
if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
else
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
}
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj _noun)]
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_noun [lookahead]
_xWILD [star match=(_advl _adv)]
# Trying broader rule. #
# _xWILD [one match=(_verb _vg)]
@@
# noun 's alpha verb
@CHECK
if (!N("noun",3) && !N("unknown",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# noun 's alpha
@CHECK
if (!N("noun",3) && !N("unknown",3))
fail();
if (N("adj",3))
fail(); # 05/08/07 AM.
if (N("verb",3))
{
if (!vconjq(N(3),"inf") && !vconjq(N(3),"-s"))
fail();
}
@POST
if (pnname(N(4)) == "_noun")
fixnpnonhead(3);
else
{
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
}
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xWILD [one lookahead fail=(_adj _xALPHA _dbldash \, )]
@@
# 's alpha prep verb
# 's alpha to verb
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
L("stem") = nodestem(N(2));
if (!finddictattr(L("stem"),"T3")) # v to v
fail();
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
group(1,1,"_be");
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
_aposS
_xALPHA
to [s lookahead]
_xWILD [one match=(_verb _vg)]
@@
# 's alpha prep
@CHECK
if (N("noun",2))
fail(); # 08/24/06 AM.
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
group(1,1,"_be");
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
_aposS
_xALPHA
_prep [lookahead]
@@
# noun 's alpha dqan verb
@CHECK
if (!N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_adj [plus lookahead]
_noun [plus]
_xWILD [one match=(_modal _verb _vg)]
@@
# noun 's
@POST
chpos(N(2),"POS");
L("tmp1") = N(1);
group(1,2,"_adj");
pncopyvars(L("tmp1"),N(1));
N("possessive",1) = 1;
N("ne",1) = N("ne type",1) = 0;
clearpos(N(1),1,0);
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xWILD [star match=(_adv _advl)]
_xWILD [one lookahead match=(_np _adj _noun
_num _quan _xNUM)]
@@
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xWILD [s one match=(most least more less)]
_xWILD [one lookahead match=(_np _noun)]
@@
# noun 's
@CHECK
if (N("mypos",2) != "POS")
fail();
@POST
group(1,2,"_adj");
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg _det _pro)]
@@
# noun 's
@CHECK
if (N("mypos",2) != "POS")
fail();
@POST
group(1,2,"_adj");
N("apos-s",1) = 1;
N("possessive",1) = 1;
@RULES
_xNIL <-
_noun
_xWILD [one match=(_aposS _aposX)]
@@
# 's not alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"-en")
&& !vconjq(N(3),"-ing"))
fail();
@POST
L("tmp1") = N(1);
pnrename(N(1),"_be");
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"VBZ");
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_aposS
_neg [s]
_xALPHA
@@
# alpha noun
@CHECK
if (N("pos num",1) != 2)
fail();
if (!N("noun",1) || !N("verb",1))
fail();
S("number") = number(N(2));
if (!vconjq(N(1),"inf"))
fail();
if (S("number") != "singular")
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
fixnoun(N(1));
@RULES
_xNIL <-
_xALPHA
_noun
_xWILD [one lookahead match=(_verb _vg)]
@@
# verb noun to alpha
# vg np to alpha
# verb patterns...
@CHECK
if (!N("verb",7))
fail();
if (!vconjq(N(7),"inf"))
fail();
L("stem") = nodestem(N(1));
if (!finddictattr(L("stem"),"V3"))
fail();
@POST
alphatovg(7,"active","VB");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_pro _np _noun)]
_xWILD [star match=(_adv _advl)]
to [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
@PRE
<1,1> varz("pos50 v-n-v t5 try");
<3,3> varne("sem","date"); # Not some kind of adverbial np.
@CHECK
if (!verbfeat(N(1),"T5"))
fail();
@POST
N("pos50 v-n-v t5 try",1) = 1;
N("glom",3) = "right";
if (pnname(N(5)) == "_verb")
{
verbtovg(5,5,"active","VBP");
}
else if (!vgassigned(N(5)))
{
fixvg(N(5),"active","VBP");
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_pro _np _noun)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# verb noun verb
# vg np vg
# verb patterns...
@PRE
<5,5> varz("pos50 vg-np-vg");
@CHECK
if (vgassigned(N(5)))
fail();
if (!vconjq(N(5),"inf"))
fail();
if (!verbfeat(N(1),"V2"))
fail();
@POST
N("pos50 vg-np-vg",5) = 1;
if (pnname(N(5)) == "_verb")
verbtovg(5,5,"active","VB");
else
fixvg(N(5),"active","VB");
# L("n") = pndown(N(5));
# "output.txt" << pnvar(L("n"),"mypos") << "\n";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_pro _np _noun)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# verb noun alpha
@CHECK
if (N("pos50 vna",3))
fail();
@POST
N("pos50 vna",3) = 1;
if (N("noun",3))
fixnphead(3);
else if (N("adv",3))
alphatoadv(3);
else if (N("adj",3))
alphatoadj(3);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np _pro)]
_xALPHA
_xWILD [lookahead one match=( \, _clausesep _xEND )]
@@
# List looks...
# noun , alpha , and alpha
@CHECK
if (N("n-alpha-alpha",6))
fail(); # Loop guard.
if (!N("noun",4) || !N("noun",7))
fail();
@POST
if (literal(N(4)))
{
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
}
if (literal(N(7)))
{
L("tmp7") = N(7);
group(7,7,"_noun");
pncopyvars(L("tmp7"),N(7));
fixnoun(N(7));
}
N("n-alpha-alpha",6) = 1;
@RULES
_xNIL <-
_xWILD [one fail=( \, _conj )]
_noun
\,
_xWILD [one match=(_xALPHA _noun)]
\, [opt]
_conj
_xWILD [one match=(_xALPHA _noun)]
_xWILD [one lookahead match=(_xEND _qEOS _prep _advl _adv
_verb _vg)]
@@
# prep np conj np to
@PRE
<1,1> varne("stem","by");
<5,5> vareq("stem","to");
@POST
group(2,4,"_np"); # np list
@RULES
_xNIL <-
_prep
_np
_conj
_np
_prep [lookahead]
@@
# vg np conj np
# Note: Semantic compatibility...
@CHECK
if (N("pos50 n-conj-n",2))
fail();
@POST
N("pos50 n-conj-n", 2) = 1;
if (N("sem",2) == "date"
&& N("sem",5) == "date")
group(2,5,"_np");
@RULES
_xNIL <-
_vg
_np
_conj
_adv [star]
_np
_xWILD [lookahead one match=(_prep)]
@@
# Some list looks....
# noun , alpha , alpha , and alpha
# verb , alpha
# adj , alpha
# adv , alpha
# Todo: complete this, add more list cases, etc.
@CHECK
if (N("pos50-and4",1)) # Loop guard.
fail();
S("nm") = pnname(N(1));
if (S("nm") == "_noun")
{
if (!N("noun",3) || !N("noun",5))
fail();
if (!N("noun",8) && !N("adj",8))
fail();
}
@POST
N("pos50-and4",1) = 1; # Loop guard.
if (S("nm") == "_noun" || S("nm") == "_np")
{
if (literal(N(3)))
{
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
}
if (literal(N(5)))
{
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
fixnoun(N(5));
}
if (literal(N(8)))
{
if (N("noun",8))
{
L("tmp8") = N(8);
group(8,8,"_noun");
pncopyvars(L("tmp8"),N(8));
fixnoun(N(8));
}
else if (N("adj",8))
{
L("tmp8") = N(8);
group(8,8,"_adj");
pncopyvars(L("tmp8"),N(8));
fixadj(N(8));
}
}
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _verb _adj _adv)]
\,
_xANY
\,
_xANY
\, [opt]
_conj
_xANY
@@
# noun conj
@CHECK
if (!N("noun",5))
fail();
if (!N("noun",4) && !N("adj",4))
fail();
@POST
L("tmp4") = N(4);
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
fixnoun(N(5));
if (N("adj",4))
{
group(4,4,"_adj");
pncopyvars(L("tmp4"),N(4));
fixadj(N(4));
}
else
{
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
}
@RULES
_xNIL <-
_xWILD [one match=(_prep _fnword)]
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_fnword _prep _conj _dbldash
_qEOS _xEND)]
@@
# noun conj alpha
@CHECK
if (!N("noun",4))
fail();
if (N("verb",4))
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
@RULES
_xNIL <-
_xWILD [one match=(_prep _fnword _det _vg _verb)]
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_xWILD [one lookahead match=(_fnword _prep _conj _dbldash
_qEOS _xEND)]
@@
# adj prep noun conj alpha
# Note: parallel construction.
@PRE
<5,5> var("adj");
@POST
alphatoadj(5);
@RULES
_xNIL <-
_xWILD [one match=(_adj _adjc)]
_prep
_xWILD [one match=(_np _noun)]
_conj
_xALPHA
_prep [lookahead]
@@
# prep noun conj alpha prep
@PRE
<4,4> var("noun");
@CHECK
if (N("verb",4))
{
if (phrprepverbq(N(4),N(5)))
fail();
}
# nn agreement checks.
@POST
fixnphead(4);
@RULES
_xNIL <-
_prep
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_prep [lookahead]
@@
# it 's alpha
@CHECK
if (N("stem",1) != "it")
fail();
if (!N("noun",3))
fail();
if (N("verb",3))
{
S("vc") = vconj(N(3));
if (S("vc") == "-ing"
|| S("vc") == "-en"
|| S("vc") == "-edn")
fail(); # Todo: Could reduce to _vg here.
}
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
nountonp(3,1);
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
clearpos(N(2),1,0);
N("verb node",2) = L("v");
fixvg(N(2),"active",0);
chpos(L("v"),"VBZ");
nountonp(1,1);
@RULES
_xNIL <-
_pro
_aposS
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# it 's dqan
@CHECK
if (N("stem",1) != "it")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
clearpos(N(2),1,0);
N("verb node",2) = L("v");
fixvg(N(2),"active",0);
chpos(L("v"),"VBZ");
nountonp(1,1);
@RULES
_xNIL <-
_pro
_aposS
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# verb prep noun
@PRE
<2,2> varz("mypos");
<3,3> vareq("sem","date");
@CHECK
if (!prepositionalverbq(N(1),N(2)))
{
fail();
}
@POST
if (N("stem",2) != "to")
N("mypos",2) = "IN"; # phr/prep/IN.
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep [lookahead]
_noun
@@
# verb prep prep
@PRE
<2,2> varz("mypos");
@CHECK
if (pnname(N(1)) == "_vg")
{
L("v") = N("verb node",1);
if (N("voice",1) != "active" && vconjq(L("v"),"-en"))
fail();
}
else
{
if (!N("mypos",1) || N("mypos",1) == "VBN")
fail();
}
@POST
if (N("stem",2) != "to")
chpos(N(2),"IN"); # prep/IN
N("pos50 vpp",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep [lookahead]
_prep
@@
# verb prep
@CHECK
if (N("mypos",2))
fail();
if (N("pos50-particle",2))
fail();
@POST
N("pos50-particle",2) = 1;
# See if kb has something. #
if (L("num") = phrprepverbq(N(1),N(2)))
{
if (L("num") == 2) # prepositional. # FIX. 06/18/06 AM.
chpos(N(2),"IN");
else # 1 or 3.
chpos(N(2),"RP"); # Default.
}
else if (copulaq(N(1)))
{
chpos(N(2),"IN"); # eg, UP % # 06/15/06 AM.
}
# else # Default.
# chpos(N(2),"RP");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep [lookahead]
_xWILD [one match=( _advl \, _dbldash _xEND)]
@@
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [s one lookahead match=(up down out)]
@@
# clausal
# verb noun prep
# up
# Check for phrasal verb.
@CHECK
if (N("pos50-vnprep",5))
fail();
if (N("mypos",5))
fail();
@POST
N("pos50-vnprep",5) = 1;
# Leaving stocks up.
# Forcing stocks up.
# verb pattern = v n adv
# Require phrasal verb otherwise.
# Prepositional is not valid because of intervening noun.
S("num") = phrprepverbq(N(1),N(5));
if (S("num") == 1 # PHRASAL.
|| S("num") == 3)
chpos(N(5),"RP");
# else
# chpos(N(5),"RB"); # Not prepositional verb.
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_prep
_xWILD [one match=(_prep _advl \, _dbldash _xEND)]
@@
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(up down out off)]
@@
# alpha prep
# Note: Need active/passive trans/intr logic.
@PRE
<1,1> var("verb");
@CHECK
S("num") = phrprepverbq(N(1),N(2));
if (!S("num"))
fail();
@POST
if (vconjq(N(1),"inf"))
L("pos") = "VBP";
alphatoverb(1,0,L("pos"));
# if (S("num") == 1 # PHRASAL.
# || S("num") == 3)
# chpos(N(2),"RP");
# else
# chpos(N(2),"IN");
@RULES
_xNIL <-
_xALPHA
_prep [lookahead]
@@
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"RB");
@RULES
_xNIL <-
_xWILD [one match=(_prep \, )]
_xWILD [s one match=(up down)]
_xWILD [one lookahead match=(_num _quan _xNUM _money)]
@@
# Trying simple stupid for now...
# so adj
@CHECK
if (N("mypos",1) == "RB")
fail(); # Loop guard.
@POST
pnrename(N(1),"_adv");
chpos(N(1),"RB"); # so/RB
@RULES
_xNIL <-
so [s]
_adj [lookahead]
@@
# so be
_xNIL <-
so [s]
_be [s lookahead]
@@
# so pro
@PRE
<1,1> varz("ellipted-so-that"); # Loop guard.
<2,2> var("prosubj");
@POST
chpos(N(1),"IN"); # so/IN
N("ellipted-so-that",1) = 1;
@RULES
_xWILD <- #350
so [s]
_pro [lookahead]
@@
# det noun '
@POST
chpos(N(3),"POS"); # Possessive.
group(2,3,"_adj");
N("possessive",2) = 1;
@RULES
_xNIL <-
_xWILD [one fail=(_noun _xALPHA)]
_noun
_aposX
@@
# 's alpha dqan
@CHECK
if (!N("verb",3))
fail();
@POST
# verb to be??
chpos(N(1),"POS"); # 's/POS
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_aposS
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_det _pro)]
@@
# 's alpha
@CHECK
if (!N("noun",2))
fail();
if (N("adj",2))
fail();
if (N("verb",2))
{
if (!vconjq(N(2),"inf")
&& !vconjq(N(2),"-s"))
fail();
}
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xWILD [one lookahead fail=(_xALPHA _adj)]
@@
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_xEND
@@
# alpha 's noun
@POST
L("tmp1") = N(1);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
fixnoun(N(1));
if (strisupper(strpiece(N("$text",1),0,0)))
chpos(N(1),"NP");
chpos(N(2),"POS");
group(1,2,"_adj");
@RULES
_xNIL <-
_xALPHA
_xWILD [one match=(_aposS _aposX)]
_noun [lookahead]
@@
@CHECK
if (N("verb",2))
fail();
if (!N("adj",2))
fail();
@POST
if (!N("mypos",1))
chpos(N(1),"POS");
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_adj [lookahead]
@@
# num
# Lone number.
@POST
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep _conj)]
_xWILD [one match=(_quan _num _xNUM)]
_xWILD [one lookahead match=(_qEOS _prep _conj _xEND)]
@@
# alpha dqan
@CHECK
if (!N("noun",1))
fail();
if (N("verb",1)
|| N("adj",1)) # 02/10/05 AM.
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
fixnoun(N(1));
@RULES
_xNIL <-
_xALPHA
\, [opt]
_xWILD [one lookahead match=(_det _quan _num _xNUM _np)]
@@
# alpha dqan
@CHECK
if (!N("verb",1))
fail();
S("vc") = vconj(N(1));
if (S("vc") != "-edn"
&& S("vc") != "-ed"
&& S("vc") != "-en")
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _np)]
@@
# fnword alpha dqan
# before eating food
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
clearpos(N(2),1,0);
fixvg(N(2),"active","VBG");
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
_xALPHA
_xWILD [one lookahead match=(_np _pro _det
_quan _num _xNUM)]
@@
# adj conj alpha
@CHECK
if (!N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
listadd(1,3,"true");
@RULES
_xNIL <-
_adjc
_conj
_xALPHA
_xWILD [one match=(_pro _np _prep)]
@@
# adj conj alpha
@CHECK
if (!N("adj",5))
fail();
@POST
alphatoadj(5);
@RULES
_xNIL <-
_xWILD [one match=(_adj _adjc)]
_xWILD [opt match=( \, _dbldash)]
_conj
_adv [star]
_xALPHA
_xWILD [one lookahead match=( \, _dbldash _qEOS _xEND)]
@@
# verb dqan alpha
# pattern: v n adj
@CHECK
if (!N("adj",6))
fail();
S("stem") = nodestem(N(1));
if (!finddictattr(S("stem"),"X7")) # keep it green ...
fail();
S("head") = lasteltnode(5);
dqaninfo(2,3,4,5);
@POST
L("tmp6") = N(6);
group(6,6,"_adj");
pncopyvars(L("tmp6"),N(6));
fixadj(N(6));
L("tmp6") = N(6);
group(6,6,"_adjc");
pncopyvars(L("tmp6"),N(6));
clearpos(N(6),1,1);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_xEND _dbldash _adv _prep \, )]
@@
# verb dqan alpha
# pattern: v n n
@CHECK
if (!N("noun",6))
fail();
if (!plural(N(6)))
fail();
S("stem") = nodestem(N(1));
if (!finddictattr(S("stem"),"D1")) # give her flowers ...
fail();
S("head") = lasteltnode(5);
dqaninfo(2,3,4,5);
@POST
fixnphead(6);
nountonp(6,1);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_xEND _dbldash _adv _prep \, )]
@@
# verb dqan alpha
@CHECK
if (N("pos50 dqan-a",6))
fail();
if (N("stem",7) == "to")
fail();
if (copulaq(N(1)))
fail();
@POST
N("pos50 dqan-a",6) = 1;
L("stem") = nodestem(N(1));
if (N("verb",6))
{
if (finddictattr(L("stem"),"V4")
&& vconjq(N(6),"-ing"))
{
alphatoverb(6,"active","VBD");
L("done") = 1;
}
else if (finddictattr(L("stem"),"V8")
&& vconjq(N(6),"-en"))
{
alphatoverb(6,"passive","VBN");
L("done") = 1;
}
else if (finddictattr(L("stem"),"V2")
&& vconjq(N(6),"inf"))
{
alphatoverb(6,"active","VB");
L("done") = 1;
}
}
N("pos50 psv7",6) = 1;
if (!L("done") && N("noun",6))
fixnphead(6);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xALPHA [star lookahead]
_xWILD [one match=(_xEND _fnword
_dbldash _adv \, )]
@@
# vg dqa alpha , np ,
# Apposition.
@PRE
<5,5> var("noun");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail();
if (N("adj",5) && pnname(N(7)) == "_adj")
fail();
@POST
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_xALPHA ### (5)
\, [lookahead] ### (6)
_xWILD [one match=(_np _noun _det _quan _num _xNUM _adj)] ### (7)
@@
# verb dqa alpha alpha
@PRE
<5,5> var("noun");
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"-ing"))
fail();
if (!verbfeat(N(1),"V4"))
fail();
@POST
alphatovg(6,"active","VBG");
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA
@@
# vg dqa alpha prep
@PRE
<5,5> var("noun");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail();
if (N("verb",5))
{
if (!vconjq(N(5),"inf"))
fail();
}
@POST
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_xALPHA ### (5)
_xWILD [one match=(_prep)] ### (6)
@@
# verb prep dqan alpha
@PRE
<7,7> varz("cap");
<8,8> varne("sem","date");
<8,8> varz("adv");
@CHECK
if (!N("noun",8))
fail();
if (!N(4) && !N(5) && !N(6) && !N(7))
fail();
if (N("verb",8))
{
if (copulaq(N(1)))
fail();
if (N(7))
{
S("head") = lasteltnode(7);
if (plural(S("head")) || pnvar(S("head"),"mass"))
fail();
}
}
@POST
L("tmp8") = N(8);
groupone(N(8),8,"_noun",1);
# pncopyvars(L("tmp8"),N(8));
fixnoun(N(8));
dqaninfo(4,5,6,7);
S("olast") = 8;
S("last") = S("lastn") = S("lastan") = N(8);
if (!N(7))
S("firstn") = S("firstan") = N(8);
if (!N(4) && !N(5) && !N(6) && !N(7)) # Just in case.
S("first") = N(8);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_prep [opt]
_xWILD [s star match=(_det _proPoss _adv)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [star] # 7
_xALPHA # 8
\, [opt lookahead] # 04/18/05 AM.
_xWILD [one fail=(_xALPHA _noun _adj _aposS _aposX _prep)]
@@
# verb prep dqan alpha
@PRE
<8,8> varne("sem","date");
<8,8> varz("adv");
@CHECK
if (!N("noun",8))
fail();
if (N("verb",8))
{
if (copulaq(N(1)))
fail();
if (vconjq(N(8),"-ing"))
fail();
if (verbfeat(N(1),"V2") && vconjq(N(8),"inf"))
fail();
}
L("n") = lasteltnode(7);
if (plural(L("n")) || pnvar(L("n"),"mass"))
fail();
if (plural(N(8)))
fail();
@POST
# Agreement.
alphatonoun(8);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_prep [opt]
_xWILD [s star match=(_det _proPoss _adv)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus] # 7 # 04/18/07 AM.
_xALPHA # 8
@@
# more than vg by
@PRE
<1,1> var("morethan");
<3,3> varz("voice");
<4,4> vareq("stem","by");
@CHECK
if (!vconjq(N(3),"-en"))
fail();
@POST
fixvg(N(3),"passive","VBD");
N("pos50 psv8",3) = 1;
@RULES
_xNIL <-
_fnword
_xWILD [star match=(_adv _advl)]
_vg
_prep [lookahead]
@@
# verb by dqan
# vg by dqan
@CHECK
if (N("pos50-v-by",2))
fail();
@POST
if (pnname(N(2)) == "_verb")
fixverb(N(2),"passive","VBN");
else
fixvg(N(2),"passive","VBN");
N("pos50-v-by",2) = 2;
@RULES
_xNIL <-
_xWILD [one match=( \, _dbldash)]
_xWILD [one match=(_verb _vg)]
by [s lookahead]
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# verb dqan aposS
# prep dqan aposS
# verb dqan 's
# prep dqan 's
@CHECK
if (N("mypos",6))
fail();
@POST
chpos(N(6),"POS"); # Assumed.
L("tmp5") = lasteltnode(5);
group(5,6,"_adj");
pncopyvars(L("tmp5"),N(5));
N("ne",5) = 0;
clearpos(N(5),1,0);
N("possessive",5) = 1;
N("adj's",5) = 1;
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb _prep _fnword
# _conj \,
)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus gp=_nouns]
_xWILD [one match=(_aposS _aposX)]
@@
# prep num , alpha
@CHECK
if (N("pos num",4) != 2)
fail();
if (!N("verb",4) || !N("adj",4))
fail();
@POST
# Could be a date...
nountonp(2,1);
L("tmp4") = N(4);
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
L("v") = N(4);
group(4,4,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp4"),N(4));
N("verb node",4) = L("v");
fixvg(N(4),"active","VBD");
clearpos(N(4),1,0);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xNUM
\,
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _np)]
@@
# adv alpha
@CHECK
if (!N("adj",2))
fail();
if (N("verb",2)
|| N("adv",2)) # 02/19/05 AM.
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_adv
_xALPHA
@@
# alpha conj det
@CHECK
if (!N("noun",1))
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
_xALPHA
_conj [lookahead]
_xWILD [one match=(_det _noun _np)]
@@
# Verb list
# Should check agreement of verbs.
# Should compose semantics for the list. voice, pos, etc.
#@CHECK
# S("arr") = verblistagree(N(1),N(7));
# if (!S("arr"))
# fail();
@POST
group(2,7,"_vg");
N("list",2) = 1; # Treat as a list...
# Should collect the list into a var.
@RULES
_xNIL <-
_xWILD [one fail=(\, _verb )]
_xWILD [one match=(_verb _vg)]
\,
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_verb _vg \, )]
_conj
_xWILD [one match=(_verb _vg)]
@@
# Partial list
# , alpha noun conj
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
@POST
L("tmp2") = N(2);
if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
_xWILD [one match=( \, )]
_xALPHA
_noun
_conj
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# Partial list
# noun alpha , dqan conj noun
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj _noun)]
_xALPHA
\, [lookahead]
_xWILD [plus match=(_det _quan _num _xNUM _adj _noun _np)]
\, [opt]
_conj
_xWILD [plus match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# noun alpha noun
# noun last noun
# last month, this month etc.
# Could check N("sem",3) == "date" for corroboration.
@CHECK
if (N("mypos",2))
fail(); # Done.
@POST
L("tmp2") = N(2);
L("txt") = strtolower(N("$text",2));
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_noun)]
_xWILD [one match=(last next)] # "that", "this" ...
_noun [lookahead]
@@
# next noun
@POST
L("tmp1") = N(1);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
next
_xWILD [one lookahead match=(_noun _num _quan _xNUM _adj)]
@@
# verb conj alpha
@CHECK
if (!N("verb",5))
fail();
if (pnname(N(1)) == "_vg")
L("v") = N("verb node",1);
else
L("v") = N(1);
S("vc") = samevconj(L("v"),N(5));
if (!S("vc"))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_verb");
pncopyvars(L("tmp5"),N(5));
N("inf",5) = N("-s",5) = N("-ed",5) = N("-edn",5) =
N("-ing",5) = N("-en",5) = 0;
if (S("vc"))
{
L("p") = vconj(N(1));
pnreplaceval(N(5),L("p"),1);
}
L("voice") = N("voice",1);
if (vconjq(N(1),"VBP"))
L("pos") = "VBP";
fixverb(N(5),L("voice"),L("pos"));
# Fix up the first verb also, if needed...
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
\, [opt]
_conj
_xALPHA
@@
# np , adv alpha
@CHECK
if (N("pos num",5) != 2)
fail();
if (!N("verb",5) || !N("adj",5))
fail();
if (!vconjq(N(5),"-en"))
fail();
@POST
alphatovg(5,"passive","VBN");
N("pos50 psv9",5) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS _prep _fnword)]
_xWILD [one match=(_noun _np)]
\,
_adv
_xALPHA
@@
# alpha about num
@CHECK
if (!N("verb",1))
fail();
@POST
L("tmp1") = N(1);
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
fixverb(N(1),"active",0);
@RULES
_xNIL <-
_xALPHA
_xWILD [s one lookahead match=(about circa approximately)]
_xWILD [one match=(_quan _num _xNUM)]
@@
_xNIL <-
_xALPHA
_xWILD [s one lookahead match=(about circa approximately)]
_xWILD [s one match=(_money _percent)]
@@
# verb yet to
@CHECK
if (N("mypos",2))
fail(); # Loop guard.
@POST
pnrename(N(2),"_adv");
chpos(N(2),"RB");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
yet [s]
_xWILD [s one lookahead match=(to \, _xEND _dbldash _conj)]
@@
# NOTE: Looks like a mixed up rule. #
# Rebuild from scratch. #
# alpha conj verb
# alpha conj vg
#@CHECK
# if (!N("verb",2))
# fail();
#
# # Check parallel construction.
# S("vc") = samevconj(N(2),N(4));
# if (S("vc"))
# succeed();
# if (N("pos num",2) != 2)
# fail();
# if (pnname(N(1)) != "_noun" || !N("noun",2))
# fail();
#@POST
# L("tmp2") = N(2);
# if (N("aspect",4) || N("tense",4) == "future") # mhb
# fixnphead(2);
# else if (S("vc"))
# {
# group(2,2,"_verb");
# pncopyvars(L("tmp2"),N(2));
# N("inf",2) = N("-s",2) = N("-ed",2) = N("-edn",2) =
# N("-ing",2) = N("-en",2) = 0;
# pnreplaceval(N(2),S("vc"),1);
# fixverb(N(2),0,0);
# # Fix up the 2nd verb also, if needed...
# }
# else
# {
# group(2,2,"_noun");
# pncopyvars(L("tmp2"),N(2));
# }
#
#@RULES
#_xNIL <-
# _xWILD [one match=(_noun _np _advl \,)]
# _xALPHA
# _xWILD [plus lookahead match=( \, _conj)]
# _xWILD [one match=(_verb _vg)]
# @@
# , alpha conj verb
# , alpha conj vg
@PRE
<2,2> var("verb");
@CHECK
if (!samevconj(N(2),N(4)))
fail();
@POST
L("voice") = N("voice",4);
L("v") = N("verb node",4);
if (L("v")) # 01/30/12 AM.
L("pos") = pnvar(L("v"),"mypos");
alphatovg(2,L("voice"),L("pos"));
@RULES
_xNIL <-
\,
_xALPHA
_conj [lookahead]
_xWILD [one match=(_verb _vg)]
@@
# pro verb conj verb
@CHECK
if (N("pos50 v-c-v",3))
fail();
@POST
N("pos50 v-c-v",3) = 1;
fixverb(N(3),"active","VBP");
fixverb(N(5),"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_pro)]
_xWILD [star match=(_advl _adv)]
_verb
_conj
_verb
@@
# alpha num
# SENTINEL. Don't want num to start noun phrase if alpha
# in front of it.
# eg, "first 3 people...."
@PRE
<1,1> varz("pos50 a-num");
@POST
N("pos50 a-num",1) = 1;
@RULES
_xNIL <-
_xALPHA
_xWILD [plus match=(_quan _num _xNUM)]
@@
# Need some context.
# dqan
@PRE
<1,1> varne("glom","left");
<5,5> varz("possessive");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail();
if (N(4))
{
L("last") = lasteltnode(4);
if (pnvar(L("last"),"advl") # 04/15/05 AM.
&& N(4) != L("last"))
fail();
}
dqaninfo(1,2,3,4);
if (!numbersagree(S("first"),S("last")))
fail(); # 05/29/07 AM.
@POST
groupnp();
@RULES
_xNIL <-
_det [plus]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [star]
_xWILD [one lookahead fail=(_xALPHA _conj \, _aposS _aposX
_dbldash _dblquote _fnword)]
@@
_xNIL <-
_det [star]
_xWILD [plus match=(_quan _num _xNUM)]
_adj [star]
_noun [star]
_xWILD [one lookahead fail=(_xALPHA _conj \, _aposS _aposX
_dbldash _dblquote _fnword)]
@@
_xNIL <-
_det [plus]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [star]
_xEND
@@
# dqan , alpha
@CHECK
if (N("noun",7))
fail();
S("head") = lasteltnode(5);
if (pnvar(S("head"),"sem") == "date"
&& (N(5) != S("head")))
fail();
@POST
if (N(4))
L("f") = N(4);
else
L("f") = N(5);
L("s") = pnvar(L("f"),"$ostart");
L("l") = lasteltnode(5);
L("e") = pnvar(L("l"),"$oend");
L("txt") = inputrange(L("s"),L("e"));
dqaninfo(2,3,4,5);
groupnp();
N("sem text",2) = L("txt");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg _prep _xSTART _qEOS)]
_xWILD [opt match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one match=( \, _conj _advl)]
_xWILD [one match=(_xALPHA _det _pro _quan _num _xNUM
_adj _verb _vg _np \, )]
@@
# dqan
@PRE
<2,2> varz("prosubj");
<2,2> varz("proobj");
<2,2> varne("glom","left");
@CHECK
S("head") = lasteltnode(5);
if (pnvar(S("head"),"advl") # 04/15/05 AM.
&& N(5) != S("head"))
fail();
if (pnvar(S("head"),"sem") == "date"
&& N(5) != S("head"))
fail();
@POST
dqaninfo(2,3,4,5);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg _prep _xSTART _qEOS
_advl \" _dblquote)]
_xWILD [opt match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead match=(_prep _verb _vg _qEOS _xEND
)]
@@
_xNIL <-
_xWILD [s one match=(either)]
_xWILD [opt match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_xALPHA)]
@@
# dqan
@PRE
<1,1> varz("prosubj");
<1,1> varz("proobj");
<1,1> varne("glom","left");
@CHECK
S("head") = lasteltnode(4);
if (pnvar(S("head"),"advl") # 04/15/05 AM.
&& N(4) != S("head"))
fail();
if (pnvar(S("head"),"sem") == "date"
&& N(4) != S("head"))
fail();
@POST
dqaninfo(1,2,3,4);
groupnp();
@RULES
_xNIL <-
_xWILD [plus match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM _adv)]
_adj [star]
_noun [plus]
_xWILD [one lookahead match=(_prep _verb _vg _qEOS _xEND
)]
@@
# Too weak. #
# alpha np
# Inflected...
#@CHECK
# if (!N("verb",1))
# fail();
# S("vc") = vconj(N(1));
# if (S("vc") == "inf"
# || S("vc") == "-s")
# fail();
#@POST
# L("tmp1") = N(1);
# group(1,1,"_verb");
# pncopyvars(L("tmp1"),N(1));
#@RULES
#_xNIL <-
# _xALPHA
# _xWILD [one lookahead match=(_np)]
# @@
# verb that noun verb
# vg that np vg
# Note: Also look for subjunctive.
# decree that John eat his peas.
@PRE
<1,1> varz("pos50 verb-that");
@POST
if (!N("mypos",3))
chpos(N(3),"IN"); # that/IN.
if (pnname(N(1)) == "_vg")
fixvg(N(1),"active","VBD");
else
fixverb(N(1),"active","VBD");
N("pos50 verb-that",1) = 1;
# subjunctive...
if (!vgassigned(N(7)))
{
L("t") = nodestem(N(1));
if (finddictattr(L("t"),"that-subjunctive")
&& vconjq(N(7),"inf"))
{
if (pnname(N(7)) == "_vg")
fixvg(N(7),"active","VB"); # Subjunctive.
else
fixverb(N(7),"active","VB"); # Subjunctive.
N("that-subjunctive",1) = 1;
N("that-subjunctive",3) = 1;
N("that-subjunctive",7) = 1;
}
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
that [s except=(_det _np)]
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg)]
@@
# that noun verb
@PRE
<2,2> varz("pos50 that-nv");
<4,4> var("verb");
@POST
N("pos50 that-nv",2) = 1; # Loop guard.
# Need a default tense for sentence, text, etc...
if (vconjq(N(4),"-ed"))
{
if (pnname(N(4)) == "_vg")
fixvg(N(4),"active","VBD");
else
fixverb(N(4),"active","VBD");
}
else if (vconjq(N(4),"inf"))
{
# demand that John do good
# think that John does good
if (singular(N(3)))
{
if (pnname(N(4)) == "_vg")
fixvg(N(4),"active","VB");
else
fixverb(N(4),"active","VB");
}
# If plural, VB/VBP ambiguous!
}
else
{
if (pnname(N(4)) == "_vg")
fixvg(N(4),0,0);
else
fixverb(N(4),0,0);
}
@RULES
_xNIL <-
_xWILD [one fail=(_fnword)]
that [s]
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
@@
@CHECK
if (N("mass",3) || plural(N(3)))
fail();
if (N("ne",3) || N("ne arr",3))
fail();
if (N("mypos",2))
fail();
@POST
pnrename(N(2),"_det"); # that -> det
chpos(N(2),"DT"); # that/DT
@RULES
_xNIL <-
_xWILD [one fail=(_np _noun _adv _advl _xALPHA)]
that [s except=(_det _np)]
_xWILD [one match=(_noun)]
_xWILD [one lookahead match=(_verb _vg)]
@@
_xNIL <- #400
_xSTART
that [s except=(_det)]
_noun
_xWILD [one lookahead match=(_verb _vg)]
@@
# Checking for agreement to enhance this.
# prep that noun
# that
@CHECK
if (N("mypos",2))
fail();
S("head") = lasteltnode(4);
if (pnvar(S("head"),"mass"))
fail(); # Don't know.
@POST
if (plural(S("head")))
{
chpos(N(2),"IN"); # that/IN.
dqaninfo(0,0,3,4);
groupnp();
}
else
{
chpos(N(2),"DT"); # that/DT
pnrename(N(2),"_det");
}
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
that [s]
_adj [star lookahead]
_noun [plus]
_xWILD [one fail=(_xALPHA _adj)]
@@
# dqan alpha that dqan
# that
@CHECK
if (!N("verb",4))
fail();
if (vconjq(N(4),"-en"))
fail(); # Could be adj also.
@POST
L("tmp4") = N(4);
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
L("v") = N(4);
group(4,4,"_vg");
mhbv(N(4),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp4"),N(4));
N("verb node",4) = L("v");
clearpos(N(4),1,0);
fixvg(N(4),0,0);
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj _noun _np)]
_xWILD [star match=(_advl _adv)]
\, [opt]
_xALPHA
that [s lookahead]
_xWILD [one match=(_det _quan _num _xNUM _adj
_noun _np _prep)]
@@
# dqan alpha that verb
# that
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_det _quan _num _xNUM _adj _noun)]
_xALPHA
that [s lookahead]
_xWILD [one match=(_verb _vg)]
@@
# alpha that verb
@CHECK
if (N("pos50 a-that-v",1))
fail();
@POST
N("pos50 a-that-v",1) = 1;
if (N("noun",1))
fixnphead(1);
@RULES
_xNIL <-
_xALPHA
that [s lookahead]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg)]
@@
# alpha that
# Note: try simplest first.
@CHECK
if (!N("verb",1))
fail();
if (!verbfeat(N(1),"T5"))
fail();
@POST
alphatoverb(1,"active",0);
@RULES
_xNIL <-
_xALPHA
that [s lookahead]
@@
# dqan outside noun
@POST
pnrename(N(2),"_adj");
@RULES
_xNIL <-
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
outside [s]
_noun [lookahead]
@@
# verb worth
@POST
L("tmp2") = N(2);
group(2,2,"_prep");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
worth
_xWILD [one lookahead match=(_qEOS _xEND \, )]
@@
# prep dqan value
@CHECK
if (!N("money",3))
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
fixnoun(N(4));
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_det
_noun
value
@@
# noun noun alpha
@CHECK
if (!N("cap",2) && N("mypos",2) != "NP")
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
nountonp(2,1);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
L("v") = N(3);
group(3,3,"_vg");
mhbv(N(3),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp3"),N(3));
N("verb node",3) = L("v");
clearpos(N(3),1,0);
fixvg(N(3),"active",0);
@RULES
_xNIL <-
_xWILD [one match=(_noun)]
_noun
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# adj alpha alpha dqan
@CHECK
if (!N("noun",2))
fail();
if (!N("verb",4))
fail();
if (N("noun",4))
{
if (plural(N(2)) && plural(N(4)))
succeed();
fail();
}
@POST
L("tmp2") = N(2);
L("tmp4") = N(4);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
L("v") = N(4);
L("neg") = mhbvadv(3,0,0,0);
group(4,4,"_vg");
mhbv(N(4),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp4"),N(4));
N("verb node",4) = L("v");
clearpos(N(4),1,0);
fixvg(N(4),"active","VBP");
@RULES
_xNIL <-
_adj
_xALPHA
_xWILD [star match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _noun)]
@@
# verb alpha dqan
@CHECK
if (!N("adv",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
_xWILD [one lookahead match=(_np _det _pro _quan _num _xNUM)]
@@
# modal alpha
@CHECK
if (!N("verb",3))
fail();
@POST
alphatoverb(3,"active","VB");
@RULES
_xNIL <-
_modal
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# det noun of noun
# dqan noun of noun
@POST
L("number") = numbersagree(N(1),N(2));
L("tmp2") = N(2);
L("tmp4") = N(4);
nountonp(4,1);
group(1,2,"_np");
pncopyvars(L("tmp2"),N(1));
clearpos(N(1),1,1);
group(1,3,"_np");
pncopyvars(L("tmp4"),N(1));
clearpos(N(1),1,0);
if (L("number"))
N("number",1) = L("number");
@RULES
_xNIL <-
_det
_noun
of [s]
_noun
_xWILD [one lookahead match=(_fnword)]
@@
# det num alpha
@CHECK
if (N("noun",3))
fail();
if (!N("verb",3))
fail();
# Check n-v number agreement etc.
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
# chpos(N(2),"NN"); # Conform Treebank.
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
group(1,2,"_np");
pncopyvars(L("tmp2"),N(1));
clearpos(N(1),1,0);
@RULES
_xNIL <-
_xWILD [star match=(_det)]
_num
_xALPHA
_xWILD [one lookahead match=(_prep _np _fnword)]
@@
# dqa alpha
# num alpha [NIBBLE]
@PRE
<1,1> varz("prosubj");
<1,1> varz("proobj");
@CHECK
if (N("pos50 num-a",3))
fail();
@POST
N("pos50 num-a",3) = 1;
if (N("noun",3))
{
if (plural(N(3)))
{
fixnphead(3);
L("done") = 1;
}
}
if (!L("done") && N("verb",3) && N(2))
{
if (!vconjq(N(3),"-s"))
alphatoverb(3,"active","VBD");
}
@RULES
_xNIL <-
_xWILD [plus match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_xALPHA
_xWILD [one fail=(_adj _noun _xALPHA)]
@@
# dqa alpha
# num alpha [NIBBLE]
@PRE
<1,1> varne("stem","one");
@CHECK
if (N("pos50 num-a",2))
fail();
if (!numbersagree(N(1),N(2)))
fail();
@POST
N("pos50 num-a",2) = 1;
if (N("noun",2))
{
if (plural(N(2)))
{
fixnphead(2);
L("done") = 1;
}
else if (singular(N(2))
&& (N("num",1) == 1 || N("number",1) == "singular") )
{
fixnphead(2);
L("done") = 1;
}
}
if (!L("done") && N("verb",2))
{
if (!vconjq(N(2),"-s"))
alphatoverb(2,"active","VBD");
}
@RULES
_xNIL <-
_xWILD [plus match=(_quan _num _xNUM)]
_xALPHA
_xWILD [one fail=(_adj _noun _xALPHA)]
@@
# Moose said
# alpha vg
# alpha verb
@CHECK
if (!N("noun",2))
fail();
L("sem") = nodesem(N(4));
if (!finddictattr(L("sem"),"agent-human"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
N("cap",2) = 1;
fixnoun(N(2));
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xCAP
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_verb _vg)]
@@
# adj and adj
@POST
group(2,4,"_adj");
N("adjs",2) = 1;
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [plus match=(_det _pro _quan _num _xNUM _advl)]
_adj
_conj
_adj
_xWILD [one lookahead match=(_noun _verb _vg _advl _adv)]
@@
########################
# Some mhbv into here...
########################
@CHECK
if (!N("verb",9))
fail();
# Agreement...
L("arr") = vgagree(N(1),N(3),N(5),N(7),N(9));
if (!L("arr")[0])
fail();
@POST
L("tmp9") = N(9);
if (pnname(N(9)) != "_verb")
{
group(9,9,"_verb");
pncopyvars(L("tmp9"),N(9));
}
L("m") = N(1);
L("h") = N(3);
L("b") = N(5);
L("being") = N(7);
L("v") = N(9);
L("neg") = mhbvadv(2,4,6,8);
if (N(7))
{
N("sem",7) = N("stem",7) = "be";
chpos(N(7),"VBG");
}
group(1,9,"_vg");
pncopyvars(L("v"),N(1));
clearpos(N(1),1,0);
mhbv(N(1),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
@RULES
_xNIL <-
_modal [s opt]
_xWILD [star match=(_adv _advl)]
_have [s]
_xWILD [star match=(_adv _advl)]
_be [s opt]
_xWILD [star match=(_adv _advl)]
being [s opt]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_xALPHA _verb)]
@@
# past
#@CHECK
# if (pnname(N(2)) == "_adj")
# fail(); # Infinite loop guard.
@POST
# if (literal(N(2)))
# {
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
# }
# else
# pnrename(N(2),"_adj");
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
past
_xWILD [one lookahead match=(_quan _num _xNUM _adj _noun)]
@@
# ^ prep dqan alpha noun ,
@POST
if (N("verb",7) && vconjq(N(7),"-ing"))
alphatovg(7,"active","VBG");
else if (N("noun",7) || N("unknown",7) || N("cap",7))
alphatonoun(7);
else if (N("adj",7))
alphatoadj(7);
else if (N("adv",7))
alphatoadv(7);
else # [DEFAULT
alphatonoun(7);
@RULES
_xNIL <-
_xSTART ### (1)
_prep ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_xALPHA ### (7)
_noun [lookahead] ### (8)
\, ### (9)
@@
# ^ prep dqan alpha
# ^ prep noun alpha
@CHECK
if (N("pos num",6) != 2)
fail();
if (!N("noun",6) || !N("verb",6))
fail();
S("noun") = lasteltnode(5);
if (!singular(S("noun")))
fail();
S("vc6") = vconj(N(6));
if (S("vc6") == "-ing"
|| S("vc6") == "-edn"
|| S("vc6") == "-en")
fail();
@POST
L("tmp6") = N(6);
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
fixnoun(N(6));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_prep
_det [star]
_adj [star]
_noun [plus]
_xALPHA
@@
# ^ prep dqa alpha
@CHECK
if (!N("noun",4) || !N("verb",4))
fail();
S("noun") = lasteltnode(3);
if (!singular(S("noun")))
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
# fixnoun(N(4));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_prep
_xWILD [plus match=(_det _quan _num _xNUM _adj)]
_xALPHA
@@
# noun alpha
# noun date
@CHECK
if (!N("advl",2)) # 04/15/05 AM.
fail();
if (!N("noun",2) && !N("adv",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
# chpos(N(2),"NN"); # Conform treebank.
chpos(N(2),"RB");
@RULES
_xNIL <-
_noun
_xALPHA
@@
# noun subject to
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
subject
to [s lookahead]
@@
# plus alpha
@CHECK
if (!N("noun",2))
fail();
if (!singular(N(2)))
fail();
@POST
# Note mass-noun here...
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
plus [s]
_xALPHA
_xWILD [one lookahead match=(_xEND _qEOS \, )]
@@
# plus
# By and large a CC in the corpus, if no other evidence.
@POST
pncopyvars(1);
S("mypos") = "CC";
single();
@RULES
_conj <-
plus
@@
# ^ alpha noun
# ^ alpha dqan
@CHECK
if (!N("adj",2))
fail();
if (!singular(N(2)))
fail();
if (N("mass",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xALPHA
_noun [lookahead]
@@
# noun alpha adj
@CHECK
if (N("adj",2))
fail();
if (!N("verb",2))
fail();
if (copulaq(N(2)))
succeed();
if (vconjq(N(2),"inf"))
fail(); # Want inflected.
@POST
alphatoverb(2,0,0);
@RULES
_xNIL <-
_noun [plus]
_xALPHA
_adv [star]
_adj [lookahead]
_xWILD [one match=(_adj _noun _fnword)]
@@
# verb adj conj verb noun
# Conflicted ving + ving parallel construction.
@CHECK
L("txt2") = nodetext(N(2));
L("txt4") = nodetext(N(4));
if (strendswith(L("txt2"),"ing")
&& strendswith(L("txt4"),"ing") )
succeed();
fail();
@POST
pnrename(N(4),"_adj"); # verb -> adj
fixadj(N(4));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_adj
_conj
_verb
_xWILD [one lookahead match=(_noun _adj)]
@@
# adj conj adj
@POST
L("tmp5") = N(5);
group(2,5,"_adj");
pncopyvars(L("tmp5"),N(2));
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg _noun _np)]
_xWILD [one match=(_adj _adjc)]
_conj
_adv [star]
_adj
_xWILD [one lookahead match=(_adj _noun _prep _xEND)]
@@
# adjc conj adj
@POST
group(1,3,"_adjc"); # Adj list...
@RULES
_xNIL <-
_adjc
_conj
_adj
@@
# dqa alpha , alpha alpha $
@PRE
<3,3> var("adj");
<5,5> var("adj");
<6,6> var("noun");
@CHECK
if (!N(1) && !N(2))
fail();
@POST
fixnphead(6);
fixnpnonhead(5);
fixnpnonhead(3);
group(3,5,"_adj"); # List.
@RULES
_xNIL <-
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_xALPHA
\,
_xALPHA
_xALPHA
_xEND
@@
# Some alphas centered around a comma.
# dqa alpha , alpha
@CHECK
if (!N("noun",2))
fail();
if (!N("verb",4))
fail();
S("vc4") = vconj(N(4));
if (S("vc4") != "-edn"
&& S("vc4") != "-en")
fail();
@POST
L("tmp2") = N(2);
L("tmp4") = N(4);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
group(4,4,"_verb");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=(_det _pro _quan _num _xNUM _adj)]
_xALPHA
\,
_xALPHA
_xWILD [one lookahead match=(_prep _advl _adv)]
@@
# alpha alpha prep
# alpha alpha adv
@PRE
<4,4> varne("stem","of");
@CHECK
if (!N("noun",1))
fail();
if (!N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "-edn"
&& S("vc") != "-en"
&& S("vc") != "-ed")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xALPHA
_xALPHA
_xWILD [star match=(_advl _adv)]
_xWILD [one lookahead match=(_prep)]
@@
# quasi-question forms.
# modal np have be being verb
# Not only will John eat....
# ...
@CHECK
if (!N(5) && !N(7) && !N(9) && !N(11))
fail();
if (N(11))
{
if (!N("verb",11))
fail();
}
L("arr") = vgagree(N(1),N(5),N(7),N(9),N(11));
if (!L("arr")[0])
fail();
@POST
if (N(11))
{
if (pnname(N(11)) != "_verb")
alphatoverb(11,0,0);
}
L("m") = N(1);
L("h") = N(5);
L("b") = N(7);
L("being") = N(9);
L("v") = N(11);
if (N(9))
{
N("sem",9) = N("stem",9) = "be";
chpos(N(9),"VBG");
}
L("neg") = mhbvadv(2,6,8,10);
group(5,11,"_vg");
mhbv(N(5),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
N("q-q remote",5) = "left";
if (pnname(N(1)) != "_vg")
{
if (N("stem",1) == "do")
fixverb(N(1),"active","VBP");
group(1,1,"_vg");
N("q-q remote",1) = "right";
N("modal",1) = 1;
}
@RULES
_xNIL <-
_xWILD [s one match=(_modal _do)]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_noun _np _pro)]
_xWILD [star match=(_advl _adv)]
_have [s opt]
_xWILD [star match=(_advl _adv)]
_be [s opt]
_xWILD [star match=(_advl _adv)]
being [s opt]
_xWILD [star match=(_advl _adv)]
_xWILD [opt match=(_verb _xALPHA)]
@@
@CHECK
if (pnname(N(1)) == "_adv" && !N(2))
fail();
# Agreement...
L("arr") = vgagree(N(2),N(4),N(6),N(8),N(10));
if (!L("arr")[0])
fail();
@POST
L("neg") = mhbvadv(3,5,7,9);
L("m") = N(2);
L("h") = N(4);
L("b") = N(6);
L("being") = N(8);
L("v") = N(10);
L("glom") = N("glom",10); # ..and other semantics.
# Should be a more compact way to do the below...
# At least group could return the grouped node.
L("neg") = mhbvadv(3,5,7,9);
if (N(8))
{
N("sem",8) = N("stem",8) = "be";
chpos(N(8),"VBG");
}
if (N(2))
{
group(2,10,"_vg");
mhbv(N(2),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
if (L("glom"))
N("glom",2) = L("glom");
}
else if (N(4))
{
group(4,10,"_vg");
mhbv(N(4),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
if (L("glom"))
N("glom",4) = L("glom");
}
else if (N(6))
{
group(6,10,"_vg");
mhbv(N(6),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
if (L("glom"))
N("glom",6) = L("glom");
}
else if (N(8))
{
group(8,10,"_vg");
mhbv(N(8),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
if (L("glom"))
N("glom",8) = L("glom");
}
else
{
group(10,10,"_vg");
mhbv(N(10),L("neg"),L("m"),L("h"),L("b"),L("being"),L("v"));
if (L("glom"))
N("glom",10) = L("glom");
}
@RULES
_xNIL <-
_xWILD [one fail=(_verb _modal _xALPHA)]
_xWILD [s opt match=(_modal _do)]
_xWILD [star match=(_adv _advl)]
_have [s opt]
_xWILD [star match=(_adv _advl)]
_be [s opt]
_xWILD [star match=(_adv _advl)]
being [s opt]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(_verb) except=(_modal _have _be _vg)]
@@
_xNIL <-
_xWILD [one fail=(_verb _modal _xALPHA)]
_xWILD [s opt match=(_modal _do)]
_xWILD [star match=(_adv _advl)]
_have [s]
_xWILD [star match=(_adv _advl)]
_be [s opt]
_xWILD [star match=(_adv _advl)]
being [s opt]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(_verb) except=(_be _vg)]
@@
# quasi-question forms.
# have np be being verb
# Not only has John eaten....
# Not only has John been eating...
@CHECK
if (!N(5) && !N(7) && !N(9))
fail();
L("arr") = vgagree(0,N(1),N(5),N(7),N(9));
if (!L("arr")[0])
fail();
@POST
L("h") = N(1);
L("b") = N(5);
L("being") = N(7);
L("v") = N(9);
if (N(7))
{
N("sem",7) = N("stem",7) = "be";
chpos(N(7),"VBG");
}
L("neg") = mhbvadv(2,4,6,8);
group(5,9,"_vg");
mhbv(N(5),L("neg"),0,L("h"),L("b"),L("being"),L("v"));
N("q-q remote",5) = "left";
if (pnname(N(1)) == "_verb")
{
group(1,1,"_vg");
N("q-q remote",1) = "right";
N("have",1) = 1;
}
@RULES
_xNIL <-
_have [s]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_noun _np _pro)]
_xWILD [star match=(_advl _adv)]
_be [s opt]
_xWILD [star match=(_advl _adv)]
being [s opt]
_xWILD [star match=(_advl _adv)]
_verb [opt]
@@
# Trap to avoid quasi-question forms.
@PRE
<1,1> varz("pos50 there-be-n");
@POST
N("pos50 there-be-n",1) = 1;
N("glom",1) = "right";
N("glom",7) = "left"; # elts are renumbered 1 2 3=vg 4=advl 5=np
if (N("stem",1) == "there")
{
L("n") = pndown(N(1));
chpos(L("n"),"EX"); # there/EX
}
L("m") = N(3);
L("h") = N(4);
L("b") = N(5);
if (pnname(N(5)) != "_vg")
{
L("neg") = mhbvadv(2,6,0,0);
group(3,5,"_vg");
mhbv(N(3),L("neg"),L("m"),L("h"),L("b"),0,0);
}
@RULES
_xNIL <-
_xWILD [s one match=(there _pro)]
_xWILD [star match=(_advl _adv)]
_modal [opt]
_have [opt]
_be [s]
_xWILD [star lookahead match=(_advl _adv)]
_xWILD [one match=(_noun _np _adj _adjc
_det _quan _num _xNUM _noun)]
@@
# question/query form.
# whword be noun verb
# whword be np mhbv
@CHECK
if (!N(7) && !N(9))
fail();
L("arr") = vgagree(0,0,N(3),N(7),N(9));
if (!L("arr")[0])
fail();
if (N("q-q remote",3))
fail();
@POST
L("b") = N(3);
L("being") = N(7);
L("v") = N(9);
if (N(7))
{
N("sem",7) = N("stem",7) = "be";
chpos(N(7),"VBG");
}
L("neg") = mhbvadv(3,5,7,9);
group(7,9,"_vg");
mhbv(N(7),L("neg"),0,0,L("b"),L("being"),L("v"));
N("q-q remote",7) = "left";
if (pnname(N(3)) == "_verb")
{
group(3,3,"_vg");
N("be",3) = 1;
}
N("q-q remote",3) = "right";
@RULES
_xNIL <-
_whword [s]
_xWILD [star match=(_advl _adv)]
_be [s]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_noun _np _pro)]
_xWILD [star match=(_advl _adv)]
being [s opt]
_xWILD [star match=(_advl _adv)]
_verb [opt]
@@
# quasi-question forms.
# be np being verb
# Not only is John eating...
@PRE
<5,5> varz("pos50: be-being"); # 09/24/13 AM.
<7,7> varz("pos50: be-being"); # 09/24/13 AM.
@CHECK
if (!N(5) && !N(7))
fail();
L("arr") = vgagree(0,0,N(1),N(5),N(7));
if (!L("arr")[0])
fail();
@POST
L("b") = N(1);
L("being") = N(5);
L("v") = N(7);
if (N(5))
{
N("sem",5) = N("stem",5) = "be";
chpos(N(5),"VBG");
}
L("neg") = mhbvadv(2,4,6,0);
group(5,7,"_vg");
++N("pos50: be-being",5); # 09/24/13 AM.
mhbv(N(5),L("neg"),0,0,L("b"),L("being"),L("v"));
N("q-q remote",5) = "left";
if (pnname(N(1)) == "_verb")
{
group(1,1,"_vg");
N("q-q remote",1) = "right";
N("be",1) = 1;
}
@RULES
_xNIL <-
_be [s]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_noun _np _pro)]
_xWILD [star match=(_advl _adv)]
being [s opt]
_xWILD [star match=(_advl _adv)]
_verb [opt]
@@
# Not only will John have been eating.
# so
# Unless we know something, tag it as adverb.
@CHECK
if (N("mypos"))
fail();
@POST
chpos(N(1),"RB"); # so/RB
@RULES
_xNIL <-
so [s]
@@
# adj -- conj adj --
@POST
group(1,5,"_adj");
@RULES
_xNIL <-
_adj
_dbldash
_conj
_adj
_dbldash
@@
# prep alpha alpha prep
# Assuming all cases of this handled here....
@CHECK
if (N("paap",1))
fail(); # Loop guard.
@POST
N("paap",1) = 1;
L("tmp2") = N(2);
L("tmp3") = N(3);
L("txt2") = N("$text",2);
# First alpha first...
if (finddictattr(L("txt2"),"prefer-verb"))
{
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
N("paap",1) = 1;
}
@RULES
_xNIL <-
_prep
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# clausal form
# vg np prep alpha alpha
@CHECK
if (!N("noun",6))
fail();
if (N("verb",6))
{
S("vc6") = vconj(N(6));
if (S("vc6") != "inf"
&& S("vc6") != "-s")
fail();
}
@POST
L("tmp6") = N(6);
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_np _noun)]
_prep
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_fnword _prep)]
@@
# clausal fragment
# verb noun prep alpha noun verb
@CHECK
if (N("stem",3) == "to")
fail();
if (!N("noun",4))
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
_prep
_xALPHA
_xWILD [one lookahead match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
@@
# np np vg
@CHECK
if (N("pos50 n-n-v",3))
fail(); # Loop guard.
if (N("eos",1))
fail();
@POST
if (!N("voice",3) || N("voice",3) == "active")
fixvg(N(3),"active","VBP");
N("pos50 n-n-v",3) = 1;
@RULES
_xNIL <-
_np
_np
_vg
@@
# do verb
@CHECK
if (!vconjq(N(3),"inf"))
fail();
@POST
L("tmp1") = N(1);
L("tmp3") = N(3);
L("neg") = mhbvadv(2,0,0,0);
group(1,3,"_vg");
mhbv(N(1),L("neg"),L("tmp1"),0,0,0,L("tmp3"));
pncopyvars(L("tmp3"),N(1));
clearpos(N(1),1,0);
N("voice",1) = "active";
N("verb node",1) = L("tmp3");
N("first verb",1) = L("tmp1");
fixvg(N(1),"active","VBP");
fixverb(L("tmp3"),"active","VB");
@RULES
_xNIL <-
_do [s]
_xWILD [star match=(_adv _advl)]
_verb
@@
# do alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"inf"))
fail();
@POST
L("tmp1") = N(1);
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
L("v") = N(3);
L("neg") = mhbvadv(2,0,0,0);
group(1,3,"_vg");
mhbv(N(1),L("neg"),L("tmp1"),0,0,0,L("v"));
pncopyvars(L("tmp3"),N(1));
clearpos(N(1),1,0);
N("voice",1) = "active";
N("verb node",1) = L("v");
N("first verb",1) = L("tmp1");
fixvg(N(1),"active","VBP");
fixverb(L("v"),"active","VB");
@RULES
_xNIL <- #450
_do [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# beginning
@CHECK
if (!N("verb",1) || !N("noun",1))
fail();
if (N("adv",1))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_prep");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"VBG"); # Conformity.
@RULES
_xNIL <-
_xALPHA
_xWILD [one match=(beginning)]
@@
# clausal pattern.
# verb np prep alpha
@CHECK
if (N("stem",5) == "to")
fail();
if (!N("noun",6))
fail();
@POST
L("tmp6") = N(6);
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_advl _adv)]
_np
_xWILD [star match=(_advl _adv)]
_prep
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# noun either verb
@CHECK
if (N("mypos",2))
fail(); # Loop guard etc.
@POST
if (!N("mypos",2))
chpos(N(2),"RB"); # Conformity.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro)]
_xWILD [s one match=(either neither)]
_xWILD [one lookahead match=(_verb _vg)]
@@
# clause dqan alpha dqan
@CHECK
if (N(2))
{
if (N("pattern",2) != "advl")
fail();
}
if (!N("verb",7))
fail();
@POST
L("tmp7") = N(7);
group(7,7,"_verb");
pncopyvars(L("tmp7"),N(7));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_clause [opt]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_det _pro _quan _num _xNUM
_adj)]
@@
# verb alpha prep
# verb forward prep
# Certain adverbs....
@CHECK
if (nonliteral(N(2)))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [s one match=(forward forwards
backward backwards ahead)]
@@
# noun cap
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
N("cap",2) = 1;
if (N("eos",1))
{
chpos(N(2),"NP");
nountonp(2,1);
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xCAP
_xWILD [one lookahead match=(_prep)]
@@
# noun alpha pro
@CHECK
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xALPHA
_pro [lookahead]
@@
# verb prep np verb
@PRE
<1,1> varne("stem","be");
<1,1> varne("stem","have");
<1,1> varne("stem","do");
@POST
group(2,3,"_advl");
N("pp",2) = 1;
@RULES
_xNIL <-
_verb
_prep
_np
_verb [lookahead]
@@
# verb
@POST
L("tmp2") = N(2);
group(2,2,"_vg");
pncopyvars(L("tmp2"),N(2));
clearpos(N(2),1,0);
N("verb node",2) = N("first verb",2) = L("tmp2");
mhbv(N(2),L("neg"),0,0,0,0,L("tmp2"));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro _prep)]
_verb
_xWILD [one lookahead match=(_noun _np _pro
# _prep # Might lead to advl. #
)]
@@
# verb prep noun verb
@CHECK
if (N("voice",1) != "active")
fail();
if (N("voice",4))
fail();
@POST
fixvg(N(4),"passive","VBN");
N("pos50 psv10",4) = 1;
@RULES
_xNIL <-
_vg
_prep
_np
_vg
_xWILD [one fail=(_det _pro _quan _num _xNUM _adj _noun _np _fnword)]
@@
# adj alpha.
@CHECK
if (!N("possessive",1))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("verb",2))
fail(); # Loop guard.
S("vc2") = vconj(N(2));
if (S("vc2") != "inf"
&& S("vc2") != "-s"
&& S("vc2") != "-ed")
fail();
@POST
N("verb",2) = 0;
L("tmp2") = N(2);
L("num") = N("pos num",2);
pnreplaceval(N(2),"pos num",--L("num"));
alphaunambigred(2);
@RULES
_xNIL <-
_adj
_xALPHA
@@
# verb adj alpha alpha
@CHECK
if (!N("adj",3) && !N("noun",3))
fail();
if (!N("adj",4) && !N("noun",4))
fail();
if (N("adv",4))
fail();
if (N("verb",3))
fail(); # For now.
if (N("verb",4))
{
if (!vconjq(N(4),"inf"))
fail();
}
@POST
fixnpnonhead(3);
if (N("verb",4))
{
N("verb",4) = 0;
L("num") = N("pos num",4);
pnreplaceval(N(4),"pos num",--L("num"));
if (L("num") == 1) # Unambiguous.
{
L("tmp4") = N(4);
# Figure out pos and do reduction.
if (N("adj",4))
group(4,4,"_adj");
else if (N("noun",4))
group(4,4,"_noun");
else if (N("adv",4))
group(4,4,"_adv");
pncopyvars(L("tmp4"),N(4));
if (pnname(N(4)) == "_adj")
fixadj(N(4));
}
}
@RULES
_xNIL <-
_verb
_adj
_xALPHA
_xALPHA
@@
# alpha adv alpha
@CHECK
if (N("verb",1))
fail();
if (!N("noun",1))
fail();
if (!N("verb",3))
fail();
@POST
L("tmp1") = N(1);
L("tmp3") = N(3);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xALPHA
_adv
_xALPHA
_xWILD [one match=(_prep _fnword)]
@@
# noun , alpha
@CHECK
if (!N("noun",5))
fail();
if (N("verb",5))
{
if (!vconjq(N(5),"inf") && !N("inf",5))
fail();
}
if (!N("cap",3) && N("mypos",3) != "NP")
fail();
if (!N("glom comma",3))
fail();
@POST
excise(4,4);
@RULES
_xNIL <-
_xWILD [one match=(_prep _det _pro)]
_xWILD [star match=(_quan _num _xNUM _adj)]
_noun
\,
_xALPHA [lookahead]
@@
# noun , alpha conj adj
@CHECK
if (!N("noun",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xALPHA
\, [opt lookahead]
_conj
_xWILD [one match=(_adj _noun _np
_adv _advl # 04/21/07 AM.
)]
@@
# , noun verb alpha dqan
@CHECK
if (!N("adj",4) || !N("verb",4))
fail();
if (N("noun",4))
fail(); # Could handle this here also...
S("vc4") = vconj(N(4));
if (S("vc4") != "-edn"
&& S("vc4") != "-en")
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_adj");
pncopyvars(L("tmp4"),N(4));
fixadj(N(4));
@RULES
_xNIL <-
_xWILD [one match=(_qEOS _xSTART \, )]
_xWILD [one match=(_noun _np)]
_xWILD [s one match=(_verb _vg) except=(_modal _have _be
being)]
_xALPHA
_xWILD [one lookahead match=(_adj _noun _xALPHA)]
@@
# noun alpha noun verb
# dqan alpha noun verb
@CHECK
if (!N("noun",6))
fail();
if (N(2))
S("start") = N(2);
else if (N(3))
S("start") = N(3);
else if (N(4))
S("start") = N(4);
else
S("start") = N(5);
if (!numbersagree(S("start"),N(6)))
fail();
@POST
fixnpnonhead(6);
@RULES
_xNIL <-
_xWILD [one match=(_qEOS _xSTART \" _dblquote
_fnword # Trying. # 06/13/06 AM.
)]
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_noun [plus lookahead]
_xWILD [one match=(_verb _vg)]
@@
# noun conj alpha conj
@CHECK
if (!N("noun",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
\, [opt]
_xWILD [one lookahead match=(_conj)]
@@
# prep alpha conj vg
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
L("v") = N(2);
pncopyvars(L("tmp2"),N(2));
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
N("first verb",2) = N("verb node",2) = L("v");
pncopyvars(L("tmp2"),N(2));
fixvg(N(2),"active","VBG");
clearpos(N(2),1,0);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_xALPHA
_conj [lookahead]
_xWILD [one match=(_verb _vg)]
@@
# num or num
@POST
L("tmp4") = N(4);
group(2,4,"_num");
N("list",2) = 1;
pncopyvars(L("tmp4"),N(2));
@RULES
_xNIL <-
_xWILD [one fail=( \, )]
_num
or [s]
_num
@@
# dqan
# det cap conj cap alpha
@CHECK
if (!N("cap",2) || !N("cap",4))
fail();
if (N("cap",5))
fail();
if (!N("noun",5))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
if (N("date",2) || N("date",4))
L("date") = 1;
group(2,4,"_noun");
if (L("date"))
N("date",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_det)]
_noun
_conj
_noun
_xALPHA
@@
# dqa alpha
# NIBBLE.
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
if (N("pos50 adj-a-a",4))
fail();
@POST
N("pos50 adj-a-a",4) = 1;
if (N("verb",4))
{
--N("pos num",4);
N("verb",4) = 0;
}
if (N("pos num",4) == 1)
alphaunambigred(4);
# else if (N("verb",5))
# fixnphead(4); # MAYBE.
else if (!N("verb",5))
fixnpnonhead(4); # adj-noun thingy.
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
_adv [star]
_adj [plus]
_xALPHA
_xALPHA
@@
# Too broad. Too old.
# dqan alpha
# dqan cap alpha
@CHECK
if (!N("cap",4))
fail();
if (N("cap",5))
fail();
if (!N("noun",5))
fail();
if (N("verb",5))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
@RULES
_xNIL <-
_xWILD [one match=(_det)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun
_xALPHA
@@
# dqa alpha prep
@CHECK
if (N("verb",2))
fail();
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_adj [plus]
_xALPHA
_xWILD [one lookahead match=(_prep _conj _xEND)]
@@
# ^ dqan alpha prep
@CHECK
if (!N("noun",6))
fail();
# Agreement...
if (vconjq(N(6),"-ing"))
fail();
L("last") = lasteltnode(5);
if (singular(L("last")) && singular(N(6)))
succeed();
fail();
@POST
fixnphead(6);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_prep \, )]
@@
# prep noun
# prep cap
@CHECK
if (!N("cap",2))
fail();
if (N("date",2))
fail();
@POST
N("date",2) = 1;
N("sem",2) = "date";
N("eventive",2) = 1; # Some kind of historic event...
if (pnname(N(1)) == "_fnword")
pnrename(N(1),"_prep");
@RULES
_xNIL <-
_xWILD [s one match=(since until till after before during)]
_noun
_xWILD [one lookahead match=( _qEOS _xPUNCT )]
@@
# Ellipted temporal.
# days before
# noun noun prep
@CHECK
if (N("cap",1))
fail();
if (N("sem",2) != "date")
fail();
@POST
nountonp(2,1);
N("sem",2) = "date";
if (pnname(N(3)) == "_fnword")
pnrename(N(3),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_noun
_xWILD [s one lookahead match=(before after)]
@@
# noun alpha prep as
@CHECK
if (!N("verb",2))
fail();
# Agreement.
if (plural(N(1)) && singular(N(2)))
succeed();
if (plural(N(2)) && singular(N(1)))
succeed();
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xALPHA
_prep [lookahead]
_fnword
@@
# np of dqan alpha prep
@CHECK
if (!N("verb",7))
fail();
if (N("noun",7)) # 09/09/05 AM.
fail();
@POST
N("t") = nodetext(N(7));
if (phrprepverbq(N(7),N(8)))
alphatoverb(7,0,"VBP");
else if (finddictattr(N("t"),"trans"))
alphatoverb(7,"passive","VBN");
else
alphatoverb(7,0,"VBP");
N("pos50 psv11",7) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_np _noun)]
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_xNUM _num _quan)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# verb alpha verb
@CHECK
if (!N("adv",2))
fail();
S("stem1") = N("stem",1);
if (S("stem1") != "be"
&& S("stem1") != "being"
&& S("stem1") != "have")
fail();
if (!singular(N(2)))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_verb
_xALPHA
_verb
@@
# noun noun
# Some things that don't glom.
@CHECK
if (N("sem",2) != "date")
fail();
if (N("stem",2) == "date") # Literal word 'date'.
fail();
@POST
L("year") = N("year",2); # 02/13/14 AM.
nountonp(2,1);
group(2,2,"_advl"); # Date adverbial... # np -> advl
if (L("year")) N("year",2) = L("year");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _verb _vg)]
_xWILD [one match=(_noun _np)]
@@
# np (date)
@PRE
<2,2> vareq("sem","date");
@POST
group(2,2,"_advl");
@RULES
_xNIL <-
_xWILD [one match=(_adj _adjc _noun _np)]
_np
@@
# prep conj prep
@POST
if (N("list",2))
listadd(2,4,"true");
else
{
group(2,4,"_prep");
clearpos(N(2),1,0);
N("list",2) = 1;
}
@RULES
_xNIL <-
_xWILD [one fail=( \, )]
_prep
_conj
_prep
@@
# ^ adj alpha noun
# ^ dqa alpha noun
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
@POST
fixnpnonhead(3);
@RULES
_xNIL <-
_xSTART
_adj [plus]
_xALPHA
_noun [lookahead]
@@
# adj alpha noun
# dqa alpha noun
@CHECK
if (!N("adj",2) || !N("noun",2))
fail();
if (N("adv",2) || N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
_adj
_xALPHA
_noun [lookahead]
@@
# adj " noun "
# Period and quote should be inverted ahead of time...
@POST
if (N(5))
{
group(6,6,"_qEOS"); # Convert to an end of sentence!
group(2,5,"_np");
}
else
group(2,6,"_np");
clearpos(N(2),1,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_adj
_xWILD [one match=( \" _dblquote)]
_noun
_qEOS [opt]
_xWILD [one match=( \" _dblquote)]
@@
# vg to vg
# wants to eat
# Semantically simpler to deal with...
@PRE
<6,6> varz("pos50 v-to-v 2");
@CHECK
if (N("voice",1) != "active")
fail();
@POST
N("pos50 v-to-v 2",6) = 1;
L("tmp1") = N(1);
L("tmp5") = N(5);
L("v") = N("first verb",5);
if (!L("v"))
L("v") = N("verb node",5);
if (L("v") && !pnvar(L("v"),"mypos"))
{
chpos(L("v"),"VB"); # infinitive.
pnreplaceval(L("v"),"inf",1);
}
# If ambiguous...
fixvg(N(1),"active","VBD");
if (N(6))
{
if (pnname(N(6)) != "_prep")
vgvgcompound(1,5,"vg-to-vg");
}
@RULES
_xNIL <-
_vg
_xWILD [star match=(_advl _adv)]
to [s]
_xWILD [star match=(_advl _adv)]
_vg
_xWILD [one lookahead fail=(_verb _vg _modal)]
@@
# verb to vg
@POST
verbtovg(1,1,"active","VBD");
vgvgcompound(1,3,"vg-to-vg");
@RULES
_xNIL <-
_verb
to [s]
_vg
@@
# verb alpha prep
@CHECK
L("txt2") = strtolower(N("$text",2));
if (!finddictattr(L("txt2"),"prefer-adj"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# np prep noun alpha
@CHECK
if (!singular(N(2)))
fail();
if (!singular(N(5)))
fail();
if (!N("adj",5))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_adj");
pncopyvars(L("tmp5"),N(5));
fixadj(N(5));
@RULES
_xNIL <-
_xSTART
_np
for [s]
_noun
_xALPHA
_adv [opt lookahead]
to [s]
@@
# both alpha and noun
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
if (N("cap",2))
chpos(N(2),"NP");
else
chpos(N(2),"NN");
@RULES
_xNIL <-
both [s]
_xALPHA
_conj [lookahead]
_xWILD [one match=(_noun _np)]
@@
# both noun
@CHECK
if (!plural(N(2)))
fail();
if (N("mypos",1))
fail();
@POST
chpos(N(1),"DT");
pnrename(N(1),"_det"); # fnword -> det
@RULES
_xNIL <-
both [s]
_noun [lookahead]
@@
# verb noun alpha to
# Note: verb (that) noun verb
@CHECK
if (!N("verb",3))
fail();
if (!verbfeat(N(1),"T5"))
fail();
@POST
alphatoverb(3,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_noun
_xALPHA
to [s lookahead]
@@
# clausal.
# vg np , vg prep
# vg that np, vg prep
# verb noun , verb prep
# ex: said (that) the man, eaten with envy...
# note: form like: vg [that] np
# that
@CHECK
S("v5") = vgambigedn(N(5)); # 06/13/06 AM.
if (!S("v5")) # 06/13/06 AM.
fail();
@POST
pnreplaceval(S("v5"),"-en",1);
pnreplaceval(S("v5"),"-edn",0);
chpos(S("v5"),"VBN");
if (!N("voice",5))
N("voice",5) = "passive";
if (!N(2))
N("ellipted-that",1) = 1;
N("pos50 psv12",5) = 1;
@RULES
_xNIL <-
_vg
that [s opt lookahead]
_np
\,
_vg
_xWILD [one match=(_prep)]
@@
@PRE
<2,2> varne("glom","left");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail();
dqaninfo(2,3,4,5);
if (!numbersagree(S("first"),S("last")))
fail(); # 06/15/06 AM.
@POST
if (N(3))
if (!nonliteral(N(3)))
{
L("tmp3") = N(3);
group(3,3,"_num");
pncopyvars(L("tmp3"),N(3));
if (S("first") == L("tmp3")) # 04/23/07 AM.
S("first") = N(3);
}
if (N(4))
L("f") = N(4);
else
L("f") = N(5);
L("s") = pnvar(L("f"),"$ostart");
L("l") = lasteltnode(5);
L("e") = pnvar(L("l"),"$oend");
L("txt") = inputrange(L("s"),L("e"));
groupnp();
N("sem text",2) = L("txt");
@RULES
_xNIL <-
_xALPHA
_xWILD [s plus match=(_det _pro) except=(_proObj)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS _aposX _conj
_dblquote )]
@@
_xNIL <-
_xALPHA
_xWILD [s star match=(_det _pro) except=(_proObj)]
_xWILD [plus match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS _aposX _conj
_dblquote)]
@@
_xNIL <-
_xWILD [one fail=(_xALPHA _det _quan _num _xNUM
_adj _noun _aposS _aposX \, )]
_xWILD [s star match=(_det _proPoss)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS _aposX _conj
_dblquote)]
@@
# ^ dqan
@POST
dqaninfo(2,3,4,5);
groupnp();
@RULES
_xNIL <-
_xSTART
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead match=(_dbldash _prep _fnword
_adv _advl _modal _verb _vg)]
@@
# her
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"PP$");
@RULES
_xNIL <-
her [s]
_xWILD [one lookahead match=(_adj _noun)]
@@
# Try the broadest possible.
# verb alpha
@CHECK
if (!N("adj",3))
fail();
S("stem") = nodestem(N(3));
S("core") = finddictsattr(S("stem"),"core");
if (!copulaq(N(1)) && S("core") != "adj")
fail();
if (N("pos_np_head",3) == "NN")
fail(); # Wants to be a noun...
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# verb one 's
# prep one 's
# note: ignoring alternative for now.
# alternative eg: thinking (that) one's hot stuff.
@CHECK
if (N("mypos",2))
fail();
@POST
pnrename(N(2),"_noun");
chpos(N(2),"NN"); # one/NN (Conform Treebank. I prefer PP.)
chpos(N(3),"POS"); # 's/POS
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg _prep _fnword)]
one [s]
_aposS
@@
# adj np
# Glom a possessive adj at beginning of noun phrase.
# (Not if np starts with a det etc.
@CHECK
if (!N("possessive",1))
fail();
@POST
listadd(2,1,"false");
@RULES
_xNIL <-
_adj
_np
@@
# 's alpha dqan verb
@CHECK
if (!N("adj",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_aposS _aposX)]
_xALPHA
_adj [star lookahead]
_noun [plus]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg)]
@@
# A possible bad possessive. (s instead of 's).
# noun alpha prep
#@CHECK
# if (!N("noun",2))
# fail();
# if (!plural(N(1)) || !plural(N(2)))
# fail();
#@POST
# # First noun should really be adj possessive.
# L("tmp2") = N(2);
# group(2,2,"_noun");
# pncopyvars(L("tmp2"),N(2));
# # Flag that this is possessive adj etc.
# N("posessive",1) = 1;
#@RULES
#_xNIL <-
# _noun
# _xALPHA
# _xWILD [one match=(_prep)]
# @@
# semi-clausal pattern
# det alpha alpha prep np verb
# dqa alpha alpha prep np verb
@CHECK
if (!vconjq(N(6),"inf")
&& !vconjq(N(6),"-s"))
fail();
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("noun",3))
fail();
if (!singular(N(2)) || !singular(N(3)))
fail();
@POST
fixnpnonhead(2);
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_prep [lookahead]
_np
_xWILD [one match=(_modal _verb _vg)]
@@
# adj alpha alpha np verb
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3))
fail();
@POST
fixnpnonhead(2);
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
@RULES
_xNIL <-
_adj
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_np _pro)]
_xWILD [one match=(_modal _verb _vg)]
@@
# clausal look.
# vg np prep dqan alpha
# verb noun prep dqan alpha
@CHECK
if (!N("adj",10) && !N("noun",10) && !N("unknown",10))
fail();
if (N("verb",10))
if (!vconjq(N(10),"inf")
&& !vconjq(N(10),"-s"))
fail();
@POST
if (N("noun",10) || N("unknown",10))
fixnpnonhead(10);
else
alphatoadj(10);
@RULES
_xNIL <-
_vg ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_np ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_prep ### (5)
_xWILD [star match=(_adv _advl)] ### (6)
_xWILD [star match=(_det _pro)] ### (7)
_xWILD [star match=(_quan _num _xNUM)] ### (8)
_xWILD [plus match=(_adj _noun)] ### (9)
_xALPHA ### (10)
_xWILD [one lookahead match=(_adj _noun)] ### (11)
@@
# noun alpha adj $
@CHECK
if (!N("verb",2))
fail();
if (!copulaq(N(2)))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_noun
_xALPHA
_adj [lookahead]
_xEND
@@
# Locatives, adverbials.
# here
@POST
if (!N("mypos",1))
chpos(N(1),"RB");
L("tmp1") = N(1);
group(1,1,"_adv");
pncopyvars(L("tmp1"),N(1));
@RULES
_xNIL <-
_xWILD [one match=(here)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=( \, _xEND)]
@@
# vg that alpha prep
@PRE
<3,3> var("noun");
@CHECK
if (!plural(N(3)))
fail();
@POST
fixnphead(3);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
that [s]
_xALPHA
_prep [lookahead]
@@
# vg that alpha
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
if (N("verb",4))
{
if (!vconjq(N(4),"inf")
&& !vconjq(N(4),"-s"))
fail();
}
@POST
L("tmp4") = N(4);
if (N("adj",4))
group(4,4,"_adj");
else
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
that [s]
_xALPHA
_noun [lookahead]
@@
# clausal
# that
# Note: Looking for transitive verb pattern.
# Note: parallel construction can give wrong answer.
# ex: Animals eat that feed and benefit from it.
# "feed" could be a verb ... Nearly ambig syntactically,
# semantically.)
# verb that alpha
# vg that alpha
@PRE
<4,4> varz("adv");
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
if (N("verb",4))
{
if (vconjq(N(4),"inf")
|| vconjq(N(4),"-s"))
fail();
}
@POST
L("tmp4") = N(4);
if (N("noun",4) || N("unknown",4))
group(4,4,"_noun");
else
group(4,4,"_adj");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
that [s]
_xALPHA
@@
# clausal.
# that
# verb that dqan and verb
# vg that dqan and vg
# Todo: check parallel verb construction.
@CHECK
if (N("mypos",3))
fail();
@POST
chpos(N(3),"DT"); # that/DT.
pnrename(N(3),"_det");
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(that) except=(_det)]
_xWILD [star lookahead match=(_adv _advl)]
_adj [star]
_noun [plus]
_xWILD [star match=(_adv _advl)]
_conj
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_vg _verb)]
@@
# clausal.
# that
# verb that dqan
# Note: Looking at some verb patterns.
# said that water is bad. (ambig).
# worked that boy (unambig).
# said that boys are bad. (unambig)
@PRE
<1,1> varne("stem","be");
@CHECK
if (N("pos50 v-that-dqan",3))
fail();
@POST
N("pos50 v-that-dqan",3) = 1;
L("head") = lasteltnode(6);
L("txt") = nodestem(N(1));
L("T5") = finddictattr(L("txt"),"T5");
if (plural(L("head")))
chpos(N(3),"IN"); # that/IN.
else if (!L("T5"))
{
chpos(N(3),"DT"); # that/DT.
pnrename(N(3),"_det");
N("that-to-det",3) = 1;
}
# else ambiguous case.
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(that) except=(_det)]
_xWILD [star lookahead match=(_adv _advl)]
_adj [star]
_noun [plus]
_xWILD [one fail=(_xALPHA _aposS _aposX _verb _vg _modal)]
@@
# clausal.
# that
# verb that adj
# Note: may want to get rid of this, require singular head noun.
#@CHECK
# if (N("pos50 v-that-adj",3))
# fail();
# L("txt") = nodestem(N(1));
# L("T5") = finddictattr(L("txt"),"T5");
# if (L("T5"))
# fail();
# if (copulaq(N(1)))
# fail();
#@POST
# N("pos50 v-that-adj",3) = 1;
# chpos(N(3),"DT"); # that/DT.
# pnrename(N(3),"_det");
#@RULES
#_xNIL <-
# _xWILD [one match=(_vg _verb)]
# _xWILD [star match=(_adv _advl)]
# _xWILD [s one match=(that) except=(_det)]
# _xWILD [star lookahead match=(_adv _advl)]
# _adj
# @@
# semi clausal.
# verb , prep dqan , alpha
@CHECK
if (!N("verb",7))
fail();
if (!vconjq(N(7),"inf"))
fail();
@POST
alphatoverb(7,"active","VB");
group(3,6,"_advl");
clearpos(N(3),1,0);
@RULES
_xNIL <-
_xWILD [s one match=(_modal _have _be _do)]
_adv [star]
\,
_prep
_xWILD [s plus match=(_det _pro _quan _num _xNUM _adj
_noun _np)]
\,
_xALPHA
@@
# , alpha noun verb
@PRE
<3,3> var("adj");
@POST
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _adv _advl)]
\,
_xALPHA
_noun [plus]
_xWILD [one match=(_modal _verb vg)]
@@
# clausal.
# , alpha noun
@CHECK
if (!N("verb",3))
fail();
if (vconjq(N(3),"inf"))
fail();
@POST
alphatoverb(3,0,0);
@RULES
_xNIL <-
\,
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_pro _np)]
@@
# verb noun like
# clausal.
@CHECK
if (N("mypos",3))
fail();
if (verbfeat(N(1),"V2"))
fail(); # Don't know which ...
@POST
chpos(N(3),"IN"); # like/IN.
pnrename(N(3),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
like [s]
@@
# ^ adv alpha
@CHECK
if (N("pos50 start-x-a",3))
fail();
@POST
N("pos50 start-x-a",3) = 1;
if (N("noun",3))
{
if (plural(N(3)))
fixnphead(3);
}
@RULES
_xNIL <-
_xSTART
_xWILD [plus match=(_adv _advl \, _clausesep)]
_xALPHA
@@
# np np like
# Ex: someone (that) I like ...
@CHECK
if (N("mypos",3))
fail();
if (pnname(N(3)) != "_fnword")
fail();
@POST
N("glom",1) = "left";
N("glom",2) = "right";
N("glom",3) = "left";
pnrename(N(3),"_verb");
fixverb(N(3),"active","VBP");
chpos(N(3),"VBP"); # like/VBP.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_np _pro)]
like [s]
@@
# ^ like dqan
@CHECK
if (N("mypos",2))
fail();
if (pnname(N(2)) != "_fnword")
fail();
@POST
pnrename(N(2),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART \, \( _dbldash)]
like [s]
_xWILD [one lookahead match=(_np _noun
_det _quan _adj _num _xNUM _prep _fnword
)]
@@
# verb noun noun like noun
# Ex: tell them something like ...
@CHECK
if (pnname(N(7)) != "_fnword")
fail();
@POST
pnrename(N(7),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun _np _pro)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
like [s]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_np _noun _det _quan _num _xNUM _pro
_adj \, )]
@@
# verb noun like noun
@CHECK
if (pnname(N(5)) != "_fnword")
fail();
@POST
pnrename(N(5),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [plus match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
like [s]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_np _noun _det _quan _num _xNUM _pro _adj)]
@@
# like
# clausal.
@CHECK
if (pnname(N(2)) != "_fnword")
fail();
@POST
pnrename(N(2),"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
like [s]
_xWILD [one lookahead match=(_noun _np _pro)]
_xWILD [one match=(_verb _modal _vg)]
@@
# clausal.
@PRE
<1,1> vareq("sem","have");
@CHECK
if (!N("noun",4))
fail();
if (!N("verb",5))
fail();
if (!vconjq(N(5),"-en"))
fail();
@POST
N("v-v-remote",1) = 1;
N("v-v-remote",5) = 1;
alphatovg(5,"active","VBN");
fixnphead(4);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_det
_xALPHA
_xALPHA
# No right context.
@@
# np alpha np
@CHECK
if (N("pos50 np-a-np",3))
fail();
@POST
N("pos50 np-a-np",3) = 1;
if (N("verb",3))
{
if (vconjq(N(3),"-ed"))
alphatoverb(3,"active","VBD");
else
alphatoverb(3,"active","VBP");
}
@RULES
_xNIL <-
_np
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star match=(_adv _advl)]
_np [lookahead]
@@
# verb noun alpha noun
# vg noun alpha noun
# clausal.
@CHECK
if (!N("noun",3))
fail();
if (N("verb",3))
{
if (!vconjq(N(3),"inf"))
fail();
if (verbfeat(N(1),"V2"))
fail();
}
if (!singular(N(2)) && !N("cap",2))
fail();
@POST
if (N("cap",4))
fixnphead(3);
else
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_noun
_xALPHA
_noun
@@
# dqan conj dqan
@POST
debug();
dqaninfo(7,8,9,10);
groupnp();
dqaninfo(2,3,4,5);
groupnp();
# group(2,4,"_np"); # List of nps.
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _clausesep _prep _verb _vg
_fnword)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_conj ### (6)
_xWILD [star match=(_det _pro)] ### (7)
_xWILD [star match=(_quan _num _xNUM)] ### (8)
_adj [star] ### (9)
_noun [plus] ### (10)
_xWILD [one lookahead match=(_xEND _clausesep _prep
_conj _fnword _vg _verb)] ### (11)
@@
# np conj dqan
@POST
dqaninfo(3,4,5,6);
groupnp();
@RULES
_xNIL <-
_np
_conj
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead match=(_xEND _clausesep _prep _conj
_fnword _vg _verb)]
@@
# last alpha verb
@CHECK
if (!N("noun",2))
fail();
@POST
fixnphead(2);
alphatoadj(1);
@RULES
_xNIL <-
last
_xALPHA # 09/24/13 AM.
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg \, _clausesep _xEND)]
@@
# last noun verb
@POST
alphatoadj(1);
@RULES
_xNIL <-
last
_noun [lookahead]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_modal _verb _vg \, _clausesep _xEND
_prep _fnword )]
@@
# last adj
_xNIL <-
last
_adj [lookahead]
@@
# last noun
# last date
@CHECK
if (N("sem",2) != "date")
fail();
@POST
alphatoadj(1);
@RULES
_xNIL <-
last
_noun [lookahead]
@@
# noun yet prep
# noun yet that
# Ex: the wildest show yet ...
# Note: Would like a superlative adj in noun phrase.
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"RB"); # yet/RB.
pnrename(N(2),"_adv");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
yet [s]
_xWILD [one lookahead match=(_prep _fnword _conj)]
@@
# det alpha alpha
# dqa alpha alpha
# NIBBLE.
@CHECK
if (N("verb",2) || N("verb",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
if (!N("noun",3) && !N("adj",3))
fail();
@POST
fixnpnonhead(2);
alphatoadj(3);
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
_xALPHA
_xALPHA
@@
# clausal.
# ^ noun verb dqan
# ^ np vg dqan
# NIBBLE.
@CHECK
if (!N(7) && !N(8) && !N(9))
fail();
if (!N("noun",10))
fail();
@POST
fixnphead(10);
L("nm") = pnname(N(5));
if (L("nm") == "_vg")
fixvg(N(5),"active","VBP");
else
fixverb(N(5),"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_xWILD [star match=(_det _pro)] # 7
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xWILD [one lookahead fail=(_noun _xALPHA)] # 04/20/07 AM.
@@
# np prep dqa alpha alpha
# noun prep dqa alpha alpha
# noun prep alpha alpha
# np prep alpha alpha
# NIBBLE
@PRE
<2,2> varne("stem","to");
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
if (!N("noun",5) && !N("adj",5))
fail();
if (!singular(N(4)))
fail();
if (N("verb",4) && !N(3) && vconjq(N(4),"-ing"))
fail();
if (N("verb",5) && !N(3) && vconjq(N(5),"-ing"))
fail();
@POST
if (vconjq(N(5),"-edn"))
fixnphead(4);
else
fixnpnonhead(4);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_prep
_adj [star]
_xALPHA
_xALPHA
@@
# np prep dqan alpha np
@CHECK
if (!N("verb",7))
fail();
if (plural(N(1)) && vconjq(N(7),"inf"))
succeed();
if (singular(N(1)) && vconjq(N(7),"-s"))
succeed();
fail();
@POST
alphatoverb(7,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM _np)]
@@
# prep dqan alpha np vg
@PRE
<6,6> var("noun");
@POST
fixnphead(6);
N("ellipted-that",7) = 1;
N("glom",7) = "right";
@RULES
_xNIL <-
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_np [lookahead]
_vg
@@
# ^ alpha alpha alpha
@PRE
<2,2> varz("pos50 ^aaa");
@POST
N("pos50 ^aaa",2) = 1; # Loop guard.
if (plural(N(3)) && plural(N(4)))
{
if (N("verb",3) && N("noun",4))
{
fixnphead(4);
alphatovg(3,"active","VBZ");
if (singular(N(2)) && !N("mass",2))
cappos(N(2),0);
else
fixnphead(2);
}
}
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_fnword _prep)]
@@
# ^ alpha alpha verb
@CHECK
if (!N("noun",3) && !N("adj",3) && !N("adv",3)
&& !N("unknown",3))
fail();
@POST
if (N("noun",3) || N("adj",3) || N("unknown",3))
fixnphead(3);
else
alphatoadv(3);
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_adv [opt lookahead]
_xWILD [one match=(_verb _vg)]
@@
_xNIL <-
\,
_xALPHA
_xALPHA
_adv [opt lookahead]
_xWILD [one match=(_verb _vg)]
@@
# ^ alpha prep
@PRE
<3,3> var("verb");
<4,4> varne("stem","by");
@POST
if (vconjq(N(3),"-en"))
alphatovg(3,"passive","VBN");
else if (vconjq(N(3),"inf"))
alphatovg(3,"active","VB");
else if (N("noun",3) && vconjq(N(3),"-ing"))
fixnphead(3);
else
alphatovg(3,"active","VB");
N("pos50 psv13",3) = 1;
@RULES
_xNIL <-
_xSTART
_xWILD [s star match=(although though while _adv)]
_xALPHA
_prep
_xWILD [one match=(_noun _np)]
_xWILD [one lookahead match=( \, _verb _vg )]
@@
# to verb noun , alpha alpha
# to vg np , alpha alpha
@CHECK
if (!N("verb",5))
fail();
if (!singular(N(5)) || !singular(N(5)))
fail();
@POST
alphatoverb(5,"active","VB");
if (N("verb",6))
{
N("verb",6) = 0;
--N("pos num",6);
alphaunambigred(6);
}
@RULES
_xNIL <-
to [s]
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_xALPHA
@@
# verb back
# vg back
# Note: Idiomatic.
@POST
alphatoadv(2);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
back
@@
# alpha adj prep
# NIBBLE
@CHECK
if (!N("noun",1) && !N("verb",1))
fail();
@POST
if (N("verb",1) && !N("noun",1))
alphatoverb(1,"active",0);
else if (N("noun",1) && !N("verb",1))
fixnphead(1);
else if (copulaq(N(1)))
alphatoverb(1,"active",0);
else
fixnphead(1);
group(2,2,"_adjc");
@RULES
_xNIL <-
_xALPHA
_adj
_xWILD [s one match=(_prep _xEND) except=(to)]
@@
# det alpha alpha adj
# dqa alpha alpha adj
# Note: post-adjective thingy.
@CHECK
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
@POST
fixnpnonhead(2);
fixnphead(3);
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
_xALPHA
_xALPHA
_adj
_xWILD [one lookahead match=(_prep)]
@@
# dqan alpha alpha verb
@CHECK
if (!N("noun",3))
fail();
if (!N("noun",2) && !N("adj",2))
fail();
@POST
fixnpnonhead(2);
fixnphead(3);
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# dqan alpha alpha
# noun alpha alpha
@PRE
<2,2> varz("pos50 n-a-a-0");
@POST
N("pos50 n-a-a-0",2) = 1;
if (plural(N(1)) && plural(N(3)) && N("verb",2))
{
if (vconjq(N(2),"-ed"))
alphatovg(2,"active","VBD");
else
alphatovg(2,"active","VBP");
}
else if (plural(N(1)) && N("verb",2))
alphatoverb(2,0,0);
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA [lookahead]
@@
# prep noun alpha , conj
@PRE
<3,3> var("noun");
@POST
fixnphead(3);
@RULES
_xNIL <-
_prep
_noun
_xALPHA
\, [lookahead]
_conj
@@
_xNIL <-
_prep
_noun
_xALPHA
_xWILD [one lookahead match=(_clausesep _xEND)]
@@
# fnword dqan alpha alpha
@CHECK
if (!N("verb",6))
fail();
if (!vconjq(N(6),"-s"))
fail();
L("n") = lasteltnode(5);
if (!pnvar(L("n"),"cap"))
fail();
@POST
alphatoverb(6,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_num _xNUM _quan)]
_adj [star]
_noun [plus]
_xALPHA
_xALPHA
@@
# , dqan alpha prep np ,
# Note: part of apposition.
@CHECK
if (!N("noun",7))
fail();
@POST
fixnphead(7);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_prep [lookahead]
_np
_xWILD [one match=( \, _conj _xEND)]
@@
# Apposition.
# np , np vg prep
@PRE
<1,1> var("cap");
@CHECK
if (N("voice",4))
fail();
if (N("det stem",3) != "a" && N("det stem",3) != "the")
fail();
if (!vconjq(N(4),"-en"))
fail();
@POST
fixvg(N(4),"passive","VBN");
N("pos50 psv14",4) = 1;
@RULES
_xNIL <-
_np
\,
_np
_vg
_xWILD [one lookahead match=(_prep)]
@@
# verb noun conj alpha alpha
# vg np conj alpha alpha
# verb dqan conj alpha alpha
# note: parallel construction.
@CHECK
if (N("verb",9) || !N("verb",8))
fail();
if (!samevconj(N(1),N(8)))
fail();
# Note: could add confidence with vscalar attribute.
@POST
if (pnname(N(1)) == "_vg")
{
L("voice") = N("voice",1);
L("v") = N("verb node",1);
L("pos") = pnvar(L("v"),"mypos");
}
alphatoverb(8,L("voice"),L("pos"));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_conj
_xALPHA
_xALPHA
@@
# verb prep noun conj alpha
@PRE
<5,5> var("verb");
@CHECK
if (pnname(N(1)) == "_vg")
{
if (N("last vg",1))
L("vg1") = N("last vg",1); # 05/22/07 AM.
else
L("vg1") = N(1);
S("v1") = pnvar(L("vg1"),"verb node"); # 05/22/07 AM.
}
else
S("v1") = N(1);
S("pos") = pnvar(S("v1"),"mypos");
S("c") = samevconj(S("v1"),N(5));
if (!S("c"))
fail();
if (N("noun",5) && (S("c") == "inf" || S("c") == "-s"))
{
if (pnname(N(6)) == "_noun" || pnname(N(6)) == "_prep")
fail(); # Could be apposition...
}
@POST
alphatovg(5,"active",S("pos"));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_xWILD [one match=(_noun _np)]
_xWILD [plus match=( \, _conj)]
_xALPHA
_xWILD [one lookahead match=(_noun _np _prep _adv _advl)]
@@
# verb noun conj alpha noun
@PRE
<4,4> var("verb");
@CHECK
if (pnname(N(1)) == "_vg")
L("v1") = N("verb node",1);
else
L("v1") = N(1);
if (!samevconj(L("v1"),N(4)))
fail();
@POST
alphatovgcopy(4,N(1));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
_xWILD [plus match=(_conj \, )]
_xALPHA
_xWILD [one lookahead match=(_noun _np)]
@@
# verb noun conj alpha
# vg np conj alpha
# verb dqan conj alpha
@CHECK
if (N("pos50 vn-conj-a",1))
fail();
if (N(3) && N("adv",5))
fail();
@POST
if (pnname(N(1)) == "_vg")
L("v1") = N("verb node",1);
else
L("v1") = N(1);
if (!N("verb",5))
{
if (singular(N(5)))
fixnpnonhead(5);
else
fixnphead(5);
}
else if (samevconj(L("v1"),N(5))) # 05/02/07 AM.
{
if (pnvar(L("v1"),"mypos") == "VBD")
alphatovg(5,"active","VBD");
# Still ambig, possible parallel construction.
}
else if (singular(N(2)) && singular(N(5))) # 04/23/07 AM.
{
fixnpnonhead(5); # 04/23/07 AM.
}
else
{
if (!samevconj(L("v1"),N(5)))
{
if (singular(N(5)))
fixnpnonhead(5);
else
fixnphead(5);
}
else
{
S("vc1") = vconj(L("v"));
alphatoverb(5,0,S("vc1"));
}
}
N("pos50 vn-conj-a",1) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_conj
_xALPHA
_xWILD [one lookahead fail=(_np _det _quan _num _xNUM _adj)]
@@
# vg dqan conj alpha
# verb dqan conj alpha
# verb noun conj alpha
@PRE
<7,7> var("noun");
@CHECK
L("n") = lasteltnode(5);
if (plural(L("n")) && !plural(N(7)))
fail();
if (singular(L("n")) && !singular(N(7)))
fail();
if (vconj(N(1)) == vconj(N(7)))
fail(); # Possible parallel verb construction.
@POST
fixnphead(7);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_conj
_xALPHA
_xWILD [one match=(_np)]
@@
# vg dqan conj alpha
# verb dqan conj alpha
# verb noun conj alpha
@PRE
<7,7> var("verb");
@CHECK
L("n") = lasteltnode(5);
if (vconj(N(1)) != vconj(N(7)))
fail();
@POST
if (pnname(N(1)) == "_vg")
L("v") = N("verb node",1);
else
L("v") = N(1);
L("pos") = pnvar(L("v"),"mypos");
alphatovg(7,0,L("pos"));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_conj ### (6)
_xALPHA ### (7)
_xWILD [one lookahead match=(_np)] ### (8)
_xWILD [one match=(_xEND _dbldash \, _prep )] ### (9)
@@
# vg np prep np to alpha
@PRE
<1,6> debug();
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"inf"))
fail();
@POST
alphatovg(6,"active","VB");
@RULES
_xNIL <-
_vg
_np
_prep
_np
to [s]
_xALPHA
@@
# vg np prep np conj alpha
# vg np prep np , alpha
@PRE
<1,1> varz("pos50 vnpn-conj-a");
<6,6> var("verb");
@CHECK
if (!N("noun",6) && !N("adj",6))
succeed();
if (samevconj(N(1),N(6)))
succeed();
if (vconjq(N(1),"-ing") && vconjq(N(6),"-ed"))
succeed();
fail();
@POST
N("pos50 vnpn-conj-a",1) = 1;
if (N("voice",1) == "active")
{
group(3,4,"_advl"); # prep np -> advl
N("pp",3) = 1;
}
else if (N("stem",3) == "by")
{
# Need by-np logic.
if (!N("sem",4) || N("sem",4) != "quantity") # etc...
{
group(3,4,"_advl");
N("pp",3) = 1;
}
else
{
group(3,4,"_advl");
N("pp",3) = 1;
N("by-np",3) = 1;
N("by-actor",3) = 1;
}
# Can fix up verbs also...
}
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
_prep ### (3)
_np ### (4)
_xWILD [one match=(_conj \, )] ### (5)
_xALPHA ### (6)
@@
# vg np prep np
@POST
group(4,5,"_advl");
N("pp",4) = 1;
@RULES
_xNIL <-
_vg
_np
_xWILD [star match=(_adv _advl)]
_prep
_np
_xWILD [one lookahead match=(_prep)]
@@
# vg np prep dqan alpha , vg
@PRE
<10,10> var("noun");
@POST
fixnphead(10);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_np ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_prep ### (5)
_xWILD [star match=(_det _pro)] ### (6)
_xWILD [star match=(_quan _num _xNUM)] ### (7)
_adj [star] ### (8)
_noun [plus] ### (9)
_xALPHA ### (10)
\, [lookahead] ### (11)
_xWILD [one match=(_verb _vg _modal)] ### (12)
@@
# vg np prep dqan alpha
@PRE
<10,10> var("noun"); # Require noun.
@CHECK
if (N("pos50 n-v-p-dqan-a",10))
fail();
if (N("verb",10))
{
if (!vconjq(N(10),"inf") && !vconjq(N(10),"-s"))
fail();
}
@POST
N("pos50 n-v-p-dqan-a",10) = 1;
L("lastn") = lasteltnode(9);
if (N("noun",10))
{
alphatonoun(10);
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_np
_xWILD [star match=(_adv _advl)]
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
@@
# vg np prep dqa alpha alpha
@PRE
<9,9> varz("pos50 v-n-p-dqa-aa");
@CHECK
if (!N(6) && !N(7) && !N(8))
fail();
if (!N("noun",9) && !N("adj",9))
fail();
@POST
N("pos50 v-n-p-dqa-aa",9) = 1;
if (N("noun",9) && !N("adj",9))
alphatonoun(9);
else if (N("adj",9) && !N("noun",9))
alphatoadj(9);
else if (N("noun",10))
{
if (plural(N(10)))
{
fixnphead(10);
fixnpnonhead(9);
}
else if (plural(N(9)))
fixnphead(9);
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_np ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_prep ### (5)
_xWILD [star match=(_det _pro)] ### (6)
_xWILD [star match=(_quan _num _xNUM)] ### (7)
_adj [star] ### (8)
_xALPHA ### (9)
_xALPHA ### (10)
@@
# vg np prep alpha
@CHECK
if (!N("adj",6))
fail();
if (N("verb",6))
{
if (vconjq(N(6),"-ing"))
fail();
}
@POST
alphatoadj(6);
@RULES
_xNIL <-
_vg
_np
_xWILD [star match=(_adv _advl)]
_xWILD [s match=(_prep) except=(to)]
_adv [star]
_xALPHA
_noun [lookahead]
@@
# vg np prep alpha dqan
# vg np prep alpha np
@PRE
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"-ing"))
fail();
if (verbfeat(N(1),"D1"))
fail();
@POST
alphatovg(6,"active","VBG");
@RULES
_xNIL <-
_vg
_np
_xWILD [star match=(_adv _advl)]
_xWILD [s match=(_prep) except=(to)]
_adv [star]
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM)]
@@
# whose dqan alpha when
# whose noun alpha when
@CHECK
if (!N("verb",6))
fail();
# Todo: agreement.
@POST
alphatoverb(6,"active","VBP");
@RULES
_xNIL <-
whose [s]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_whword [s lookahead]
@@
# when noun alpha alpha
@CHECK
if (!N("noun",3) || !N("verb",4))
fail();
if (plural(N(2)) || N("mass",2))
fail();
if (!plural(N(3)))
fail();
if (vconjq(N(4),"-s"))
fail();
@POST
alphatoverb(4,"active","VBP");
fixnphead(3);
@RULES
_xNIL <-
when [s]
_noun
_xALPHA
_xALPHA
_xWILD [one match=(_xEND)]
@@
# alpha whose [NIBBLE]
@CHECK
if (!N("noun",1))
fail();
@POST
fixnphead(1);
@RULES
_xNIL <-
_xALPHA
\,
whose [s]
@@
# noun alpha alpha noun
@CHECK
if (N("pos50 n-aa-n",2))
fail();
@POST
N("pos50 n-aa-n",2) = 1;
if (N("verb",3) && vconjq(N(3),"-s"))
{
alphatoverb(3,"active","VBZ");
}
else if (plural(N(1)) && plural(N(4)))
{
if (N("verb",2) && vconjq(N(2),"-ing"))
alphatoverb(2,"active","VBG");
}
else if (!N("noun",2) && N("verb",2))
{
alphatovg(2,"active","VBP");
}
@RULES
_xNIL <-
_noun
_xALPHA
_xALPHA
_noun [lookahead]
@@
# noun alpha alpha that
@PRE
<3,3> var("verb");
<4,4> var("noun");
#@CHECK
# if (!N("mass",4) && !plural(N(4)))
# fail();
@POST
fixnphead(4);
alphatovg(3,"active","VBP");
@RULES
_xNIL <-
_np
_adv [star]
_xALPHA
_xALPHA
that [s lookahead]
@@
# fnword alpha verb
# conj alpha verb
@CHECK
if (!N("noun",2) && !N("unknown",2))
fail();
# if (vconjq(N(2),"-ing")
# || vconjq(N(2),"-edn"))
# fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(_fnword _conj)]
_xALPHA
_xWILD [one lookahead match=(_verb _vg _modal)]
@@
# , which alpha noun alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(5),"-ing"))
fail();
@POST
alphatoverb(3,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=( \, _clausesep)]
which [s]
_xALPHA
_noun
_xALPHA
@@
# noun prep alpha
@PRE
<3,3> var("noun");
@CHECK
if (!N("noun",3) && !N("unknown",3))
fail();
if (N("verb",3))
{
if (!vconjq(N(3),"inf"))
fail(); # Too ambig.
}
if (N("pos50 n-prep-a-x",3))
fail();
@POST
N("pos50 n-prep-a-x",3) = 1;
if (N("noun",3) || N("unknown",3))
fixnphead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [s one match=(of for)]
_xALPHA
_xWILD [one lookahead match=(_adv _advl _prep _vg _verb _modal)]
@@
#########################################################
# adj alpha $ [DEFAULTS]
@CHECK
if (N("pos50 j-a-end",2))
fail();
@POST
N("pos50 j-a-end",2) = 1;
if (N("noun",2))
fixnphead(2);
@RULES
_xNIL <-
_adj
_xALPHA
_xEND
@@
# noun alpha $ [DEFAULTS]
# Note: from frequent alpha patterns.
# Note: Assumes more context-rich patterns have applied.
@CHECK
if (N("pos50 n-a-end",2))
fail();
@POST
N("pos50 n-a-end",2) = 1;
# If cap, not sure of singular plural.
if (plural(N(1)) && !N("cap",1))
{
if (N("noun",2))
{
if (plural(N(2)))
fixnphead(2);
else
alphatoverb(2,0,0);
}
}
else
{
if ((N("noun",2) || N("unknown",2))
&& singular(N(2)))
fixnphead(2);
else # Reduce anyway, assume verb would be trans...
fixnphead(2); # 05/22/07 AM.
}
@RULES
_xNIL <-
_noun
_xALPHA
_xWILD [one lookahead match=( \) _dbldash _xEND )]
@@
# noun alpha np [DEFAULTS]
# Note: somewhat weak rule, but (some parts) statistically good.
@CHECK
if (N("noun",2))
fail();
@POST
if (N("verb",2))
alphatoverb(2,0,0);
else if (N("adv",2))
alphatoadv(2);
else if (N("adj",2))
alphatoadj(2);
else # default
alphatoadv(2); # [DEFAULT]
@RULES
_xNIL <-
_noun
_xALPHA
_np [lookahead]
@@
# to alpha prep [DEFAULTS]
# Note: a default rule.
@PRE
<1,1> vareq("stem","to");
@CHECK
if (N("pos50 to-a-p",2))
fail();
@POST
N("pos50 to-a-p",2) = 1;
if (N("verb",2) && vconjq(N(2),"inf"))
alphatovg(2,"active","VB");
@RULES
_xNIL <-
_prep
_xALPHA
_prep
@@
# prep alpha prep [DEFAULTS]
# Note: a default rule.
@PRE
<1,1> varne("stem","to");
@CHECK
if (N("pos50 p-a-p",2))
fail();
@POST
N("pos50 p-a-p",2) = 1;
if (N("verb",2))
{
if (vconjq(N(2),"-ing"))
alphatovg(2,"active","VBG");
}
else if (N("noun",2))
fixnphead(2);
@RULES
_xNIL <-
_prep
_xALPHA
_prep
@@
# verb prep alpha $
# vg prep alpha $
# Todo: Need clause-level patterns for this.
# Eg: "product came to market"
@PRE
<1,1> var("cap");
<3,3> vareq("stem","to");
<4,4> var("noun");
@POST
fixnphead(4);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
_prep
_xWILD [one match=(market)]
_xEND
@@
# prep alpha $ [DEFAULTS]
@CHECK
if (N("pos50 p-a-end",2))
fail();
@POST
N("pos50 p-a-end",2) = 1;
if (N("verb",2))
{
if (N("stem",1) == "to")
{
if (vconjq(N(2),"inf"))
{
alphatoverb(2,"active","VB");
L("done") = 1;
}
}
else if (vconjq(N(2),"-ing"))
L("done") = 1; # Ambig.
else if (vconjq(N(2),"-en"))
{
alphatoverb(2,"passive","VBN");
N("pos50 psv15",2) = 1;
L("done") = 1;
}
}
if (N("noun",2) && !L("done"))
{
fixnphead(2);
}
@RULES
_xNIL <-
_prep
_xALPHA
_xEND
@@
# prep alpha alpha [DEFAULTS]
# Note: Seems to work out that 2nd alpha is a noun, mainly.
@PRE
<1,1> varne("stem","to"); # 05/02/07 AM.
@CHECK
if (N("pos50 p-a-a",3))
fail();
@POST
N("pos50 p-a-a",3) = 1;
if (N("noun",3))
fixnphead(3);
@RULES
_xNIL <-
_prep
_xALPHA
_xALPHA
_xWILD [one lookahead fail=(_xALPHA _noun)]
@@
#_xNIL <-
# _prep
# _xALPHA
# _xALPHA
# @@
# noun alpha prep [DEFAULTS]
@CHECK
if (N("pos50 n-a-p",2))
fail();
@POST
N("pos50 n-a-p",2) = 1;
if (N("noun",2) && N("verb",2))
; # Dunno.
else if (N("noun",2))
fixnphead(2);
else if (N("verb",2))
alphatoverb(2,0,0);
@RULES
_xNIL <-
_noun
_xALPHA
_prep [lookahead]
@@
# adv alpha noun [DEFAULTS]
@CHECK
if (N("pos50 x-a-n",2))
fail();
@POST
N("pos50 x-a-n",2) = 1;
if (N("verb",2))
alphatoverb(2,0,0);
else if (N("adj",2))
alphatoadj(2);
@RULES
_xNIL <-
_xWILD [plus match=(_adv _advl)]
_xALPHA
_noun [lookahead]
@@
# noun alpha noun [DEFAULTS]
@CHECK
if (N("pos50 n-a-n",2))
fail();
if (N("noun",2) && N("verb",2))
fail();
@POST
N("pos50 n-a-n",2) = 1;
if (N("verb",2))
{
alphatoverb(2,"active","VBP");
}
else if (N("noun",2))
fixnpnonhead(2);
else if (N("verb",2))
alphatoverb(2,0,0);
else if (N("adj",2))
alphatoadj(2);
@RULES
_xNIL <-
_noun
_xALPHA
_noun [lookahead]
@@
# still
@POST
alphatoadv(1);
@RULES
_xNIL <-
still
@@
# alpha noun prep np vg
# Note: ellipted verb (that) ...
@PRE
<1,1> var("verb");
@CHECK
if (!verbfeat(N(1),"T5"))
fail();
@POST
alphatoverb(1,0,0);
@RULES
_xNIL <-
_xALPHA
_noun
_prep
_np
_vg
@@
# alpha alpha dqan verb
@CHECK
if (!N("noun",1))
fail();
if (!N("verb",2))
fail();
if (!verbfeat(N(2),"T5"))
fail();
@POST
alphatoverb(2,"active","VBP");
@RULES
_xNIL <-
_xALPHA
_xALPHA
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_num _quan _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# ^ dqan , which alpha
@CHECK
if (!N("verb",8))
fail();
# N-V checks...
@POST
alphatoverb(8,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
\,
which [s]
_xALPHA
@@
# alpha noun vg
# Note: ellipted verb (that) ...
@PRE
<1,1> var("verb");
@CHECK
if (!verbfeat(N(1),"T5"))
fail();
@POST
alphatoverb(1,0,0);
@RULES
_xNIL <-
_xALPHA
_noun
_xWILD [star match=(_adv _advl)]
_vg
@@
# Capitalization issues.
@PRE
<1,1> var("cap");
@POST
pnrename(N(1),"_noun"); # adj -> noun
if (N("cap",1))
cappos(N(1),0);
N("pos50 adj->noun",1) = 1;
@RULES
_xNIL <-
_adj
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
_xNIL <-
_adj
\, [opt lookahead]
_adjc
@@
# which alpha that
@PRE
<3,3> var("verb");
@POST
alphatoverb(3,"active","VBP");
@RULES
_xNIL <-
which [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
that [s]
@@
# np vg vg
@PRE
<3,3> varz("pos50 n-v-v pass"); # Loop guard.
#<4,4> var("by-actor");
@CHECK
if (vgassigned(N(3)) && vgassigned(N(5)))
fail();
if (!vconjq(N(3),"-en"))
fail();
if (!vconjq(N(5),"-ed") && !vconjq(N(5),"inf")
&& !vconjq(N(5),"-s"))
fail();
@POST
N("pos50 n-v-v pass",3) = 1;
# N("by-actor",4) = 1;
if (vgassigned(N(3)))
{
if (N("tense",3))
{
if (vconjq(N(5),"-en"))
fixvg(N(5),"passive","VBN");
else
fixvg(N(5),"active","VBP");
}
}
else if (vgassigned(N(5)))
{
if (vconjq(N(3),"-en"))
fixvg(N(3),"passive","VBN");
}
else # Neither assigned.
{
if (vconjq(N(3),"-en"))
fixvg(N(3),"passive","VBN");
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_vg
_xWILD [plus match=(_adv _advl)]
_vg
@@
# np advl vg advl alpha
@PRE
<3,3> varz("pos50 n-a-v-a-alpha");
<5,5> var("verb");
@CHECK
if (N("voice",3) && N("voice",3) != "passive")
fail();
if (!vconjq(N(3),"-en"))
fail();
if (!vconjq(N(5),"-ed"))
fail();
@POST
N("pos50 n-a-v-a-alpha",3) = 1;
alphatovg(5,"active","VBD");
fixvg(N(3),"passive","VBN");
@RULES
_xNIL <-
_np
_xWILD [star match=(_advl _adv)]
_vg
_xWILD [plus match=(_advl _adv)]
_xALPHA
_xWILD [one lookahead match=(_np _prep _det _quan
_num _xNUM)]
@@
# note: heard of asbestos having (active...)
@PRE
<1,1> varz("voice");
<1,1> varz("pos50 v-p-n-v");
<5,5> vareq("voice","active");
@CHECK
L("v") = N("verb node",1);
if (!vconjq(L("v"),"-en"))
fail();
L("v5") = N("verb node",5);
if (vconjq(L("v5"),"-ing"))
fail();
@POST
N("pos50 v-p-n-v",1) = 1; # Loop guard.
fixvg(N(1),"passive","VBN");
@RULES
_xNIL <-
_vg
_prep
_np
_xWILD [star match=(_adv _advl)]
_vg
@@
# , advl ,
@POST
group(1,3,"_advl");
@RULES
_xNIL <-
\,
_advl
\,
@@
# fnword noun vg
@POST
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
_noun [plus]
_xWILD [one lookahead match=(_vg)]
@@
# verb np adjc
# Note: Scalar converts to an adverbial.
# Ex: "looks 2 times heavier"
@PRE
<3,3> vareq("sem","scalar");
@POST
group(3,3,"_advl");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_adv _advl)]
_np
_adjc [lookahead]
@@
# Lone noun.
@POST
pncopyvars(2);
sclearpos(1,1); # Zero out token info.
singler(2,2);
@RULES
_np <-
_xWILD [one match=(_prep _vg)]
_noun
_xWILD [one lookahead match=(_prep _conj _qEOS _fnword _xEND)]
@@
# now verb
@CHECK
if (N("mypos",2))
fail();
@POST
pnrename(N(2),"_adv");
N("mypos",2) = "RB"; # now/RB.
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
now [s]
_xWILD [one lookahead match=(_verb _vg _modal _adv)]
@@
# noun that noun
# products that year.
# Note: ellipted adverbial "products during that year..."
@CHECK
if (N("mypos",3))
fail();
if (N("sem",4) != "date")
fail();
if (!singular(N(4)))
fail();
@POST
pnrename(N(3),"_det");
N("mypos",3) = "DT"; # that/DT.
group(3,4,"_advl"); # np -> advl
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
that [s]
_xWILD [one lookahead match=(_noun _np)]
_xWILD [one match=( \, )]
@@
# now
# Note: default.
@PRE
<1,1> varz("mypos");
@POST
chpos(N(1),"RB"); # now/RB. [default].
pnrename(N(1),"_adv");
@RULES
_xNIL <-
now [s]
@@
# prep alpha ,
@CHECK
if (!N("noun",2) && !N("unknown",2))
fail();
if (N("verb",2))
if (vconjq(N(2),"-ing"))
fail();
if (singular(N(2)) && !massnoun(N(2)))
fail();
@POST
alphatonoun(2);
@RULES
_xNIL <-
_xWILD [s one match=(_prep) except=(to)]
_xALPHA
_xWILD [one lookahead match=(\, _clausesep)]
@@
# noun adj noun
# Note: Simplify np formation.
@CHECK
if (!N("cap",1) && !N("num",1) && !N("possessive",2))
fail();
@POST
pnrename(N(1),"_adj"); # noun -> adj
@RULES
_xNIL <-
_noun
_adj [lookahead]
_noun
@@
# Note: Fix noun list.
@PRE
<3,3> var("adj");
@POST
fixnpnonhead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_adj [star lookahead]
_noun [plus]
\,
_conj
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# Clause level patterns.
@PRE
<3,3> varz("pos50 dbldbl"); # Opt: don't do the long match twice.
<6,6> var("verb");
# check agreement...
@POST
N("pos50 dbldbl",3) = 1;
alphatoverb(6,"active","VBP");
@RULES
_xNIL <-
_xSTART
_np
_dbldash
_xWILD
_dbldash
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_adj _noun _np)]
@@
# if noun alpha alpha
@CHECK
if (!N("noun",3) || !N("verb",3))
fail();
if (N("pos num",3) != 2)
fail();
if (!N("noun",4) || !N("verb",4))
fail();
if (N("pos num",4) != 2)
fail();
if (!plural(N(3)))
fail();
if (!singular(N(4)))
fail();
@POST
fixnphead(3);
alphatoverb(4,"active","VBP");
@RULES
_xNIL <-
if [s]
_noun
_xALPHA
_xALPHA
\, [lookahead]
@@
# ^ np vg np , alpha
# Note: clausal.
@PRE
<10,10> var("verb");
@CHECK
if (!vconjq(N(10),"-ing"))
fail();
@POST
alphatoverb(10,"active","VBG");
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_xWILD [one match=(_np _pro)] ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_vg ### (5)
_xWILD [star match=(_adv _advl)] ### (6)
_xWILD [one match=(_np _pro)] ### (7)
_xWILD [star match=(_adv _advl)] ### (8)
\, ### (9)
_xALPHA ### (10)
@@
# np vg $
@CHECK
if (N("voice",2))
fail();
S("v") = N("verb node",2);
S("t") = nodetext(S("v"));
if (!finddictattr(S("t"),"trans"))
fail();
if (!vconjq(N(2),"-en"))
fail();
@POST
fixvg(N(2),"passive","VBN");
N("pos50 psv16",2) = 1;
@RULES
_xNIL <-
_np
_vg
_xEND
@@
# that alpha alpha adj to verb
# Verb pattern... X7 ... x finds y (to be) happy to comply.
@PRE
<2,2> var("noun");
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"inf") && !vconjq(N(3),"-ed"))
fail();
@POST
chpos(N(1),"WDT"); # that/WDT.
alphatoverb(3,"active","VBP");
fixnphead(2);
@RULES
_xNIL <-
that [s]
_xALPHA
_xALPHA
_adj [lookahead]
to [s]
_xWILD [one match=(_verb _vg)]
@@
# adv conj alpha
@PRE
<3,3> var("adv");
@POST
alphatoadv(3);
@RULES
_xNIL <-
_adv
_conj
_xALPHA
_xWILD [one lookahead match=(_prep _xEND)]
@@
# noun list
# np conj noun
@CHECK
S("sem1") = N("sem",1);
S("sem3") = N("sem",3);
if (!S("sem1"))
fail();
if (S("sem1") != S("sem3"))
fail();
@POST
nountonp(3,1);
if (N("sem val",1))
L("vals")[0] = N("sem val",1);
if (N("sem val",3))
L("vals")[1] = N("sem val",3);
group(1,3,"_np");
N("sem",1) = S("sem1");
if (L("vals"))
N("sem vals",1) = L("vals");
@RULES
_xNIL <-
_np
_conj
_noun
_xWILD [one lookahead match=( \, _conj )]
@@
# np conj noun
@POST
fixnphead(3);
nountonp(3,1);
if (N("sem val",1))
L("vals")[0] = N("sem val",1);
if (N("sem val",3))
L("vals")[1] = N("sem val",3);
group(1,3,"_np");
if (S("sem1"))
N("sem",1) = S("sem1");
if (L("vals"))
N("sem vals",1) = L("vals");
@RULES
_xNIL <-
_np
_conj
_noun
\, [lookahead]
_np
@@
# noun list
# vg np conj np
@POST
if (N("sem text",2))
L("arr")[0] = N("sem text",2);
if (N("sem text",4))
L("arr")[1] = N("sem text",4);
group(2,4,"_np");
if (L("arr"))
N("sem texts",2) = L("arr");
@RULES
_xNIL <-
_vg
_np
_conj
_np
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [s one fail=(_xALPHA _verb _vg of _adj _adjc)]
@@
# adj prep np vg
@CHECK
if (N("pos50 a-p-n-v",4))
fail();
S("v") = N("verb node",4);
if (!S("v"))
fail();
if (pnvar(S("v"),"mypos"))
fail();
@POST
N("pos50 a-p-n-v",4) = 1;
fixvg(N(4),"passive","VBN");
@RULES
_xNIL <-
_adj
_prep
_np
_vg
@@
# np , though alpha
@PRE
<5,5> var("verb");
@CHECK
if (!vconjq(N(5),"-ing"))
fail();
@POST
alphatovg(5,"active","VBG");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
though [s]
_xWILD [star match=(_adv _advl)]
_xALPHA
\, [lookahead]
@@
# Ellipted pattersn.
# noun advl pro verb alpha
# "the cost of the goods they purchased rose"
@PRE
<2,2> var("pp"); # prep phrase.
<5,5> var("verb");
@CHECK
if (!vconjq(N(5),"-ed"))
fail();
@POST
alphatovg(5,"active","VBP");
fixverb(N(4),"active","VBP");
group(3,4,"_clause"); # An embedded clause.
setunsealed(3,"true"); # 07/10/12 AM.
group(3,3,"_advl");
setunsealed(3,"true"); # 07/10/12 AM.
N("ellipted-that-clause",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_np _noun)]
_advl
_proSubj [s]
_verb
_xALPHA
@@
# verb that noun that noun verb
@PRE
<1,1> varz("pos50 that-n-that");
@POST
N("pos50 that-n-that",1) = 1; # Loop guard.
chpos(N(2),"DT"); # that/DT
chpos(N(4),"IN"); # that/IN
@RULES
_xNIL <-
_vg
that [s lookahead]
_np
that [s]
_np
_vg
@@
# det alpha alpha noun
@PRE
<2,2> var("noun");
<3,3> var("verb");
<3,3> varz("noun");
@POST
alphatovg(3,"active","VBP");
fixnphead(2);
@RULES
_xNIL <-
_det
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_np _det _quan _num _xNUM)]
@@
# ^ np vg np
@PRE
<5,5> varz("voice");
@POST
fixvg(N(5),"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_np _pro)]
_xWILD [star match=(_advl _adv)]
_vg
_xWILD [star match=(_advl _adv)]
_xWILD [one match=(_np _pro)]
_xWILD [star match=(_advl _adv)]
_xWILD [s one lookahead match=(to \, _conj)]
@@
# verb that much
@PRE
<2,2> varz("pos50 v-that-much");
@POST
N("pos50 v-that-much",2) = 1; # Loop guard.
chpos(N(2),"RB"); # that/RB
chpos(N(3),"RB"); # much/RB
pnrename(N(2),"_adv");
pnrename(N(3),"_adv");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
that [s]
much [s]
@@
# verb prep np vg
@PRE
<4,4> varz("pos50 v-p-n-v");
@CHECK
if (!vconjq(N(4),"-ed"))
fail();
@POST
N("pos50 v-p-n-v",4) = 1;
fixvg(N(4),"active","VBP");
@RULES
_xNIL <-
_verb
_prep
_np
_vg
@@
# parens.
@POST
group(1,3,"_advl");
@RULES
_xNIL <-
\(
_xWILD [plus fail=(_xALPHA \) )]
\)
@@
# verb , adv alpha ,
# verb -- adv alpha --
@PRE
<4,4> var("verb");
@CHECK
if (!vconjq(N(1),"-en"))
fail();
@POST
alphatovg(4,"passive","VBN");
N("pos50 psv17",4) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
\,
_adv
_xALPHA
\,
@@
_xNIL <-
_xWILD [one match=(_verb _vg)]
_dbldash
_adv
_xALPHA
_dbldash
@@
# Note: ellipted-that patterns.
@CHECK
if (!copulaq(N(3)))
fail();
if (!verbfeat(N(1),"T5"))
fail();
@POST
group(2,6,"_clause");
setunsealed(2,"true"); # 07/10/12 AM.
group(2,2,"_advl");
setunsealed(2,"true"); # 07/10/12 AM.
N("ellipted-that",2) = 1;
@RULES
_xNIL <-
_vg
_xWILD [one match=(_np _pro)]
_vg
_xWILD [star match=(_adv _advl)]
_adjc
_np
@@
# , adv adj
@POST
group(2,3,"_adj");
@RULES
_xNIL <-
\,
_adv [plus]
_adj
@@
# vg np prep np conj dqan alpha
# vg prep np conj dqan alpha
@PRE
<10,10> var("noun");
@POST
fixnphead(10);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_np [opt] ### (2)
_prep ### (3)
_np ### (4)
_conj ### (5)
_xWILD [star match=(_det _pro)] ### (6)
_xWILD [star match=(_quan _num _xNUM)] ### (7)
_adj [star] ### (8)
_noun [plus] ### (9)
_xALPHA ### (10)
_xWILD [one match=(_dbldash _prep _fnword _xEND)] ### (11)
@@
# vg np prep np conj alpha
@PRE
<5,5> var("noun");
@POST
fixnphead(5);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_np
_conj
_xALPHA
_xWILD [one match=(_dbldash _xEND)]
@@
# vg np prep np conj alpha noun
@PRE
<6,6> var("verb");
@CHECK
L("v1") = N("verb node",1);
if (!samevconj(L("v1"),N(6)))
fail();
@POST
alphatovg(6,"active","VBP");
@RULES
_xNIL <-
_vg
_np
_prep
_np
_conj
_xALPHA
_noun [lookahead]
@@
# np conj alpha $
# noun conj alpha $
@PRE
<3,3> var("noun");
@POST
fixnphead(3);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_xEND
@@
_xNIL <-
_xWILD [one match=(_noun _np)]
_conj
_xALPHA
_xWILD [s one lookahead match=(who whom)]
@@
# verb noun , noun alpha conj
# Note: noun lists.
@PRE
<5,5> var("noun");
@CHECK
if (plural(N(2)) && plural(N(5)))
succeed();
if (singular(N(2)) && singular(N(5)))
succeed();
fail();
@POST
fixnphead(5);
group(4,5,"_np");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
\,
_noun
_xALPHA
_xWILD [one lookahead match=( _conj \, )]
@@
# noun , noun conj alpha alpha
@PRE
<6,6> var("noun");
@CHECK
if (!N("noun",5) && !N("adj",5))
fail();
if (plural(N(3)) && plural(N(6)))
succeed();
if (singular(N(3)) && singular(N(6)))
succeed();
fail();
@POST
fixnphead(6);
fixnpnonhead(5);
group(5,6,"_np");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [plus match=( \, _conj)]
_xWILD [one match=(_noun _np)]
_xWILD [plus match=( \, _conj)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=( \, _conj)]
@@
# np , np and np
@POST
group(1,5,"_np");
@RULES
_xNIL <-
_np
_xWILD [plus match=(\, _conj)]
_np
_xWILD [plus match=(\, _conj)]
_np
@@
# before vg np
@POST
if (!N("mypos",1))
chpos(N(1),"IN");
group(1,3,"_clause");
setunsealed(1,"true");
group(1,1,"_advl");
setunsealed(1,"true");
@RULES
_xNIL <-
before [s]
_vg
_np
_conj [lookahead]
_vg
@@
# ^ np , prep vg np , alpha
@PRE
<9,9> var("verb");
@CHECK
if (!vconjq(N(9),"inf") && !vconjq(N(9),"-s")
&& !vconjq(N(9),"-ed"))
fail();
@POST
alphatovg(9,"active","VBP");
@RULES
_xNIL <-
_xWILD [one match=( _xSTART \, )] ### (1)
_xWILD [one match=(_noun _np)] ### (2)
\, ### (3)
_prep ### (4)
_vg ### (5)
_np ### (6)
_xWILD [star match=(_adv _advl)] ### (7)
\, ### (8)
_xALPHA ### (9)
@@
# ^ np , alpha
@PRE
<5,5> var("verb");
<5,5> var("adj");
<5,5> vareq("pos num",2);
@POST
if (vconjq(N(5),"-en") && verbfeat(N(5),"jj-pref"))
alphatoadj(5);
else
alphatovg(5,0,0);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
\,
_xALPHA
@@
# , adv alpha ,
@PRE
<3,3> var("adj");
@CHECK
if (N("verb",3))
{
if (!vconjq(N(3),"inf"))
fail();
}
@POST
alphatoadj(3);
@RULES
_xNIL <-
\,
_adv [plus]
_xALPHA
\, [lookahead]
@@
# verb prep np verb
@CHECK
if (!copulaq(N(1)))
fail();
S("v") = N("verb node",5);
if (pnvar(S("v"),"mypos"))
fail();
if (!vconjq(S("v"),"-en"))
fail();
@POST
fixvg(N(5),"passive","VBN");
N("pos50 psv18",5) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_adv [star]
_prep
_np
_vg
@@
# noun adj
# num adj
@CHECK
if (!N("num",2) && !N("possessive",3) && !N("cap",2))
fail();
@POST
pnrename(N(2),"_num"); # noun -> num
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_noun
_xWILD [one match=(_adj _noun)]
@@
# vg np prep np vg
# verb noun prep noun verb
# vg prep np vg
@PRE
<6,6> varz("pos50 vnpnv");
@CHECK
if (!verbfeat(N(1),"D1"))
fail();
if (pnname(N(6)) == "_vg")
{
if (N("voice",6))
fail();
}
else if (N("mypos",6))
fail();
if (!vconjq(N(6),"-en"))
fail();
@POST
N("pos50 vnpnv",6) = 1;
fixvg(N(6),"passive","VBN");
N("voice",6) = "passive"; # Loop guard too!
N("pos50 psv19",6) = 1;
N("ellipted-that-is",6) = 1;
N("glom",6) = "right";
N("glom",5) = "left";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [star match=(_np _noun)]
_xWILD [star match=(_adv _advl)]
_prep
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
_xWILD [one lookahead match=(_np _noun _det _quan _num _xNUM _adj)]
@@
# vg prep vg
# vg to vg
# Note: May need semantic grounds, or at least trans/intr check.
@PRE
<2,2> varz("voice");
<3,3> vareq("stem","to");
@CHECK
if (!vconjq(N(2),"-en"))
fail();
if (phrprepverbq(N(2),N(3)))
fail();
if (verbfeat(N(2),"intr") || verbfeat(N(2),"T3"))
fail();
@POST
fixvg(N(2),"passive","VBN");
N("pos50 psv20",2) = 1;
@RULES
_xNIL <-
_xWILD [one fail=(_xALPHA _noun _np _conj)]
_vg
_prep
_vg
@@
# more adv than alpha conj noun
@CHECK
if (!N("noun",4) && !N("unknown",4))
fail();
@POST
fixnphead(4);
@RULES
_xNIL <-
more [s]
_adv [plus]
than [s]
_xALPHA
_xWILD [plus lookahead match=(_conj \, )]
_adv [star]
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# Note: use apposition to assume verb.
@PRE
<6,6> var("verb");
@CHECK
# Todo: nvagree.
L("n") = lasteltnode(5);
if (plural(L("n")) && vconjq(N(6),"-s"))
fail();
if (singular(L("n")) && vconjq(N(6),"inf"))
fail();
@POST
alphatovg(6,0,0);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_prep [lookahead]
_np
\,
_xWILD [one match=(_det _quan _num _xNUM _adj _noun _np)]
@@
# noun alpha
@PRE
<2,2> var("verb");
@CHECK
if (!vconjq(N(2),"-ing"))
fail();
if (!plural(N(1)))
fail();
@POST
alphatovg(2,"active","VBG");
@RULES
_xNIL <-
_noun
_xALPHA
@@
# Note: Need more general embedded clause handling.
@PRE
<4,4> varz("pos50 n-that-v-adj");
@POST
N("pos50 n-that-v-adj",4) = 1; # Loop guard.
if (!N("mypos",3))
chpos(N(3),"WDT"); # that/WDT
if (N("mypos",5) == "VBG")
chpos(N(5),"JJ"); # vbg -> jj
group(2,5,"_clause");
setunsealed(2,"true"); # 07/10/12 AM.
group(1,3,"_advl");
setunsealed(1,"true"); # 07/10/12 AM.
@RULES
_xNIL <-
_xWILD [one match=(_dbldash)]
_np
that [s]
_vg
_adj
_xWILD [one match=(_dbldash)]
@@
# --
@POST
group(1,3,"_advl");
@RULES
_xNIL <-
_dbldash
_xWILD [plus fail=(_xALPHA _verb _noun _dbldash)]
_dbldash
@@
# advl who , vg
# np who , vg
@PRE
<4,4> varz("voice");
@CHECK
if (!vconjq(N(4),"-en"))
fail();
@POST
N("voice",4) = "passive";
fixvg(N(4),"passive","VBN");
N("pos50 psv21",4) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_advl _np)]
who [s lookahead]
\,
_vg
@@
# np , alpha np
# noun , alpha noun
# Note: A kind of apposition with a verb.
@PRE
<3,3> var("verb");
@CHECK
if (!N("cap",1) && !N("cap",4))
fail(); # Some appos validation.
# if (!vconjq(N(3),"-en"))
# fail();
@POST
if (vconjq(N(3),"-en")
&& N("adj",3) && verbfeat(N(3),"jj-pref"))
alphatoadj(3);
else
alphatovg(3,0,0);
N("pos50 psv22",3) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [one match=( \, \( )] # 04/25/10 AM.
_xALPHA
_xWILD [one lookahead match=(_noun _np)] # 04/25/10 AM.
@@
# vg prep prep np alpha
@PRE
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"-ing"))
fail();
@POST
alphatovg(6,"active","VBG");
@RULES
_xNIL <-
_vg
_prep
_prep
_np
_adv [star]
_xALPHA
_xWILD [one lookahead match=(_conj)]
@@
# np , vg np , alpha
# Add restrictions as needed. #
@PRE
<11,11> var("verb");
@POST
if (vconjq(N(11),"-en")
&& N("adj",11) && verbfeat(N(11),"jj-pref"))
alphatoadj(11);
else
alphatovg(11,0,0);
N("pos50 n,vn,a");
@RULES
_xNIL <-
_np ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
\, ### (3)
_xWILD [star match=(_adv _advl)] ### (4)
_vg ### (5)
_xWILD [star match=(_adv _advl)] ### (6)
_np ### (7)
_xWILD [star match=(_adv _advl)] ### (8)
\, ### (9)
_xWILD [star match=(_adv _advl)] ### (10)
_xALPHA ### (11)
@@
# vg np , alpha np
# vg np , alpha dqan
@PRE
<4,4> var("verb");
@POST
if (vconjq(N(4),"inf"))
alphatovg(4,"active","VB");
else
alphatovg(4,0,0);
@RULES
_xNIL <-
_vg
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_xWILD [one match=(_np _det _pro _pdt _quan _num _xNUM)]
@@
# vg np prep np , alpha alpha that
@PRE
<6,6> var("verb");
<7,7> var("noun");
@CHECK
if (!vconjq(N(6),"-ing"))
fail();
@POST
fixnphead(7);
alphatovg(6,"active","VBG");
@RULES
_xNIL <-
_vg
_np
_prep
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_xALPHA
that [s lookahead]
@@
# vg np prep np , alpha noun , vg
# Note: Trying to decide on verb list.
@PRE
<6,6> var("verb");
@CHECK
if (samevconj(N(1),N(6)))
{
S("voice") = N("voice",1);
L("v") = N("verb node",1);
S("pos") = pnvar(L("v"),"mypos");
succeed();
}
if (samevconj(N(6),N(9)))
{
S("voice") = N("voice",9);
L("v") = N("verb node",9);
S("pos") = pnvar(L("v"),"mypos");
succeed();
}
fail();
@POST
alphatovg(6,S("voice"),S("pos"));
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
_prep ### (3)
_xWILD [one match=(_noun _np)] ### (4)
\, ### (5)
_xALPHA ### (6)
_xWILD [one lookahead match=(_np _noun)] ### (7)
\, ### (8)
_xWILD [one match=(_verb _vg)] ### (9)
@@
# verb alpha , alpha conj adv
@PRE
<2,2> var("adv");
<4,4> var("adv");
@POST
alphatoadv(4);
alphatoadv(2);
group(2,6,"_advl");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xALPHA
\,
_xALPHA
_xWILD [plus match=( \, _conj)]
_adv
_xWILD [one lookahead match=(_xEND)]
@@
# prep dqa , alpha
# adj , alpha
@PRE
<6,6> var("adj");
@POST
alphatoadj(6); # 06/08/07 AM.
@RULES
_xNIL <-
_xWILD [one match=(_prep)] ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [plus] ### (4)
\, ### (5)
_xALPHA ### (6)
@@
# det adj , alpha alpha
@PRE
<4,4> varz("verb");
<4,4> var("adj");
<5,5> varz("verb");
<5,5> var("noun");
@POST
fixnpnonhead(4);
@RULES
_xNIL <-
_det
_adj
\,
_xALPHA
_xALPHA
@@
# vg np advl vg
@PRE
<4,4> vareq("voice","passive");
@POST
group(4,5,"_clause");
setunsealed(4,"true"); # 07/10/12 AM.
group(4,4,"_advl");
setunsealed(4,"true"); # 07/10/12 AM.
@RULES
_xNIL <-
_vg
_np
_advl [plus]
_vg
_advl [star]
@@
# vg np advl that np vg np
@PRE
<4,4> varz("mypos");
@POST
chpos(N(4),"IN"); # that/IN
N("pos50 vna-that-nvn",4) = 1;
@RULES
_xNIL <-
_vg
_np
_advl [plus]
that [s]
_np [lookahead]
_vg
_np
@@
# vg np advl that
#@PRE
#<4,4> varz("mypos");
#@POST
# chpos(N(4),"WDT"); # that/WDT
# N("pos50 vna-that",4) = 1;
#@RULES
#_xNIL <-
# _vg
# _np
# _advl [plus]
# that [s]
# @@
# , adv prep np of np alpha prep np ,
@PRE
<7,7> var("verb");
@POST
alphatovg(7,0,"VBP");
@RULES
_xNIL <-
\,
_adv [plus]
_prep
_np
of [s]
_xWILD [one match=(_noun _np)]
_xALPHA
_prep
_np
\,
@@
# , det noun noun verb
@PRE
<4,4> var("cap");
<5,5> varz("mypos");
@POST
fixverb(N(5),"active","VBP");
@RULES
_xNIL <-
\,
_det
_noun
_noun
_verb
_xWILD [one lookahead match=(_dbldash)]
@@
# np vg that adj
# Note: adverbial 'that'
@PRE
<5,5> varz("mypos");
<5,5> varz("that-adv");
@CHECK
if (!copulaq(N(3)))
fail();
@POST
N("that-adv",5) = 1;
chpos(N(5),"RB"); # that/RB.
pnrename(N(5),"_adv"); # that -> adv. fnword -> adv.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
_adv [star]
that [s]
_xWILD [one lookahead match=(_adj _adjc)]
_xWILD [one match=(_prep _fnword)]
@@
# np vg that
# noun verb that
@PRE
<3,3> varz("pos50 nvthat");
@CHECK
L("v") = N("verb node",3);
if (!L("v"))
{
"err.txt" << "No verb node=" << phrasetext() << "\n";
fail(); # 03/04/10 AM.
}
if (pnvar(L("v"),"mypos"))
fail();
@POST
N("glom",2) = "right";
fixvg(N(3),"active","VBP");
N("pos50 nvthat",3) = 1; # Fix. # 10/14/10 AM.
@RULES
_xNIL <-
_xWILD [one match=( \, _conj)]
_np
_vg
that [s]
@@
# vg prep dqa alpha alpha prep vg
# Note: Context for favoring noun phrase.
@PRE
<3,3> varz("prosubj");
<3,3> varz("proobj");
<7,7> var("noun");
@CHECK
if (!N("noun",6) && !N("adj",6))
fail();
@POST
fixnphead(7);
fixnpnonhead(6);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_prep ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_xALPHA ### (6)
_xALPHA ### (7)
_prep [lookahead] ### (8)
_xWILD [one match=(_verb _vg _modal)] ### (9)
@@
# verb prep dqa alpha alpha by
# vg prep dqa alpha alpha by
@PRE
<7,7> var("noun");
@CHECK
if (!N("noun",6) && !N("adj",6))
fail();
if (!vconjq(N(1),"-en"))
fail();
@POST
if (!N("voice",1))
{
if (pnname(N(1)) == "_vg")
fixvg(N(1),"passive","VBN");
else
fixverb(N(1),"passive","VBN");
}
N("pos50 psv23",1) = 1;
fixnphead(7);
fixnpnonhead(6);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA
by [s lookahead]
@@
# verb prep dqan conj alpha
# vg prep dqan conj alpha
@PRE
<8,8> var("noun");
@CHECK
L("n") = lasteltnode(6);
if (plural(L("n")) && !plural(N(8)))
fail();
if (singular(L("n")) && !singular(N(8)))
fail();
if (N("verb",8))
{
if (vconj(N(1)) == vconj(N(8)))
fail(); # Parallel construction.
if (phrprepverbq(N(8),N(9)))
fail(); # Verb-particle good combo.
}
@POST
fixnphead(8);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_conj
_xALPHA
_xWILD [one match=(_prep)]
@@
# vg np prep dqan conj alpha np
@PRE
<9,9> var("verb");
@CHECK
if (pnname(N(10)) == "_noun"
&& plural(N(10))
&& vconjq(N(9),"inf"))
{
if (N("adj",9) || N("noun",9))
fail(); # Could still be noun or adj.
}
S("vc") = vconj(N(1));
if (S("vc") != vconj(N(9)))
fail();
@POST
L("v") = N("verb node",1);
if (L("v"))
L("pos") = pnvar(L("v"),"mypos");
alphatovg(9,N("voice",1),L("pos"));
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
_prep ### (3)
_xWILD [star match=(_det _pro)] ### (4)
_xWILD [star match=(_quan _num _xNUM)] ### (5)
_adj [star] ### (6)
_noun [plus] ### (7)
_conj ### (8)
_xALPHA ### (9)
_xWILD [one lookahead match=(_np _noun)] ### (10)
@@
# alpha np np
@PRE
<1,1> var("verb");
@CHECK
if (!verbfeat(N(1),"D1"))
fail();
@POST
alphatoverb(1,"active",0);
@RULES
_xNIL <-
_xALPHA
_np [lookahead]
_xWILD [one match=(_np _det _pro _quan _num _xNUM _adj _noun)]
@@
# of np vg
@CHECK
L("v") = N("verb node",3);
if (!L("v"))
fail();
if (pnvar(L("v"),"mypos"))
fail();
if (!vconjq(L("v"),"-en"))
fail();
@POST
fixvg(N(3),"passive","VBN");
N("pos50 psv24",3) = 1;
@RULES
_xNIL <-
of [s]
_np
_vg
_xWILD [s one lookahead match=(that by)]
@@
# prep dqan alpha that
@CHECK
if (!N("verb",6) && !N("adj",6))
fail();
if (N("noun",6) && numbersagree(N(1),N(6)))
fail();
@POST
if (N("verb",6))
alphatovg(6,0,0);
else
{
alphatoadj(6);
group(6,6,"_adjc");
}
@RULES
_xNIL <-
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [s one lookahead match=(that)]
@@
# np noun alpha verb
@PRE
<2,2> var("cap");
<3,3> var("verb");
@POST
nountonp(2,1);
alphatovg(3,"active","VBP");
group(2,3,"_clause");
setunsealed(2,"true"); # 07/10/12 AM.
N("ellipted-that",2) = 1;
group(2,2,"_advl");
setunsealed(2,"true"); # 07/10/12 AM.
@RULES
_xNIL <-
_np
_noun
_xALPHA
_vg
@@
# noun , dqan conj alpha+ prep
# Apposition
@CHECK
if (N(10))
{
if (!N("noun",10))
fail();
if (!N("noun",9) && !N("adj",9))
fail();
if (!N("noun",8) && !N("adj",8))
fail();
}
else if (N(9))
{
if (!N("noun",9))
fail();
if (!N("noun",8) && !N("adj",8))
fail();
}
else if (!N("noun",8))
fail();
@POST
if (N(10))
{
fixnphead(10);
fixnpnonhead(9);
fixnpnonhead(8);
}
else if (N(9))
{
fixnphead(9);
fixnpnonhead(8);
}
else
{
fixnphead(8);
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)] ### (1)
\, ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_conj ### (7)
_xALPHA ### (8)
_xALPHA [opt] ### (9)
_xALPHA [opt] ### (10)
_xWILD [one lookahead match=(_prep \, )] ### (11)
@@
# Apposition
# np , dqa , dqa alpha ,
@PRE
<7,7> var("noun");
@CHECK
if (!numbersagree(N(3),N(7)))
fail();
@POST
fixnphead(7);
group(4,6,"_adj");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)] ### (1)
\, ### (2)
_det ### (3)
_adj ### (4)
\, ### (5)
_adj ### (6)
_xALPHA ### (7)
\, [lookahead] ### (8)
@@
# aposS alpha , noun ,
@PRE
<2,2> var("noun");
@CHECK
if (!singular(N(2)))
fail();
@POST
fixnphead(2);
@RULES
_xNIL <-
_xWILD [one match=(_aposS)]
_xALPHA
\, [lookahead]
_xWILD [one match=(_noun _np)]
\,
@@
# Apposition
# np vg alpha
@PRE
<2,2> varz("pos50 nvathat");
<2,2> vareq("voice","passive");
<4,4> var("verb");
@CHECK
# Should be verbfeat(N(4),"T5")
succeed();
@POST
N("pos50 nvathat",2) = 1;
alphatovg(4,"active","VBP");
N("glom",4) = "right";
if (!N("mypos",5))
chpos(N(5),"IN"); # that/IN
@RULES
_xNIL <-
_np
_vg
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [s one lookahead match=(that)]
@@
# np for np , alpha for np
# Note: parallel construction.
@PRE
<5,5> var("noun");
@POST
fixnphead(5);
@RULES
_xNIL <-
_np
_prep
_np
\,
_xALPHA
_prep
_np
@@
# ^ alpha+ verb
@CHECK
if (N(5))
{
if (!N("noun",5))
fail();
S("o") = 5;
}
else if (N(4))
{
if (!N("noun",4))
fail();
S("o") = 4;
}
else if (N(3))
{
if (!N("noun",3))
fail();
S("o") = 3;
}
else
{
if (!N("noun",2))
fail();
S("o") = 2;
}
@POST
L("n") = fixnphead(S("o"));
while (--S("o") > 1)
{
L("n") = pnprev(L("n"));
if (pnvar(L("n"),"noun") || pnvar(L("n"),"adj"))
L("n") = fixnpnonhead(S("o"));
else
S("o") = 1; # Terminate loop.
}
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA [opt]
_xALPHA [opt]
_xALPHA [opt]
_xWILD [star lookahead match=(_adv _advl)]
_vg
@@
# vg np so much
@PRE
<4,4> varz("mypos");
@POST
alphatoadv(4);
pnrename(N(3),"_adv"); # so -> adv
@RULES
_xNIL <-
_vg
_np
so [s]
much
@@
# be prep noun conj verb
# Note: Remote linkage.
@PRE
<1,1> vareq("stem","be");
<1,1> varz("v-v-remote");
<6,6> varz("mypos");
@CHECK
if (!vconjq(N(6),"-en"))
fail();
@POST
fixverb(N(6),"passive","VBN");
N("v-v-remote",1) = N("v-v-remote",6) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_adv [star]
_prep
_noun
_xWILD [plus match=( \, _conj)]
_verb
@@
# Apposition
# noun , dqan 's alpha
@PRE
<1,1> var("cap");
<8,8> var("noun");
@POST
fixnphead(8);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_num _quan _xNUM)]
_adj [star]
_noun [plus]
_aposS
_xALPHA
_xWILD [one lookahead match=( \, _xEND _dbldash)]
@@
# dqan adj , adj , noun
@POST
group(2,6,"_adj");
@RULES
_xNIL <-
_xWILD [one match=(_det _pro _quan _num _xNUM _adj)]
_adj
_xWILD [plus match=( \, _conj)]
_adv [star]
_adj
\, [opt]
_noun [lookahead]
@@
# to dqan alpha $
# verb dqan alpha $
@CHECK
dqaninfo(3,4,5,6);
if (!numbersagree(S("first"),N(7)))
fail();
@POST
fixnphead(7);
@RULES
_xNIL <-
_xWILD [s one match=(_vg _verb to)]
_xWILD [star match=(_adv _advl)]
_xWILD [opt match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
_xWILD [one lookahead match=(_dbldash _xEND \, )]
@@
# the noun conj alpha that
@PRE
<4,4> var("noun");
@CHECK
if (!plural(N(4)))
fail();
@POST
fixnphead(4);
@RULES
_xNIL <-
the [s]
_noun
_conj
_xALPHA
that [s lookahead]
@@
# verb prep conj alpha
@PRE
<4,4> var("verb");
@CHECK
if (pnname(N(1)) == "_vg")
L("v") = N("verb node",1);
else
L("v") = N(1);
S("pos") = pnvar(L("v"),"mypos");
if (!S("pos"))
fail();
L("m") = vtreebanktopos(S("pos"));
if (!vconjq(N(4),L("m")))
fail();
@POST
alphatovg(4,0,S("pos"));
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_conj
_xALPHA
@@
# np , dqan alpha to vg
# Note: Appositive like, ellipted-copula.
@PRE
<8,8> vareq("stem","to");
@CHECK
if (!N("adj",7) && !N("verb",7))
fail();
if (N("verb",7) && !N("adj",7))
{
if (!vconjq(N(7),"-ing") && !vconjq(N(7),"-en"))
fail();
}
@POST
if (N("verb",7))
{
if (vconjq(N(7),"-ing"))
alphatovg(7,"active","VBG");
else
alphatovg(7,"passive","VBN");
N("ellipted-that-is",7) = 1;
N("pos50 n-,-n-a-to-v",7) = 1;
}
else
{
alphatoadj(7);
N("ellipted-copula",7) = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)] ### (1)
_xWILD [plus match=(_conj \, )] ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_xALPHA ### (7)
_prep ### (8)
_xWILD [one match=(_verb _vg _modal)] ### (9)
@@
# np that vg
# noun that verb
@PRE
<2,2> varz("mypos");
@CHECK
if (N("voice",4) != "active")
fail();
if (!vgassigned(N(4)))
fail();
@POST
L("pos") = vgpos(N(4));
if (L("pos") == "VBG")
chpos(N(2),"IN"); # that/IN. # 06/02/07 AM.
else
chpos(N(2),"WDT"); # that/WDT # 05/22/07 AM.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
that [s]
_xWILD [star match=(_adv _advl)]
_xWILD [one lookahead match=(_verb _vg _modal)]
@@
# , prep dqan , alpha
@POST
if (N("stem",2) == "by")
L("by-np") = 1;
dqaninfo(3,4,5,6);
groupnp();
group(1,4,"_advl"); # prep np -> advl
N("pp",2) = 1;
if (L("by-np"))
N("by-np",2) = 1;
@RULES
_xNIL <-
\,
_prep
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
\,
_xALPHA
_xWILD [one lookahead match=(_prep _fnword _xEND)]
@@
# vg prep dqan conj alpha* vg
@PRE
<9,9> var("noun");
@CHECK
if (!N("adj",8) && !N("noun",8))
fail();
if (!plural(N(9)))
fail();
@POST
fixnphead(9);
fixnpnonhead(8);
group(7,7,"_fnword"); # Keep nps from glomming.
@RULES
_xNIL <-
_vg ### (1)
_prep ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_conj ### (7)
_xALPHA ### (8)
_xALPHA ### (9)
_vg ### (10)
@@
# np vg dqan alpha
# vg dqan alpha
@PRE
<3,3> varz("pos50 n-v-dqan-alpha-p");
@POST
N("pos50 n-v-dqan-alpha-p",3) = 1;
if (N("adj",9) && verbfeat(N(3),"X7"))
alphatovg(9,"active","VB");
else if (N("noun",9) && N("verb",9))
{
if (vconjq(N(9),"-en")) # (ambig verb...)
alphatovg(9,"passive","VBN");
else if (verbfeat(N(3),"V2") && vconjq(N(9),"inf"))
alphatovg(9,"active","VB");
else if (vconjq(N(9),"inf") || vconjq(N(9),"-s"))
fixnphead(9);
}
else if (N("adj",9) && N("adv",9))
{
if (!N("noun",0) && !N("verb",9))
alphatoadv(9);
}
@RULES
_xNIL <-
_np [opt] ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_vg ### (3)
_adv [star] ### (4)
_xWILD [star match=(_det _pro)] ### (5)
_xWILD [star match=(_quan _num _xNUM)] ### (6)
_adj [star] ### (7)
_noun [plus] ### (8)
_xALPHA ### (9)
_xWILD [one lookahead match=(_prep _conj)] ### (10)
@@
# ^ following np by
@POST
chpos(N(2),"IN"); # following/IN [NO Conform Treebank]
group(2,2,"_fnword");
@RULES
_xNIL <-
_xSTART
following
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
by [s lookahead]
@@
# ^ np , noun conj np , alpha
# Note: Apposition.
@CHECK
if (N("verb",8) && !N("adj",8))
fail();
@POST
group(4,6,"_np");
group(2,5,"_np");
@RULES
_xNIL <-
_xSTART ### (1)
_np ### (2)
\, ### (3)
_xWILD [one match=(_noun _np)] ### (4)
_conj ### (5)
_xWILD [one match=(_noun _np)] ### (6)
\, [lookahead] ### (7)
_xALPHA ### (8)
@@
# ^ np alpha
@PRE
<3,3> var("verb");
@POST
alphatovg(3,"active","VBP");
@RULES
_xNIL <-
_xSTART
_np
_xALPHA
@@
# ^ np prep dqan alpha
@PRE
<9,9> var("verb");
@CHECK
if (literal(N(10)))
{
if (!N("noun",10) && !N("adj"))
fail();
}
@POST
alphatovg(9,"active","VBP");
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_xWILD [one match=(_noun _np _det)] ### (3)
_prep ### (4)
_xWILD [plus match=(_det _pro)] ### (5)
_xWILD [star match=(_quan _num _xNUM)] ### (6)
_adj [star] ### (7)
_noun [plus] ### (8)
_xALPHA ### (9)
_xWILD [one lookahead match=(_xALPHA _det _quan _num _xNUM
_adj _noun _np)]
@@
# vg np conj np vg prep
@PRE
<6,6> varz("voice");
@CHECK
if (!vconjq(N(6),"-en"))
fail();
if (!vconjq(N(1),"-ing"))
fail();
# Want "ellipted verb", no auxiliaries on it.
L("n") = pndown(N(1));
if (pnname(L("n")) != "_verb")
fail();
L("n") = pndown(L("n"));
if (!literal(L("n")))
fail();
@POST
fixvg(N(6),"passive","VBN");
N("pos50 psv25",6) = 1;
@RULES
_xNIL <-
_vg ### (1)
_np [lookahead] ### (2)
_conj ### (3)
_np ### (4)
_xWILD [star match=(_adv _advl)] ### (5)
_vg ### (6)
_xWILD [one match=(_prep)] ### (7)
@@
# vg np conj vg
@PRE
<4,4> varz("pos50 vn-conj-v");
@CHECK
if (!vgassigned(N(1)) || vgassigned(N(4)))
fail();
@POST
N("pos50 vn-conj-v",4) = 1;
if (N("compound-vg",1))
L("vg") = N("last vg",1);
else
L("vg") = N(1);
L("pos") = vgpos(L("vg"));
L("voice") = pnvar(L("vg"),"voice");
fixvg(N(4),L("voice"),L("pos"));
@RULES
_xNIL <-
_vg
_np
_conj
_vg
@@
# , np vg ,
@POST
group(4,4,"_clausesep");
group(1,1,"_clausesep");
@RULES
_xNIL <-
\,
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
\,
@@
# ^ np , clause , vg
@PRE
<8,8> varz("voice");
@POST
fixvg(N(8),"active","VBD");
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_adv _advl)]
_np
_clausesep
_xWILD [star match=(_noun _np _verb _vg _adv _advl _prep)]
_clausesep
_xWILD [star match=(_adv _advl)]
_vg
@@
# vg by np as np vg
@CHECK
if (vgassigned(N(6))) # New fn. # 05/27/07 AM.
fail();
@POST
fixvg(N(6),"passive","VBN");
N("pos50 psv26",6) = 1;
N("glom",6) = "right";
N("glom",5) = "left";
if (pnname(N(4)) == "_fnword" && !N("renamed",4))
{
pnrename(N(4),"_prep"); # as -> prep
N("renamed",4) = 1;
}
@RULES
_xNIL <-
_vg
by [s]
_np
as [s]
_np [lookahead]
_vg
@@
# vg prep np by np
@PRE
<2,2> varz("pos50 zzz");
@POST
N("zzz",2) = 1;
L("num") = phrprepverbq(N(2),N(3));
if (L("num") == 1 || L("num") == 3)
{
if (N("stem",3) != "to")
chpos(N(3),"RP"); # prep/RP.
groupvgprep(2,3);
}
else if (L("num") == 2)
{
if (N("stem",3) != "to")
chpos(N(3),"IN"); # prep/IN.
groupvgprep(2,3);
}
else
{
if (!vgassigned(N(2)))
{
if ((N("stem",3) == "by" && semactornode(N(4)))
|| (N("stem",5) == "by" && semactornode(N(6))) )
{
if (vconjq(N(2),"-en"))
fixvg(N(2),"passive","VBN");
else
fixvg(N(2),"active","VBP");
}
}
prepnptoadvl(3,4); # prep np -> advl
}
@RULES
_xNIL <-
_xWILD [plus match=( \, _conj _np)]
_vg
_prep
_np
_prep
_xWILD [one match=(_noun _np)]
@@
# vg by np
@PRE
<1,1> varz("pos50 psvx1");
<1,1> varz("voice");
<3,3> vareq("stem","by");
@CHECK
if (!semactornode(N(5)))
fail();
@POST
N("pos50 psvx1",1) = 1;
fixvg(N(1),"passive","VBN");
@RULES
_xNIL <-
_vg
_xWILD [star lookahead match=(_adv _advl)]
_prep
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_np)]
@@
# prep noun , conj vg
@POST
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_noun
_xWILD [plus lookahead match=( \, _conj)]
_vg
@@
# np , as alpha ,
@PRE
<4,4> var("verb");
@CHECK
if (!vconjq(N(4),"-en"))
fail();
@POST
alphatovg(4,"passive","VBN");
N("pos50 psv27",4) = 1;
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
\,
as [s]
_xALPHA
\, [lookahead]
@@
# vg that noun
@CHECK
if (!plural(N(3)) && N("mypos",2) != "IN")
fail();
@POST
dqaninfo(0,0,0,3);
groupnp();
@RULES
_xNIL <-
_vg
that [s]
_noun
_xWILD [one lookahead match=(_prep)]
@@
# that np prep np and np
@POST
group(3,4,"_advl");
N("pp",3) = 1;
group(2,5,"_np"); # List of nps.
@RULES
_xNIL <-
that [s]
_np
_prep
_np
_conj
_np
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_verb _vg)]
@@
# alpha dqan alpha
@PRE
<1,1> var("verb");
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"inf"))
fail();
if (!verbfeat(N(1),"V2")) # eg, let x go
fail();
@POST
alphatovg(6,"active","VB");
alphatovg(1,"active",0);
@RULES
_xNIL <-
_xALPHA ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_xALPHA ### (6)
@@
# vg np before vg np or alpha
@PRE
<7,7> var("verb");
@CHECK
S("vc") = samevconj(N(1),N(7));
if (S("vc"))
{
L("v") = N("verb node",1);
S("pos") = pnvar(L("v"),"mypos");
succeed();
}
S("vc") = samevconj(N(4),N(7));
if (S("vc"))
{
L("v") = N("verb node",4);
S("pos") = pnvar(L("v"),"mypos");
succeed();
}
fail();
@POST
alphatovg(7,0,S("pos"));
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
before [s] ### (3)
_vg ### (4)
_np ### (5)
or [s] ### (6)
_xALPHA ### (7)
_xWILD [one lookahead match=(_np)] ### (8)
@@
# of alpha np
# of alpha dqan
@PRE
<2,2> var("verb");
@CHECK
if (!vconjq(N(2),"-ing"))
fail();
@POST
alphatovg(2,"active","VBG");
@RULES
_xNIL <-
of [s]
_xALPHA
_xWILD [one lookahead match=(_np _det _pro _quan _num _xNUM)]
@@
# Note: no x, no y = implicit there is.
@PRE
<2,2> varz("pos50 np-,-np");
<2,2> var("neg");
@POST
N("pos50 np-,-np",2) = 1;
X("implicit-there-is") = 1;
@RULES
_xNIL <-
_xSTART
_np
\, [lookahead]
_np
@@
# np vg prep prep np
@PRE
<1,1> var("neg");
<2,2> varz("voice");
@CHECK
if (!X("implicit-there-is"))
fail();
if (!vconjq(N(2),"-en"))
fail();
@POST
fixvg(N(2),"passive","VBN");
N("pos50 psv28",2) = 1;
@RULES
_xNIL <-
_np
_vg
_prep
_prep
_np
@@
# np , np conj np , vg
@POST
group(2,7,"_np");
N("apposition",2) = 1;
N("pos50 n,n&n,v",2) = 1;
@RULES
_xNIL <-
_xWILD [one match=( \, )] ### (1)
_np ### (2)
\, ### (3)
_xWILD [one match=(_noun _np)] ### (4)
_conj ### (5)
_xWILD [one match=(_noun _np)] ### (6)
\, ### (7)
_adv [star lookahead] ### (8)
_vg ### (9)
@@
# , dqan alpha
@PRE
<3,3> varz("prosubj");
<3,3> varz("proobj");
<7,7> var("noun");
@CHECK
L("n") = lasteltnode(6);
if (plural(L("n")))
fail();
if (N(3))
L("first") = N(3);
else if (N(4))
L("first") = N(4);
else if (N(5))
L("first") = N(5);
else
L("first") = N(6);
if (!numbersagree(L("first"),N(7)))
fail();
if (plural(N(7)) && !N(4))
fail();
if (N("verb",7))
{
if (vconjq(N(7),"-ed") || vconjq(N(7),"-ing"))
fail();
}
@POST
fixnphead(7);
@RULES
_xNIL <-
\, ### (1)
_xWILD [star match=(_adv _advl)] ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_xALPHA ### (7)
_xWILD [one lookahead fail=(_xALPHA _noun _np
_det _quan _num _xNUM _adj)] ### (8)
@@
# np that np vg
# Note: A specialized subjunctive pattern.
# decree that John eat his peas.
@CHECK
if (vgassigned(N(4)))
fail();
L("t") = nodestem(N(1));
if (!finddictattr(L("t"),"that-subjunctive"))
fail();
if (!vconjq(N(4),"inf"))
fail();
@POST
fixvg(N(4),"active","VB"); # Subjunctive.
N("that-subjunctive",1) = 1;
N("that-subjunctive",2) = 1;
N("that-subjunctive",4) = 1;
if (!N("mypos",2))
chpos(N(2),"IN"); # that/IN
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
that [s]
_np
_vg
@@
# det noun , adj noun
# det cap , adj noun
@PRE
<2,2> var("cap");
@POST
pnrename(N(2),"_adj"); # noun -> adj
group(2,4,"_adj"); # adj list
@RULES
_xNIL <-
_xWILD [s one match=(a an the)]
_noun
\,
_adj
_noun [lookahead]
@@
# vg pro alpha alpha
# vg her alpha alpha [vg np alpha alpha]
# her
# Note: 'her' is ambiguous PP/PP$
@PRE
<2,2> vareq("stem","her");
<3,3> var("verb");
<3,3> varz("pos50 vg-her-aa");
@CHECK
if (!vconjq(N(3),"-ing"))
fail();
@POST
N("pos50 vg-her-aa",3) = 1; # Loop guard.
if (!N("noun",3) && N("adj",3) && (N("noun",4) || N("adj",4)))
{
fixnpnonhead(3);
}
else if (N("noun",3) && N("adj",4))
{
# Still ambig, but may be a predicate adjective.
alphatoadj(4);
group(4,4,"_adjc");
N("predicate-adj",4) = 1;
fixnphead(3);
chpos(N(2),"PP$"); # her/PP$
dqaninfo(2,0,0,3);
groupnp();
}
else if (verbfeat(N(1),"V4"))
{
alphatovg(3,"active","VBG");
if (!N("mypos",2))
chpos(N(2),"PP"); # her/PP.
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_pro
_xALPHA
_xALPHA
@@
# her
# Note: 'her' is ambiguous PP/PP$
@PRE
<2,2> vareq("stem","her");
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"-ing"))
fail();
if (!verbfeat(N(1),"V4"))
fail();
@POST
alphatovg(3,"active","VBG");
if (!N("mypos",2))
chpos(N(2),"PP"); # her/PP.
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_pro
_xALPHA
@@
# adj to alpha
# adjc to alpha
@PRE
<2,2> vareq("stem","to");
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"inf"))
fail();
@POST
alphatovg(3,"active","VB");
@RULES
_xNIL <-
_xWILD [one match=(_adjc _adj)]
_prep
_xALPHA
@@
# verb much alpha $
# vg much alpha $
@PRE
<3,3> var("adv");
@CHECK
L("t") = nodestem(N(3));
if (advconj(L("t")) != "RBR")
fail();
@POST
alphatoadv(3);
alphatoadv(2);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(much)]
_xALPHA
_xEND
@@
# verb by
# vg by
@PRE
<2,2> varne("mypos","RB");
<2,2> varz("by-adv");
@CHECK
if (pnname(N(1)) == "_vg")
{
if (!vgassigned(N(1)))
fail();
if (N("voice",1) == "passive")
fail();
}
else if (!N("mypos",1) || N("mypos",1) == "VBN")
fail();
if (phrprepverbq(N(1),N(2)))
fail();
@POST
N("by-adv",2) = 1;
chpos(N(2),"RB"); # by/RB
pnrename(N(2),"_adv"); # by -> adv. prep -> adv.
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
by [s]
_xWILD [one lookahead fail=(_noun _np _xALPHA _det _quan _num
_xNUM _adj)]
@@
# vg to alpha noun (that)
# Note: Decide or gather evidence for noun/verb.
@PRE
<2,2> vareq("stem","to");
<3,3> varz("pos50 vtan-that");
#<5,5> vareq("stem","that");
@POST
N("pos50 vtan-that",3) = 1;
if (!N("noun",3) && !N("verb",3))
{
if (N("adj",3))
alphatoadj(3);
else if (N("adv",3))
alphatoadv(3);
}
else if (!N("verb",3))
fixnpnonhead(3);
else if (!N("noun",3))
{
# patterns: stooped to eating garbage....
# verb adj adv
if (vconjq(N(3),"-en") || vconjq(N(3),"-s"))
fixnpnonhead(3);
else if (vconjq(N(3),"-ing"))
alphatovg(3,"active","VBG");
else # inf
{
# [DEFAULT] for now. Can do more logic here.
if (!verbfeat(N(1),"T3") && !verbfeat(N(3),"D5"))
fixnpnonhead(3);
else
alphatovg(3,"active","VB");
}
}
else # both noun and verb possible.
{
# noun verb (adj adv)
if (vconjq(N(3),"-en") || vconjq(N(3),"-s"))
fixnpnonhead(3);
else if (vconjq(N(3),"-ing"))
alphatovg(3,"active","VBG");
else # inf
{
if (!verbfeat(N(1),"T3") && !verbfeat(N(3),"D5"))
# fixnpnonhead(3); # Need more context.
;
else
alphatovg(3,"active","VB");
}
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_prep ### (2)
_xALPHA ### (3)
_noun [lookahead] ### (4)
# _fnword ### (5)
@@
# verb to alpha
# be to alpha
# have to alpha
@PRE
<3,3> var("verb");
<3,3> varz("noun");
@CHECK
if (N("stem",1) != "be" && N("stem",1) != "have")
fail();
if (!vconjq(N(3),"inf"))
fail();
@POST
alphatoverb(3,"active","VB");
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
_prep
_xALPHA
@@
# vg np dqan conj alpha
# verb noun dqan conj alpha
@PRE
<8,8> varz("pos50 vn-dqan-conj-a");
@POST
N("pos50 vn-dqan-conj-a",8) = 1;
if (!N("noun",8) && !N("verb",8))
{
if (N("adv",8))
alphatoadv(8);
else if (N("adj",8))
alphatoadj(8);
}
else if (!N("verb",8))
fixnphead(8);
else if (!N("noun",8))
{
if (samevconj(N(1),N(8)))
{
if (pnname(N(1)) == "_vg")
L("v") = N("verb node",8);
else
L("v") = N(8);
if (L("v"))
L("m") = pnvar(L("v"),"mypos");
alphatovg(8,L("voice"),L("m"));
}
else
alphatovg(8,0,0);
}
else # Both verb and noun.
{
if (samevconj(N(1),N(8)))
{
if (pnname(N(1)) == "_vg")
L("v") = N("verb node",1);
else
L("v") = N(1);
if (L("v"))
L("m") = pnvar(L("v"),"mypos");
alphatovg(8,L("voice"),L("m"));
}
else
alphatovg(8,0,0);
}
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_xWILD [one match=(_noun _np)] ### (2)
_xWILD [plus match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
_conj ### (7)
_xALPHA ### (8)
_xWILD [one match=(_pro)] ### (9)
@@
# dqa alpha+ prep which
# Note: back-end patterns...
@CHECK
if (!N(1) && !N(2) && !N(3))
fail();
if (!N("adj",4) && !N("noun",4))
fail();
@POST
L("n") = eltnode(5);
if (L("n"))
{
if (pnvar(L("n"),"noun") || pnvar(L("n"),"adj"))
{
alphatoadj(4);
}
else
fixnphead(4);
}
else
fixnphead(4);
@RULES
_xNIL <-
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA [star]
_prep [lookahead]
_whword [s]
@@
# vg np who alpha noun conj alpha noun
@PRE
<3,3> vareq("stem","who"); # who. # whom.
<4,4> varz("pos50 vnpro-an-conj-an");
<4,4> var("verb");
<8,8> var("verb");
@CHECK
S("vc") = samevconj(N(4),N(8));
if (!S("vc"))
fail();
if (S("vc") == "-en")
fail();
@POST
N("pos50 vnpro-an-conj-an",4) = 1;
if (S("vc") == "inf")
{
alphatovg(8,"active","VBP");
alphatovg(4,"active","VBP");
}
else
{
alphatovg(8,"active","VBD");
alphatovg(4,"active","VBD");
}
@RULES
_xNIL <-
_vg ### (1)
_np ### (2)
_fnword ### (3)
_xALPHA ### (4)
_xALPHA [star] ### (5)
_noun [plus] ### (6)
_conj ### (7)
_xALPHA ### (8)
_xALPHA [star lookahead] ### (9)
_noun [plus] ### (10)
_xWILD [one match=( \, _clausesep)] ### (11)
_xWILD [one match=(_fnword _pro _np _det _quan _num _xNUM _adj)] ### (12)
@@
##### VERBLESS
@CHECK
# Agreement...
L("arr") = vgagree(N(1),N(3),N(5),N(7),0);
# "err.txt" << L("arr") << phrasetext() << "\n";
if (!L("arr")[0])
fail();
S("voice") = "active";
@POST
L("m") = N(1);
L("h") = N(3);
L("b") = N(5);
L("being") = N(7);
if (N(7))
{
N("sem",7) = N("stem",7) = "be";
chpos(N(7),"VBG");
}
L("neg") = mhbvadv(2,4,6,0);
group(1,7,"_vg");
mhbv(N(1),L("neg"),L("m"),L("h"),L("b"),L("being"),0);
N("voice",1) = S("voice");
@RULES
_xNIL <-
_xWILD [s one match=(_modal _do) except=(_vg _np)]
_xWILD [star match=(_adv _advl)]
_have [s opt]
_xWILD [star match=(_adv _advl)]
_be [s opt]
_xWILD [star match=(_adv _advl)]
being [s opt]
_xWILD [one lookahead fail=(_xALPHA _adv)]
@@
@CHECK
# Agreement...
L("arr") = vgagree(0,N(1),N(3),N(5),0);
S("verb node") = N(1);
if (!L("arr")[0])
fail();
# S("voice") = "active";
@POST
L("h") = N(1);
L("b") = N(3);
L("being") = N(5);
if (N(5))
{
N("sem",5) = N("stem",5) = "be";
chpos(N(5),"VBG");
}
L("neg") = mhbvadv(2,4,0,0);
if (N(5))
group(1,5,"_vg");
else if (N(3))
group(1,3,"_vg");
else
group(1,1,"_vg");
mhbv(N(1),L("neg"),0,L("h"),L("b"),L("being"),0);
@RULES
_xNIL <-
_xWILD [s one match=(_have) except=(_vg)]
_xWILD [star match=(_adv _advl)]
_be [s opt]
_xWILD [star match=(_adv _advl)]
being [s opt]
@@
@CHECK
# Agreement...
L("arr") = vgagree(0,0,N(1),N(3),0);
if (!L("arr")[0])
fail();
S("voice") = "active";
if (N(3))
{
N("sem",3) = N("stem",3) = "be";
chpos(N(3),"VBG");
}
@RULES
_vg <-
_be [s]
_xWILD [star match=(_adv _advl)]
being [s]
@@
# be
@POST
L("tmp1") = N(1);
L("neg") = mhbvadv(2,0,0,0);
group(1,1,"_vg");
mhbv(N(1),L("neg"),0,0,L("tmp1"),0,0);
pncopyvars(L("tmp1"),N(1));
N("voice",1) = "active";
clearpos(N(1),1,0);
@RULES
_xNIL <-
_be [s except=(_vg)]
_adv [star]
_xWILD [one fail=(_xALPHA _verb)]
@@
_xNIL <-
_be [s except=(_vg)]
_adv [star]
_xEND
@@
# noun alpha to verb
# np alpha to vg
@PRE
<2,2> var("verb");
@CHECK
if (!verbfeat(N(2),"T3"))
fail();
@POST
alphatoverb(2,0,"VBP");
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro)]
_xALPHA
to [s lookahead]
_xWILD [one match=(_verb _vg _modal)]
@@
# noun , dqan , conj alpha
@POST
if (N("adj",9) && N("noun",9))
{
if (singular(N(9)) && !N("mass",9))
fixnpnonhead(9);
}
else if (N("adj",9))
alphatoadj(9);
else if (N("noun",9))
alphatonoun(9);
dqaninfo(3,4,5,6);
groupnp();
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _pro)] ### (1)
\, ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_noun [plus] ### (6)
\, [opt] ### (7)
_conj ### (8)
_xALPHA ### (9)
@@
# alpha verb
@PRE
<1,1> var("noun");
@CHECK
if (!plural(N(1)))
fail();
@POST
fixnphead(1);
@RULES
_xNIL <-
_xALPHA
_xWILD [one lookahead match=(_verb _vg)]
@@
# vg prep dqa alpha
@PRE
<6,6> var("noun");
@CHECK
if (!N(3) && !N(4) && !N(5))
fail();
@POST
if (!N("mypos",2))
{
L("rp") = phrprepverbq(N(1),N(2));
if (L("rp") == 1 || L("rp") == 3)
chpos(N(2),"RP"); # prep/RP.
else
chpos(N(2),"IN"); # prep/IN.
}
fixnphead(6);
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)] ### (1)
_prep ### (2)
_xWILD [star match=(_det _pro)] ### (3)
_xWILD [star match=(_quan _num _xNUM)] ### (4)
_adj [star] ### (5)
_xALPHA ### (6)
_xWILD [star lookahead match=(_adv _advl)] ### (7)
_xWILD [one match=(_prep _fnword)] ### (8)
@@
# vg prep dqan
@PRE
<1,1> varz("pos50 v-p-check");
<2,2> varz("mypos");
@CHECK
S("num") = phrprepverbq(N(1),N(2));
if (!S("num"))
fail();
@POST
N("pos50 v-p-check",1) = 1;
if (S("num") == 1 || S("num") == 3)
chpos(N(2),"RP"); # prep/RP
else
chpos(N(2),"IN"); # prep/IN
@RULES
_xNIL <-
_vg
_prep
_xWILD [one match=(_det _pro _quan _num _xNUM _adj _noun _np)]
@@
# prep np
# vg prep np
@PRE
<3,3> varne("stem","after");
<3,3> varne("stem","by"); # Trying to keep these out. # 05/25/07 AM.
<3,3> varne("mypos","RP"); # Attached to verb.
@POST
if (N("stem",3) == "by")
L("by") = 1;
L("tmp4") = N(4);
group(3,4,"_advl"); # prep np -> advl
pncopyvars(L("tmp4"),N(3));
clearpos(N(3),1,0);
N("pp",3) = 1;
if (L("by"))
{
if (N("sem",3) != "date" && N("sem",3) != "geoloc"
&& N("sem",3) != "quantity")
N("by-actor",3) = 3;
N("by-np",3) = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(_pro _np _vg)]
_xWILD [star match=(_adv _advl)]
_prep
_np
_xWILD [one lookahead fail=( \, _conj _prep)]
@@
#####################
# DEALING WITH DOUBLE QUOTES. #
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
@POST
fixnpnonhead(2);
group(1,4,"_np");
@RULES
_xNIL <-
_dblquote
_xALPHA
_noun
_dblquote
@@
# Zap double quotes.
@POST
excise(1,1);
@RULES
_xNIL <-
_dblquote
@@
|
# Add numeric value number as a string to concept concept's attribute named nameString
@CODE
G("Malibu") = makeconcept(findroot(), "Malibu");
addsval(G("Malibu"),"Route",1+2); |
@NODES _LINE
@RULES
_CompleteSchoolName <- _university @@
@POST
singler(3,3)
@RULES
_CompleteSchoolName [base] <-
The [s]
_xWHITE [s star]
_SchoolNamePhrase [t]
@@
_CompleteSchoolName [base] <-
_xSTART
_xWHITE [s star]
_SchoolNamePhrase [t]
@@
@RULES
_CompleteSchoolName [base] <-
_PostalState [s]
_xWHITE [s]
_SchoolNamePhrase [t]
@@
_CompleteSchoolName [base] <-
_SchoolNamePhrase [s]
_xWHITE [s star]
of [s]
_xWHITE [s star]
_CityName [s]
@@
|
@CODE
prlit("ctc.txt", "\n CONTACT ZONE LINES DUMP\n");
prlit("ctc.txt", "----------------------------------\n");
@@CODE
@PATH _ROOT _contactZone
# @INI
# Dump contact zone node's data once.
# xdump()
# @@INI
@POST
ndump("ctc.txt",1);
prlit("ctc.txt", "-------\n");
@RULES
_xNIL <- _LINE @@
|
@CODE
# Iterate over note_words, getting codes associated with
# tagged euis
L("iter") = down(G("note_words"));
while (L("iter")) {
L("euis") = findvals(L("iter"), "bases");
if (L("euis")) {
addattr(L("iter"), "codes");
}
# Iterate over euis, adding to section codes attribute
L("eui_iter") = L("euis");
while (L("eui_iter")) {
L("codes") = GetCodes(getstrval(L("eui_iter")));
# Add list of codes to pred_codes and as section attr
if (L("codes")) {
AddVals(L("iter"), "codes", L("codes"));
while (L("codes")) {
AddUniqueCon(G("pred_codes"), getstrval(L("codes")));
L("codes") = nextval(L("codes"));
}
}
L("eui_iter") = nextval(L("eui_iter"));
}
# Advance iterator
L("iter") = next(L("iter"));
}
L("iter") = down(G("note_words"));
while (L("iter")) {
L("codes") = findvals(L("iter"), "codes");
L("euis") = findvals(L("iter"), "bases");
L("note_euis") = StrValsToArray(L("euis"));
while (L("codes")) {
L("code_str") = getstrval(L("codes"));
L("code_rank") = rankCode(L("code_str"), L("note_euis"));
L("code_con") = getconcept(G("pred_codes"), L("code_str"));
if (L("code_con")) {
addstrval(L("code_con"), conceptname(L("iter")), str(L("code_rank")));
}
L("codes") = nextval(L("codes"));
}
# Advance iterator
L("iter") = next(L("iter"));
}
@@CODE |
@NODES _term
@POST
excise(1,1);
noop();
@RULES
_xNIL <-
_xWILD [one matches=(_xNUM _xPUNCT)]
@@
@PRE
<1,1> var("stopword");
@POST
excise(1,1);
noop();
@RULES
_xNIL <-
_xANY
@@ |
@PATH _ROOT _edge
@POST
excise(2,2);
single();
@RULES
_child <-
_xWILD [fails=(\t)] ### (1)
\t
@@
|
# Todo: DROP NLP++ regions into their own segments, to
# be able to say @PATH _ROOT _NLPPP
@NODES _NLPPP
@POST
# movesem(1) #
makestmts(1) # Renamed. # 05/29/01 AM.
single()
@RULES
# NEED THE BASE, OR INFINITE LOOP!
_STMTS [base] <- _xWILD [s one match=(_STMT _EXPR
_BLOCK # 12/15/99 AM.
)] @@
@POST
addstmt(1, 2)
single()
@RULES
_STMTS [base] <- _STMTS _xWILD [s one match=(_STMT _EXPR
_BLOCK # 12/15/99 AM.
)] @@
@POST
movesem(2)
single()
@RULES
_BLOCK <- \{ [s] _STMTS \} [s] @@
# EMPTY BLOCK. #
@RULES
_BLOCK <- \{ [s] \} [s] @@ # 08/01/00 AM.
@POST
ifstmt(1, 2)
single()
@RULES
_IFSTMT <-
_IFPART
_xWILD [s one match=(_BLOCK _STMT _EXPR)]
@@
# WHILE STATEMENT #
@POST
movesem(2)
single()
@RULES
_WHILECOND <- _WHILE _EXPR @@
# Should make sure expr is parenthesized.
@POST
whilestmt(1, 2)
single()
@RULES
_STMT <- _WHILECOND _xWILD [s one match=(_BLOCK _STMT
_EXPR # 05/29/01 AM.
)]
@@
@POST
# movesem(2) #
makestmt(2) # 05/11/00 AM.
single()
@RULES
# Need better action to handle _NUM, _STR.
_ELSEPART <-
_ELSE
_xWILD [s one match=(_BLOCK _STMT _EXPR)]
@@
@POST
ifelse(1, 2)
single()
@RULES
_STMT <-
_IFSTMT _ELSEPART
@@
# Probably need to check for no "else" as next elt.
@POST
movesem(1)
singler(1, 1)
@RULES
_STMT [base] <- _IFSTMT
_xWILD [s one match=( _xANY _xEND _xEOF ) except=( _ELSE )]
@@
|
@PATH _ROOT _equation
@POST
X("type",2) = N("label type");
X("name",2) = N("label name");
addEquationName(N("label name"));
@RULES
_xNIL <-
_label [s]
@@
|
@NODES _LINE
@PRE
<1,1> cap();
<4,4> cap();
@RULES
# Ex: N.\_H.
_statePhrase [layer=(_Caps )] <- N [s] \. [s] _xWHITE [star s] H [s] \. [s] @@
# Ex: N.\_M.
_statePhrase [layer=(_Caps )] <- N [s] \. [s] _xWHITE [star s] M [s] \. [s] @@
# Ex: N.\_J.
_statePhrase [layer=(_Caps )] <- N [s] \. [s] _xWHITE [star s] J [s] \. [s] @@
# Ex: N.\_D.
_statePhrase [layer=(_Caps )] <- N [s] \. [s] _xWHITE [star s] D [s] \. [s] @@
# Ex: N.\_C.
_statePhrase [layer=(_Caps )] <- N [s] \. [s] _xWHITE [star s] C [s] \. [s] @@
# Ex: R.\_I.
_statePhrase [layer=(_Caps )] <- R [s] \. [s] _xWHITE [star s] I [s] \. [s] @@
# Ex: S.\_C.
_statePhrase [layer=(_Caps )] <- S [s] \. [s] _xWHITE [star s] C [s] \. [s] @@
# Ex: S.\_D.
_statePhrase [layer=(_Caps )] <- S [s] \. [s] _xWHITE [star s] D [s] \. [s] @@
# Ex: W.\_V.
_statePhrase [layer=(_Caps )] <- W [s] \. [s] _xWHITE [star s] V [s] \. [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: Cal.
_statePhrase [layer=(_Caps )] <- Cal [s] \. [s] @@
# Ex: Calif.
_statePhrase [layer=(_Caps )] <- Calif [s] \. [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: New\_Hampshire
_statePhrase [layer=(_Caps )] <- New [s] _xWHITE [star s] Hampshire [s] @@
# Ex: New\_Mexico
_statePhrase [layer=(_Caps )] <- New [s] _xWHITE [star s] Mexico [s] @@
# Ex: New\_Jersey
_statePhrase [layer=(_Caps )] <- New [s] _xWHITE [star s] Jersey [s] @@
# Ex: North\_Dakota
_statePhrase [layer=(_Caps )] <- North [s] _xWHITE [star s] Dakota [s] @@
# Ex: North\_Carolina
_statePhrase [layer=(_Caps )] <- North [s] _xWHITE [star s] Carolina [s] @@
# Ex: Puerto\_Rico
_statePhrase [layer=(_Caps )] <- Puerto [s] _xWHITE [star s] Rico [s] @@
# Ex: South\_Dakota
_statePhrase [layer=(_Caps )] <- South [s] _xWHITE [star s] Dakota [s] @@
# Ex: South\_Carolina
_statePhrase [layer=(_Caps )] <- South [s] _xWHITE [star s] Carolina [s] @@
# Ex: West\_Virginia
_statePhrase [layer=(_Caps )] <- West [s] _xWHITE [star s] Virginia [s] @@
# Ex: Rhode\_Island
_statePhrase [layer=(_Caps )] <- Rhode [s] _xWHITE [star s] Island [s] @@
|
# Find string attribute's value up the hierarchy. con is the current concept, topCon is the root to be searched up to, and name is the name of the attribute. If hier==0, goes to root of KB.
L("return_str") = inheritval(L("con"), L("name"), L("topCon")); |
@CODE
G("enclosed by") = "\"";
G("separator") = ",";
G("lineTerminator") = "\n";
G("escaped by") = "\"";
@@CODE
@NODES _ROOT
@RULES
_enclosedBy <-
\" ### (1)
@@
_separator <-
\, ### (1)
@@
_lineTerminator <-
\r [optional] ### (1)
\n ### (2)
@@
_escapedBy <-
\\ ### (1)
@@
|
@CODE
G("people") = getconcept(findroot(),"people");
rmchildren(G("people"));
@@CODE |
@NODES _ROOT
@POST
excise(1,1);
@RULES
_xNIL <-
_rest ### (1)
@@
|
@CODE
L("hello") = 0;
@@CODE
@NODES _sent
# another ugly tiptoeing
# adj verb
# Sometimes it's a verb, sometimes ving noun.
@CHECK
L("n") = pndown(N(2));
if (!vconjq(L("n"),"-ing"))
fail();
@POST
pnrename(N(2),"_noun"); # verb -> noun
posacct(N(2));
@RULES
_xNIL <-
_adj
_verb
@@
# prep noun alpha prep
@CHECK
if (!N("adj",3))
fail();
if (N("verb",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
@RULES
_xNIL <-
_prep
_xWILD [one match=(_noun _np)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_prep)]
@@
# prep noun
@POST
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_prep)]
_noun
_xWILD [one lookahead match=(_qEOS _xEND)]
@@
# prep adv alpha noun
@CHECK
if (!N("adj",3) && !N("noun",3))
fail();
@POST
L("tmp3") = N(3);
if (N("adj",3))
group(3,3,"_adj");
else
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
if (pnname(N(3)) == "_adj")
fixadj(N(3));
@RULES
_xNIL <-
_xWILD [s one match=(_prep) except=(to)]
_adv [opt]
_xALPHA
_xWILD [one lookahead match=(_noun)]
@@
# adj conj adj
@POST
L("tmp4") = N(4);
group(2,4,"_adj");
pncopyvars(L("tmp4"),N(2));
clearpos(N(2),1,0); # Zero out token info.
@RULES
_xNIL <-
_xWILD [one match=(_xALPHA)]
_adj
_conj
_adj
@@
# how much power he
# how much alpha pro
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
nountonp(2,1);
@RULES
_xNIL <-
_howmuch [s]
_xALPHA
_xWILD [one lookahead match=(_pro)]
@@
# , det noun alpha ,
@CHECK
if (!N("noun",4))
fail();
# Todo: Check for noun-verb DISagreement.
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
@RULES
_xNIL <-
_xWILD [one match=( \, )]
_det
_noun
_xALPHA
_xWILD [one lookahead match=( \, )]
@@
# alpha noun verb
@CHECK
if (!N("verb",1))
fail();
if (N("pos num",2) != 2)
fail();
if (!N("adj",1) && !N("noun",1))
fail();
@POST
L("tmp1") = N(1);
if (N("noun",1))
group(1,1,"_noun");
else
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
if (pnname(N(1)) == "_adj")
fixadj(N(1));
@RULES
_xNIL <-
_xALPHA
_noun [lookahead]
_xWILD [one match=(_verb _vg)]
@@
# det noun '
@POST
chpos(N(3),"POS"); # Possessive.
group(2,3,"_adj");
@RULES
_xNIL <-
_xWILD [one match=(_det _quan _num _xNUM _adj)]
_noun
_aposX
@@
# once verb
@POST
pnrename(N(1),"_adv");
@RULES
_xNIL <-
once [s]
_xWILD [one lookahead match=(_verb _vg)]
@@
# Some pos tagging.
# noun verb
@POST
if (vconjq(N(2),"inf"))
chpos(N(2),"VBP");
@RULES
_xNIL <-
_xWILD [one match=(_noun _quan _num _xNUM)]
_verb
@@
# that
@CHECK
if (N("voice",1))
fail();
# if (N("mypos",2))
# fail();
@POST
# chpos(N(2),"IN"); # that/IN
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
that [s]
@@
# that
#@CHECK
# if (N("mypos",2))
# fail();
#@POST
# chpos(N(2),"WDT");
# N("bracket",2) = 1;
#@RULES
#_xNIL <-
# _xWILD [one match=(_noun _adj _np)]
# that [s]
# @@
# more adv
# more adj
@POST
chpos(N(1),"RBR");
pnrename(N(1),"_adv");
@RULES
_xNIL <-
_xWILD [s one match=(more less)]
_xWILD [one lookahead match=(_adv _advl _adj)]
@@
# verb alpha
# foreshadow rising
#@CHECK
# if (!N("adj",2))
# fail();
# if (!N("adv",2) && !N("verb",2))
# fail();
# S("vc2") = vconj(N(2));
# if (S("vc2") == "inf"
# || S("vc2") == "-s")
# fail();
#@POST
# L("tmp2") = N(2);
# group(2,2,"_adj");
# pncopyvars(L("tmp2"),N(2));
# fixadj(N(2));
#@RULES
#_xNIL <-
# _xWILD [s one match=(_verb _vg) except=(_modal _have _be being)]
# _xALPHA
# @@
# noun verb alpha
@CHECK
if (!N("verb",7))
fail();
if (!N("noun",7))
fail(); # For now, just handle n-v ambig.
@POST
L("tmp7") = N(7);
group(7,7,"_noun");
pncopyvars(L("tmp7"),N(7));
fixnoun(N(7));
if (pnname(N(5)) == "_vg")
if (!N("voice",5))
N("voice",5) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun _np)]
_xWILD [star match=(_adv _advl)]
_xWILD [s one match=(_verb _vg) except=(_modal _have _be)]
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# adj alpha that
# that
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
# chpos(N(3),"WDT"); # ...
# N("bracket",3) = 1;
@RULES
_xNIL <-
_adj
_xALPHA
that [s]
@@
|
@PATH _ROOT _equation _split
@RULES
_separator <-
\\ ### (1)
\\ ### (2)
NEWLINE ### (3)
@@
@POST
excise(1,1);
single();
@RULES
_equal <-
\& ### (1)
_xWHITE [optional] ### (2)
\= ### (3)
@@ |
@NODES _ROOT
@POST
S("sent") = MakeCountCon(G("parse"),"sentence");
addstrval(S("sent"),"text",N("$text",1));
"sents.txt" << "'" << N("$text",1) << "'\n";
single();
@RULES
_sentence [unsealed] <-
_xWILD [s plus fails=(\. \? \! _paragraphSeparator)] ### (1)
_xWILD [s one matches=(\. \? \!)] ### (2)
@@
_sentence [unsealed] <-
_xWILD [s plus fails=(_paragraphSeparator)] ### (1)
@@
|
@CODE
L("hello") = 0;
@@CODE
@NODES _TEXTZONE
@POST
S("sem") = S("ne type") = "organization";
S("ne type conf") = 95;
S("stem") = phrasetext();
S("ne") = 1;
single();
@RULES
_caps [layer=_noun] <-
united [s]
_xWHITE [star]
nations [s]
@@
_caps [layer=_noun] <-
tamil [s]
_xWHITE [star]
_xANY
@@
@PRE
<1,1> cap();
@POST
pncopyvars();
S("sem") = S("ne type") = "organization";
S("ne type conf") = 95;
S("stem") = phrasetext();
S("ne") = 1;
S("number") = "singular";
single();
@RULES
_caps [layer=_noun] <- _xWILD [s one match=(
un unicef bloomberg reuters
bbc nbc cnn
)] @@
@POST
S("sem") = S("ne type") = "organization";
S("ne type conf") = 95;
S("stem") = "EU";
S("ne") = 1;
single();
@RULES
_caps [layer=_noun] <-
eu
@@
# US States
# n.h.
@POST
S("ne") = 1;
S("ne type") = "location";
S("ne type conf") = 85;
S("sem") = "us_state";
S("stem") = phrasetext();
S("mypos") = "NP";
if (N(5))
S("pos eos") = 1;
S("cap") = 1;
single();
@RULES
_usstate [layer=_noun] <-
n
\.
_xWHITE [star]
h
\. [opt]
@@
# US States
@PRE
<1,1> cap();
@POST
S("sem") = "us_state";
S("ne type") = "location";
S("ne type conf") = 85;
S("ne") = 1;
pncopyvars(1);
S("stem") = phrasetext();
S("mypos") = "NP";
if (N(2))
S("poss eos") = 1;
single();
@RULES
_usstate [layer=_noun] <-
_xWILD [s one match=(
AA
AE
AK
AL
Ala
Alb
Alabama
#Alas
Alaska
AP
AR
Ari
Ariz
Arizona
#Ark
Arkan
Arkansas
#AS
AZ
CA
Cal
Calif
California
#CO
Col
Colo
Colorado
Columbia
Conn
Connecticut
CT
Dak
Dakota
DC
#DE
Del
Dela
Delaware
FL
Fla
Flo
Flor
Florida
#FM
GA
Geo
Georg
Georgia
GU
Guam
Haw
Hawaii
#HI
IA
#ID
Ida
Idaho
#IL
#Ill
Illin
Illinois
#IN
Ind
Indiana
Iowa
Kan
Kans
Kansas
#Ken
#Kent
Kentucky
KS
KY
#LA
#Louis
Louisiana
#MA
Maine
#Mary
Maryland
#Mass
Massachusetts
#MD
#ME
MH
MI
Mich
Michigan
Micronesia
Minn
Minnesota
#Miss
Mississippi
Missouri
MN
MO
Mon
Mont
Montana
MP
#MS
MT
NC
ND
NE
Neb
Nebraska
Nev
Nevada
NH
NJ
NM
NSW
NV
NY
#OH
Ohio
#OK
Okl
Okla
Oklahoma
#OR
#Ore
Oreg
Oregon
Orpalau
#PA
Penn
Pennsyl
Pennsylvania
PR
PW
RI
Samoa
SC
SD
Tenn
Tennessee
Tex
Texas
TN
TX
UT
Utah
VA
Ver
Verm
Vermont
#VI
Vir
Virg
Virgin
Virginia
VT
WA
Wash
Washington
WI
Wis
Wisc
Wiscon
Wisconsin
WV
WY
Wyo
Wyoming
)]
\. [opt]
@@
# Hard-wiring some plural proper names. #
@PRE
<1,1> cap();
@POST
N("mypos") = "NPS"; # cap/NPS
@RULES
_xNIL <-
_xWILD [s one match=(
americans
beatles
communists
germans
jews
russians
soviets
virginians
)]
@@
# Hard-wiring some development stuff so I don't keep
# stumbling across it as an error. #
# composer name name
# jobtype cap cap
@POST
if (N("cap",1))
{
N("cap",1) = 0;
setbase(1,"true"); # Don't look for the capitalization here.
}
N("mypos",2) = N("mypos",3) = "NP";
group(2,3,"_noun");
N("ne type",2) = "person";
N("ne type conf",2) = 95;
N("ne",2) = 1;
N("sem",2) = "person";
@RULES
_xNIL <-
_xWILD [s one match=(composer)]
_xCAP [s]
_xCAP [s]
_xWILD [s one lookahead fail=(_xCAP)]
@@
|
@CODE
G("line_count") = 0;
@@CODE
|
@CODE
G("bookkeeping") = findconcept(findroot(),"bookkeeping");
if (!G("bookkeeping")) G("bookkeeping") = makeconcept(findroot(),"bookkeeping");
G("cites") = findconcept(G("bookkeeping"),"cites");
if (!G("cites")) G("cites") = makeconcept(G("bookkeeping"),"cites");
rmchildren(G("cites"));
G("figures") = findconcept(G("bookkeeping"),"figures");
if (!G("figures")) G("figures") = makeconcept(G("bookkeeping"),"figures");
rmchildren(G("figures"));
G("equations") = findconcept(G("bookkeeping"),"equations");
if (!G("equations")) G("equations") = makeconcept(G("bookkeeping"),"equations");
rmchildren(G("equations"));
G("references") = findconcept(G("bookkeeping"),"references");
if (!G("references")) G("references") = makeconcept(G("bookkeeping"),"references");
rmchildren(G("references"));
G("bibliography") = findconcept(G("bookkeeping"),"bibliography");
if (!G("bibliography")) G("bibliography") = makeconcept(G("bookkeeping"),"bibliography");
rmchildren(G("bibliography"));
@@CODE
|
@NODES _LINE
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: IBM\_RS\_
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] RS [s] _xWHITE [star s] 6000 [s] @@
# Ex: IBM\_PC\_compatibles
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] PC [s] _xWHITE [star s] compatibles [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: PDP\_11\_
_hardware [layer=(_Caps )] <- PDP [s] _xWHITE [star s] 11 [s] _xWHITE [star s] @@
# Ex: IBM\_
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] 370 [s] _xWHITE [star s] @@
_hardware [layer=(_Caps )] <- HP [s] _xWHITE [star s] 9000 [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Dec\_Alpha\_
_hardware [layer=(_Caps )] <- Dec [s] _xWHITE [star s] Alpha [s] _xWHITE [star s] @@
@RULES
# Ex: logic\_analyzer\_
_hardware [layer=(_Caps )] <- logic [s] _xWHITE [star s] analyzer [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Protocol\_Analyzer\_
_hardware [layer=(_Caps )] <- Protocol [s] _xWHITE [star s] Analyzer [s] _xWHITE [star s] @@
# Ex: Alpha\_AXP\_
_hardware [layer=(_Caps )] <- Alpha [s] _xWHITE [star s] AXP [s] _xWHITE [star s] @@
# Ex: PROM\_Burner\_
_hardware [layer=(_Caps )] <- PROM [s] _xWHITE [star s] Burner [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
@RULES
# Ex: PCI\_bus\_
_hardware [layer=(_Caps )] <- PCI [s] _xWHITE [star s] bus [s] _xWHITE [star s] @@
@RULES
# Ex: hardware\_design\_
_hardware [layer=(_Caps )] <- hardware [s] _xWHITE [star s] design [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
<3,3> cap();
<5,5> cap();
@RULES
# Ex: In-Circuit\_Emulator\_
_hardware [layer=(_Caps )] <- In [s] \- [s] Circuit [s] _xWHITE [star s] Emulator [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Silicon\_Graphics\_
_hardware [layer=(_Caps )] <- Silicon [s] _xWHITE [star s] Graphics [s] _xWHITE [star s] @@
# Ex: Apple\_II\_
_hardware [layer=(_Caps )] <- Apple [s] _xWHITE [star s] II [s] _xWHITE [star s] @@
# Ex: IBM\_PC\_
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] PC [s] _xWHITE [star s] @@
# Ex: Pentium\_PC's
_hardware [layer=(_Caps )] <- Pentium [s] _xWHITE [star s] PC [s] \' [s] s [s] @@
# Ex: DEC\_PDP-8
_hardware [layer=(_Caps )] <- DEC [s] _xWHITE [star s] PDP [s] \- [s] 8 [s] @@
# Ex: Sun\_Workstation
_hardware [layer=(_Caps )] <- Sun [s] _xWHITE [star s] Workstation [s] @@
@RULES
_hardware [layer=(_Caps )] <- 486 [s] _xWHITE [star s] @@
@PRE
<3,3> cap();
@RULES
# Ex: 10-BaseT\_
_hardware [layer=(_Caps )] <- 10 [s] \- [s] BaseT [s] _xWHITE [star s] @@
@PRE
<1,1> cap();
@RULES
# Ex: AMD\_
_hardware [layer=(_Caps )] <- AMD [s] _xWHITE [star s] @@
# Ex: Apollo\_
_hardware [layer=(_Caps )] <- Apollo [s] _xWHITE [star s] @@
# Ex: Apple\_
_hardware [layer=(_Caps )] <- Apple [s] _xWHITE [star s] @@
_hardware [layer=(_Caps )] <- AS [s] \/ [s] 400 [s] _xWHITE [star s] @@
# Ex: ATM\_
_hardware [layer=(_Caps )] <- ATM [s] _xWHITE [star s] @@
# Ex: AXP\_
_hardware [layer=(_Caps )] <- AXP [s] _xWHITE [star s] @@
# Ex: Banyan\_
_hardware [layer=(_Caps )] <- Banyan [s] _xWHITE [star s] @@
# Ex: Bull\_
_hardware [layer=(_Caps )] <- Bull [s] _xWHITE [star s] @@
# Ex: BURROUGHS\_
_hardware [layer=(_Caps )] <- BURROUGHS [s] _xWHITE [star s] @@
# Ex: CDPD\_
_hardware [layer=(_Caps )] <- CDPD [s] _xWHITE [star s] @@
# Ex: CRAY\_
_hardware [layer=(_Caps )] <- CRAY [s] _xWHITE [star s] @@
# Ex: DASD
_hardware [layer=(_Caps )] <- DASD [s] @@
# Ex: DEC
_hardware [layer=(_Caps )] <- DEC [s] @@
# Ex: Ethernet
_hardware [layer=(_Caps )] <- Ethernet [s] @@
# Ex: FDDI
_hardware [layer=(_Caps )] <- FDDI [s] @@
# Ex: Gateways
_hardware [layer=(_Caps )] <- Gateways [s] @@
# Ex: HP
_hardware [layer=(_Caps )] <- HP [s] @@
# Ex: Hubs
_hardware [layer=(_Caps )] <- Hubs [s] @@
# Ex: IBM
_hardware [layer=(_Caps )] <- IBM [s] @@
# Ex: Intel
_hardware [layer=(_Caps )] <- Intel [s] @@
# Ex: Intergraph
_hardware [layer=(_Caps )] <- Intergraph [s] @@
# Ex: LAN
_hardware [layer=(_Caps )] <- LAN [s] @@
# Ex: Mac
_hardware [layer=(_Caps )] <- Mac [s] @@
# Ex: Macintosh
_hardware [layer=(_Caps )] <- Macintosh [s] @@
@RULES
# Ex: mainframe
_hardware [layer=(_Caps )] <- mainframe [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: MIPS
_hardware [layer=(_Caps )] <- MIPS [s] @@
# Ex: Motorola
_hardware [layer=(_Caps )] <- Motorola [s] @@
# Ex: NCR
_hardware [layer=(_Caps )] <- NCR [s] @@
# Ex: NEC
_hardware [layer=(_Caps )] <- NEC [s] @@
# Ex: NeXT
_hardware [layer=(_Caps )] <- NeXT [s] @@
# Ex: OmniSQL
_hardware [layer=(_Caps )] <- OmniSQL [s] @@
@RULES
# Ex: oscilloscope
_hardware [layer=(_Caps )] <- oscilloscope [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: PC
_hardware [layer=(_Caps )] <- PC [s] @@
# Ex: PCI
_hardware [layer=(_Caps )] <- PCI [s] @@
# Ex: PCs
_hardware [layer=(_Caps )] <- PCs [s] @@
# Ex: PDP-11
_hardware [layer=(_Caps )] <- PDP [s] \- [s] 11 [s] @@
# Ex: Pentium/
_hardware [layer=(_Caps )] <- Pentium [s] \/ [s] 486 [s] @@
# Ex: PowerPC
_hardware [layer=(_Caps )] <- PowerPC [s] @@
# Ex: Routers
_hardware [layer=(_Caps )] <- Routers [s] @@
_hardware [layer=(_Caps )] <- RS [s] \/ [s] 6000 [s] @@
_hardware [layer=(_Caps )] <- RS [s] 232 [s] @@
_hardware [layer=(_Caps )] <- RS [s] \- [s] 232 [s] @@
_hardware [layer=(_Caps )] <- RS [s] \- [s] 422 [s] @@
_hardware [layer=(_Caps )] <- RS [s] \- [s] 485 [s] @@
_hardware [layer=(_Caps )] <- RS [s] 6000 [s] @@
# Ex: Sparcstation
_hardware [layer=(_Caps )] <- Sparcstation [s] @@
# Ex: Spectrometer
_hardware [layer=(_Caps )] <- Spectrometer [s] @@
# Ex: Symbolics
_hardware [layer=(_Caps )] <- Symbolics [s] @@
# Ex: Unisys
_hardware [layer=(_Caps )] <- Unisys [s] @@
# Ex: USB
_hardware [layer=(_Caps )] <- USB [s] @@
# Ex: Vax
_hardware [layer=(_Caps )] <- Vax [s] @@
# Ex: VAX
_hardware [layer=(_Caps )] <- VAX [s] @@
# Ex: Workstations
_hardware [layer=(_Caps )] <- Workstations [s] @@
@RULES
# Ex: x86
_hardware [layer=(_Caps )] <- x [s] 86 [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: PC-AT
_hardware [layer=(_Caps )] <- PC [s] \- [s] AT [s] @@
# Ex: PC-EGA
_hardware [layer=(_Caps )] <- PC [s] \- [s] EGA [s] @@
# Ex: IBM\_PC/x86
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] PC [s] \/ [s] x [s] 86 [s] @@
@PRE
<2,2> cap();
@RULES
_hardware [layer=(_Caps )] <- 320 [s] C [s] 240 [s] @@
_hardware [layer=(_Caps )] <- 68 [s] HC [s] 812 [s] @@
@PRE
<1,1> cap();
@RULES
_hardware [layer=(_Caps )] <- HP [s] 730 [s] @@
# Ex: DEC6xxx
_hardware [layer=(_Caps )] <- DEC [s] 6 [s] xxx [s] @@
_hardware [layer=(_Caps )] <- S [s] \/ [s] 390 [s] @@
# Ex: IBM\_
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] 3090 [s] @@
# Ex: IBM\_
_hardware [layer=(_Caps )] <- IBM [s] _xWHITE [star s] 4380 [s] @@
@RULES
# Ex: hubs
_hardware [layer=(_Caps )] <- hubs [s] @@
# Ex: switches
_hardware [layer=(_Caps )] <- switches [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: SCSI
_hardware [layer=(_Caps )] <- SCSI [s] @@
# Ex: T1
_hardware [layer=(_Caps )] <- T [s] 1 [s] @@
# Ex: CSU
_hardware [layer=(_Caps )] <- CSU [s] @@
@RULES
# Ex: router
_hardware [layer=(_Caps )] <- router [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: RFL
_hardware [layer=(_Caps )] <- RFL [s] @@
# Ex: Radian
_hardware [layer=(_Caps )] <- Radian [s] @@
# Ex: Z-80
_hardware [layer=(_Caps )] <- Z [s] \- [s] 80 [s] @@
@RULES
# Ex: routers
_hardware [layer=(_Caps )] <- routers [s] @@
@PRE
<1,1> cap();
@RULES
_hardware [layer=(_Caps )] <- PDP [s] \- [s] 11 [s] \/ [s] 84 [s] @@
_hardware [layer=(_Caps )] <- PDP [s] \- [s] 11 [s] \/ [s] 34 [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: MicroVax\_II
_hardware [layer=(_Caps )] <- MicroVax [s] _xWHITE [star s] II [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: UNISYS\_
_hardware [layer=(_Caps )] <- UNISYS [s] _xWHITE [star s] 2200 [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Smart\_Card
_hardware [layer=(_Caps )] <- Smart [s] _xWHITE [star s] Card [s] @@
# Ex: Pentium-II
_hardware [layer=(_Caps )] <- Pentium [s] \- [s] II [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: Pentium
_hardware [layer=(_Caps )] <- Pentium [s] @@
# Ex: Laptops
_hardware [layer=(_Caps )] <- Laptops [s] @@
# Ex: PS/2
_hardware [layer=(_Caps )] <- PS [s] \/ [s] 2 [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: PC/AT
_hardware [layer=(_Caps )] <- PC [s] \/ [s] AT [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: VAX\_
_hardware [layer=(_Caps )] <- VAX [s] _xWHITE [star s] 11 [s] \/ [s] 780 [s] @@
# Ex: Apollo
_hardware [layer=(_Caps )] <- Apollo [s] @@
# Ex: Amiga
_hardware [layer=(_Caps )] <- Amiga [s] @@
# Ex: Tandem
_hardware [layer=(_Caps )] <- Tandem [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Intel-PC
_hardware [layer=(_Caps )] <- Intel [s] \- [s] PC [s] @@
# Ex: Power\_PC
_hardware [layer=(_Caps )] <- Power [s] _xWHITE [star s] PC [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: PDAs
_hardware [layer=(_Caps )] <- PDAs [s] @@
# Ex: PCMCIA
_hardware [layer=(_Caps )] <- PCMCIA [s] @@
# Ex: OCR
_hardware [layer=(_Caps )] <- OCR [s] @@
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: CD-ROM
_hardware [layer=(_Caps )] <- CD [s] \- [s] ROM [s] @@
@PRE
<1,1> cap();
@RULES
# Ex: RISC
_hardware [layer=(_Caps )] <- RISC [s] @@
# Ex: SGI
_hardware [layer=(_Caps )] <- SGI [s] @@
# Ex: PentiumPro
_hardware [layer=(_Caps )] <- PentiumPro [s] @@
# Ex: T3
_hardware [layer=(_Caps )] <- T [s] 3 [s] @@
# Ex: X.25
_hardware [layer=(_Caps )] <- X [s] \. [s] 25 [s] @@
# Ex: Burroughs
_hardware [layer=(_Caps )] <- Burroughs [s] @@
# Ex: Convergent
_hardware [layer=(_Caps )] <- Convergent [s] @@
# Ex: Wang
_hardware [layer=(_Caps )] <- Wang [s] @@
# Ex: Altos
_hardware [layer=(_Caps )] <- Altos [s] @@
# Ex: Honeywell
_hardware [layer=(_Caps )] <- Honeywell [s] @@
@PRE
<3,3> cap();
@RULES
# Ex: 3COM\_LAN
_hardware [layer=(_Caps )] <- _companyPhrase [s] _xWHITE [star s] LAN [trig s] @@
@PRE
<1,1> cap();
@RULES
# Ex: LCD\_graphics\_display
_hardware [layer=(_Caps )] <- LCD [s] _xWHITE [star s] graphics [s] _xWHITE [star s] display [s] @@
# Ex: Flatbed\_image\_scanner
_hardware [layer=(_Caps )] <- Flatbed [s] _xWHITE [star s] image [s] _xWHITE [star s] scanner [s] @@
# Ex: Motherboard
_hardware [layer=(_Caps )] <- Motherboard [s] @@
# Ex: ASIC
_hardware [layer=(_Caps )] <- ASIC [s] @@
# Ex: ISDN
_hardware [layer=(_Caps )] <- ISDN [s] @@
# Ex: Dialup
_hardware [layer=(_Caps )] <- Dialup [s] @@
# Ex: PRI
_hardware [layer=(_Caps )] <- PRI [s] @@
# Ex: BRI
_hardware [layer=(_Caps )] <- BRI [s] @@
# Ex: Merced
_hardware [layer=(_Caps )] <- Merced [s] @@
# Ex: Amdahl
_hardware [layer=(_Caps )] <- Amdahl [s] @@
# Ex: Pyramid
_hardware [layer=(_Caps )] <- Pyramid [s] @@
|
@CODE
G("kb") = getconcept(findroot(),"kb");
SaveKB("mykb.kbb",G("kb"),2);
@@CODE |
@CODE
fileout("exp_anchors1.txt"); # 06/10/00 AM.
@@CODE
@PATH _ROOT _experienceZone
@CHECK
#Ngt(1, "nanchors", 0) # If it's an anchor line.
if (N("nanchors",1) <= 0) fail();
@POST
"exp_anchors1.txt" << "-------------------" << "\n";
ndump("exp_anchors1.txt",1);
# Transfer data up to anchor!
S("lineno") = N("lineno", 1);
single();
@RULES
_ANCHOR <- _LINE [s] @@
|
@NODES _LINE
@PRE
<1,1> cap();
<5,5> cap();
@RULES
# Ex: Activities\_and\_Honors
_OtherHeaderPhrase [layer=(_header )] <- Activities [s] _xWHITE [star s] and [s] _xWHITE [star s] Honors [s] @@
|
@DECL
########
# FUNC: LOADATTR
# SUBJ: Load attr into dictionary concept.
# RET:
# NOTE: Updates dictionary in kb.
########
loadattr(
L("entry"),
L("attr_s"),
L("val")
)
{
if (!L("entry") || !L("attr_s"))
return;
L("lc") = strtolower(L("entry"));
L("dict concept") = dictgetword(L("lc"));
if (!L("dict concept"))
return;
replaceval(L("dict concept"),L("attr_s"),L("val"));
}
@CODE
# One-time load for this analyzer instance.
if (numval(findroot(),"geo-kb"))
exitpass();
else
replaceval(findroot(),"geo-kb",1); # Mark as loaded now.
# As an example.
# Thai attr(country) => Thailand
loadattr("Bangladeshi","country","Bangladesh");
loadattr("Burma","country","Myanmar");
loadattr("Burmese","country","Myanmar");
loadattr("Cambodian","country","Cambodia");
loadattr("Chinese","country","China");
loadattr("Indian","country","India");
loadattr("Indonesian","country","Indonesia");
loadattr("Japanese","country","Japan");
loadattr("Laotian","country","Laos");
loadattr("Malaysian","country","Malaysia");
loadattr("Myanmaran","country","Myanmar");
loadattr("Singaporan","country","Singapore");
loadattr("Singaporean","country","Singapore");
loadattr("Soviet","country","Soviet Union");
loadattr("Sri Lankan","country","Sri Lanka");
loadattr("Thai","country","Thailand");
loadattr("Viet","country","Vietnam");
loadattr("Vietnamese","country","Vietnam");
@@CODE
|
@CODE
if (!G("pretagged"))
exitpass();
@@CODE
@NODES _ROOT
@POST
excise(1,1);
@RULES
_xNIL <- _xWILD [one match=(
_COMMENT
)] @@
@POST
splice(1,1);
@RULES
_xNIL <-
_LINE
@@
|
@NODES _ROOT
@POST
singler(2,2);
@RULES
_endSent <-
_xWILD [s one match=(_xALPHA _xNUM)] ### (1)
_xWILD [one trig match=(\. \? \!)] ### (2)
_xWILD [one match=(_xWHITE _xALPHA _xNUM _xEND _xCTRL)] ### (3)
@@
|
@NODES _paragraph
@POST
excise(1,1);
noop();
@RULES
_xNIL <-
_xWHITE [s] ### (1)
@@
|
@CODE
G("corporate") = findconcept(findroot(),"corporate");
if (!G("corporate"))
G("corporate") = makeconcept(findroot(),"corporate");
G("parse") = findconcept(G("corporate"),"parse");
if (G("parse"))
rmchildren(G("parse"));
else
G("parse") = makeconcept(G("corporate"),"parse");
G("companies") = findconcept(G("corporate"),"companies");
G("section number") = 0;
# IF COMPANIES NOT IN KB, ADD THEM HERE. #
if (G("companies"))
exitpass();
G("companies") = getconcept(G("corporate"),"companies");
L("co") = getconcept(G("companies"),"TAI");
getconcept(L("co"),"Text Analysis International");
L("co") = getconcept(G("companies"),"AMRP");
getconcept(L("co"),"American Medical Records Processing");
@@CODE |
# put attribute and value in the KB
if(findconcept(findroot(),"apple"))
rmconcept(findconcept(findroot(),"apple"));
G("apple") = makeconcept(findroot(), "apple");
addstrval(G("apple"),"color","red");
# access
"output.txt" << "getstrval = " <<
getstrval(findvals(G("apple"), "color")) << "\n";
"output.txt" << "strval = " <<
strval(G("apple"),"color") << "\n";
"output.txt" << "attrwithval = " <<
attrwithval(G("apple"), "color", "red") << "\n";
# change
attrchange(G("apple"), "color", "red", "blue")
"output.txt" << "changing apples' color from red to blue\n" << "attrwithval = " <<
attrwithval(G("apple"), "color", "blue") << "\n";
# remove attrs, vals and concept
rmattrval(G("apple"), "color", "red");
rmconcept(G("apple"); |
@NODES _ROOT
@POST
if (N("type") == "figure")
{
S("image") = N("image");
S("label") = N("label");
S("caption") = N("caption");
}
else
{
S("text") = N("text");
S("title") = N("title");
}
S("type") = N("type");
single();
@RULES
_item <-
_xWILD [one match=(_abstract _sectionBlock _figure)] ### (1)
@@
|
@NODES _LINE
# NOTE: Hand building some generalized email handling. #
@POST
singler(3,5);
@RULES
_email <- _emailHeader [s]
_xWILD [s star match=(\ \t \n _whtSEP)]
_xWILD [s fails=(\ \t \n _whtSEP _emailHeader \@)]
\@ [s trig]
_xWILD [s fails=(\ \t \n _whtSEP)]
_xWILD [s star match=(\ \t \n _whtSEP)]
@@
@POST
singler(3,5);
@RULES
_email <- _emailHeader [s]
_xWILD [s star match=(_xWHITE _whtSEP)]
_xWILD [s fails=(_xWHITE _whtSEP _emailHeader \@)]
\@ [s trig]
_xWILD [s fails=(_xWHITE _whtSEP)]
_xWILD [s star match=(_xWHITE _whtSEP)]
@@
# (_xwhite in match list not implemented yet.)
@POST
singler(2,4);
@RULES
_email <-
# Trying to make this match the interp rule below. #
_xWILD [s star match=( \ \t \n _whtSEP _whtIndent)]
_xWILD [s plus fails=( \ \t \n _whtSEP _whtIndent \, \[ \] _WebLinks)]
\@ [s trigger]
_xWILD [s plus fails=( \ \t \n _whtSEP \, \[ \] _Weblinks)]
_xWILD [s star match=( \ \t \n _whtSEP)]
@@
@POST
singler(2,4);
@RULES
_email <-
# _xWILD [s star match=(_xWHITE _whtSEP _whtIndent)]
_xWILD [s plus fails=(_xWHITE _whtSEP _whtIndent \, \[ \] _WebLinks)]
\@ [s trigger]
_xWILD [s plus fails=(_xWHITE _whtSEP \, \[ \] _Weblinks)]
_xWILD [s star match=(_xWHITE _whtSEP)]
@@
|
@DECL
# Add code to word
# code: alphanumeric code
# type: procedure or diagnosis
# word: word to add code to
addCodeToWord(L("code"), L("type"), L("word")) {
L("word_concept") = AddUniqueCon(G("word2codes"),L("name"));
}
# Add word to code
#
@@DECL |
@PATH _ROOT _defZone _definition _LINE
@RULES
_xNIL <-
_item ### (1)
@@
|
@CODE
L("hello") = 0;
@@CODE
@NODES _ROOT
# First paragraph.
@PRE
<3,3> varz("PAR");
@RULES
_TEXTZONE [unsealed] <-
_xSTART
_LINE
_LINE [star]
@@
@PRE
<1,1> var("PAR");
<2,2> varz("PAR");
@RULES
_TEXTZONE [unsealed] <-
_LINE
_LINE [star]
@@
# Excise blank lines.
@POST
excise(1,1);
@RULES
_xNIL <-
_xWILD [plus match=(_BLANKLINE _xWHITE)]
@@
|
@CODE
G("persist") = findconcept(findroot(),"persist");
if (!G("persist")) G("persist") = makeconcept(findroot(),"persist");
G("counter") = IncrementCount(G("persist"),"count");
G("file") = openfile(FileNameNumbered("outfile",5));
@@CODE |
# Remove value val from attribute attr.
rmval(L("attr"), L("val")); |
@PATH _ROOT _paragraph _sentence
@RULES
_officer <-
CEO [s] ### (1)
_conj [s] ### (2)
president [s] ### (3)
@@
_moneyType <-
assumed [s] ### (1)
debt [s] ### (2)
@@
_commentor <-
analysts [s] ### (1)
predictions [s] ### (2)
@@ |
@CODE
DictionaryStart();
@@CODE
|
@NODES _sentence
@PRE
<1,1> cap();
<2,2> cap();
@RULES
# Ex: Los\_Angeles
_city <- Los [s] Angeles [s] @@
|
@CODE
L("hello") = 0;
@@CODE
#@PATH _ROOT _TEXTZONE _sent
#@NODES _sent
|
@MULTI _ROOT _section _looseText _sentence _subsection _item
@PRE
<1,1> var("bases");
@POST
if (strcontains(",", N("bases", 1))) {
L("base_list") = split(N("bases", 1), ",");
L("base_lens") = 0;
if (!L("parent_keys")) {
pnmakevar(X(),"bases", L("child_keys")[L("i")]);
L("i")++;
}
while (L("i") < L("keys_len")) {
pnrpushval(X(), "bases", L("child_keys")[L("i")]);
L("i")++;
}
pnreplaceval(N(1), "bases", 0);
}
@RULES
_xNIL <-
_xANY
# FOR RECURSIVE IMPLEMENTATION, REMOVE THIS
# Stop list includes non-leaf nodes and common medical terms
# We exclude _subsection since this may give us relevant title
# _xWILD [one fails=(
# _xEND
# _section
# _subsection
# _sentence
# _item
# patient
# )]
@@
# @PRE
# <1,1> var("bases");
# @POST
# if (N("section_title", 1)) {
# pnrpushval(N(1), "bases", N("section_title", 1));
# }
# if (N("subsection", 1)) {
# pnrpushval(N(1), "bases", N("subsection", 1));
# }
# if (X("section_title")) {
# pnrpushval(N(1), "bases", X("section_title"));
# }
# if (X("subsection")) {
# pnrpushval(N(1), "bases", X("subsection"));
# }
# @RULES
# _xNIL <-
# _xWILD [one matches=(
# _section
# _sentence
# _subsection
# _item
# _looseText
# )]
# @@
|
@NODES _ROOT
@POST
excise(1,1);
@RULES
_xNIL <-
_BLANKLINE [s] ### (1)
@@
|
@CODE
exitpass(); # IGNORE THIS PASS! # 09/02/01 AM.
prlit("dump.txt", "Before sort\n");
"dump.txt" << G("Words");
# fprintvar("dump.txt", "Words", " ");
prlit("dump.txt", "\n\n");
sortvals("Words"); # Sort alphabetically.
prlit("dump.txt", "After sort\n");
"dump.txt" << G("Words");
prlit("dump.txt", "\n\n");
gtolower("Words");
prlit("dump.txt", "Lowercased\n");
"dump.txt" << G("Words");
prlit("dump.txt", "\n\n");
guniq("Words");
prlit("dump.txt", "Uniqued\n");
"dump.txt" << G("Words");
prlit("dump.txt", "\n\n");
# Lookup words in dict file.
lookup("Words", # The variable.
"doom.words", # The word file.
"word"); # Bit-flag to update in symbol table entry.
@@CODE
#@PATH _ROOT _LINE
#@PRE
#<1,1> unknown() # Reduce if alphabetic is not a plain English word.
#@RULES
# _UNK <- _xALPHA [s] @@
# Should always fail!
#_xNIL <- zzzzzzxxyyA @@
|
@CODE
# Count the total number of cityStateZip matches.
# For use in determining best address block.
#var("cityStateZip", "0")
G("cityStateZip") = 0;
@@CODE
# Scour each line for all its addressParts and count them.
@MULTI _LINE
@POST
++X("addressParts");
@RULES
_xNIL <- _addressPart @@
# Record instances of city,state,zip pattern. #
@POST
++G("cityStateZip");
@RULES
_xNIL <- _cityStateZip @@
|
@PATH _ROOT _CODE _NLPPP
@POST
# rfacode(1) #
rfaactions(1) # 01/14/00 AM.
single()
@RULES
#_CODE [base] <- _xWILD [plus match=(_ACTION)] @@
_CODE [base] <- _STMTS [plus] @@ # 01/14/00 AM.
@POST
noop()
@RULES
_xNIL <- _xWILD [one match=(
_xPUNCT
_WHILECOND
)] @@
@POST
rfberror(1)
single()
@RULES
_ERROR <- _xANY [plus] @@
|
@NODES _ROOT
@RULES
_defZone <-
_definition ### (1)
_example ### (2)
@@
|
@NODES _LINE
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Customer\_Service
_jobPhrase [layer=(_Caps )] <- Customer [s] _xWHITE [star s] Service [s] @@
# Ex: Customer\_Support
_jobPhrase [layer=(_Caps )] <- Customer [s] _xWHITE [star s] Support [s] @@
|
# Find mother in the kb, then print parent concept name to output.txt, then go back to mother concept and print, then print daughter concept name, then print daughter's previous sibling name.
# The output should look like:'\nfemale\nmother\ndaughter\nmother'
@CODE
# Find 'mother' in the KB
G("First") = findhierconcept("mother", findroot());
# goto parent concept (female)
G("Second") = up(G("First"));
"output.txt" << conceptname(G("Second")) << "\n";
# go back down to 'mother'
G("First") = down(G("Second"));
"output.txt" << conceptname(G("First")) << "\n";
# find mother's first sibling, which should be 'daughter'
G("Second") = next(G("First"));
"output.txt" << conceptname(G("Second")) << "\n";
# find daughter's previous sibling, which should be mother
G("First") = prev(G("Second"));
"output.txt" << conceptname(G("First")) << "\n";
|
@PATH _ROOT _contactZone _LINE _humanName
# Just some samples here. These could be done by a RUG hierarchy!!
# If name is unlabeled, meaning first middle, last haven't been set.
@PRE
<2,2> lengthr(2,20) # Title must be length > 1.
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("prefixName") = N("$text", 2);
X("firstName") = N("$text", 5);
X("middleName") = N("$text", 8);
X("lastName") = N("$text", 11);
X("suffixName") = N("$text", 14);
# noop()
@RULES
# Dr. John A. Blipper [, Jr.]
_xNIL <-
_xSTART
_xALPHA [s layer=(_prefixName)]
\. [s]
_xWHITE [s star]
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_middleName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@PRE
<2,2> lengthr(2,20) # Title must be length > 1.
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("prefixName") = N("$text", 2);
X("firstName") = N("$text", 5);
X("lastName") = N("$text", 8);
X("suffixName") = N("$text", 11);
# noop()
@RULES
# Dr. John Blipper [, Jr.]
_xNIL <-
_xSTART
_xALPHA [s layer=(_prefixName)]
\. [s]
_xWHITE [s star]
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("firstName") = N("$text", 2);
X("middleName") = N("$text", 5);
X("lastName") = N("$text", 8);
X("suffixName") = N("$text", 11);
# noop()
@RULES
_xNIL <-
_xSTART
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_middleName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("firstName") = N("$text", 2);
X("lastName") = N("$text", 5);
X("suffixName") = N("$text", 8);
# noop();
@RULES
_xNIL <-
_xSTART
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
|
@PATH _ROOT _translations _headerZone _LINE
@POST
L("con") = MakeCountCon(X("language",3),"translation");
addstrval(L("con"),"text",N("$text",1));
@RULES
_xNIL <-
_translation ### (2)
@@
|
@NODES _ROOT
@RULES
_attrBlock <-
_startAttr ### (1)
_xWILD [fail=(_endAttr)]
_endAttr
@@
|
#@NODES _REZZONE
@PATH _ROOT _REZZONE
@POST
excise(1,1)
@RULES
_xNIL <- _BLANKLINE @@
|
@NODES _ROOT
@POST
excise(1,1);
@RULES
_xNIL <-
_rest ### (1)
@@
|
@PATH _ROOT _paragraph _sentence
@POST
S("value") = N("value",1);
S("asset") = N("$text",2);
single();
@RULES
_money <-
_money ### (1)
_asset ### (2)
@@
@POST
S("value") = num(N("$text",1)) * num(N("numeric",2));
single();
@RULES
_money <-
_xNUM ### (1)
_number ### (2)
_currency ### (3)
@@
@POST
S("value") = N("value",1);
single();
@RULES
_money <-
_money ### (1)
_number ### (2)
_currency ### (3)
@@
|
@PATH _ROOT _RULES
@RECURSE listarg
@POST
rfaarg(1)
single()
@RULES
_ARG [base] <- _NONLIT @@
_ARG [base] <- _LIT @@
_ARG [base] <- _STR @@
_ARG [base] <- _NUM @@
@@RECURSE listarg
@RECURSE argtolist
@POST
rfaargtolist(1)
single()
@RULES
_LIST <- _ARG @@
@@RECURSE argtolist
@POST
rfapair(1, 3)
single()
@RULES
_PAIR [base] <- _LIT \= [trig] _xWILD [min=1 max=1 match=(_LIT _NONLIT _STR _NUM _LIST) recurse=(listarg argtolist)] @@
|
@NODES _LINE
# Fix it up as a capitalized phrase!
# Special terminations
@RULES
_companyRoot <-
of [s] _xWHITE [s star]
North [s opt] _xWHITE [s star]
America [s] @@
@POST
S("len") = 2;
single()
@RULES
_Caps [base unsealed] <-
\, [s] _xWHITE [s star]
_parenList _xWHITE [s star]
_companyRoot [s trig]
@@
# NORMAL TERMINATION CASE.
@POST
S("len") = 1;
single()
@RULES
_Caps [base unsealed] <-
\, [s opt]
_xWHITE [s star]
_companyRoot [s trig]
@@
# If multiple company roots, handle that.
@POST
++N("len",1);
listadd(1,4)
@RULES
_xNIL <- _Caps \, [s opt] _xWHITE [s star] _companyRoot [s] @@
@POST
++N("len",3);
listadd(3,1) # REVERSE list-building. # 12/05/99 AM.
@RULES
# Note: need a separate rule for punct. _xANY before them is also
# glommed in.
_xNIL <-
_xWILD [s one match=( _xCAP \& \* \! \- \/
of and
_possessive _hardware
_companyMod _companyRoot _educationPart # Some crud.
_xNUM _parenList
)]
_xWHITE [s star]
_Caps [trig] @@
# Single char abbreviations.
@PRE
<1,1> length(1)
@POST
++N("len",4);
listadd(4,1)
@RULES
_xNIL <- _xCAP [s] \. [s] _xWHITE [s star] _Caps [trig] @@
# DOT-COM companies!
@RULES
_companyMod <- \. _xWHITE [s star]
_xWILD [s one match=(Com Net)] @@
# The important INTERNATIONAL word.
_companyMod <- int [s] \' [s] l [s] @@
# Parenthesized list.
@RULES
_parenList <- \( [s]
_xWILD [s min=1 max=6 fail=( \) _DateRange _SingleDate )]
\) [s] @@
|
# Create apple concept with color and weight attributes, then print out values
if (findconcept(findroot(),"apple"))
# kill them (to start fresh)
rmconcept(findconcept(findroot(),"apple"));
# Create the apple concept
G("apple") = makeconcept(findroot(),"apple");
# Apple's color's red
addstrval(G("apple"),"color","red");
# MacIntosh apples are also red
addstrval(makeconcept(G("apple"),"MacIntosh"),"color","red");
# Print out apple's attribute and value
if(attrname(findattrs(G("apple")))){
"output.txt" << "1) apples " << attrname(findattrs(G("apple")));
"output.txt" << " " << getstrval(attrvals(findattrs(G("apple")))) << "\n";
}
# Test for a given attribute and value
"output.txt" << "2) are apples red? " <<
attrwithval(G("apple"), "color", "red") << "\n";
"output.txt" << "3) are apples blue? " <<
attrwithval(G("apple"), "color", "blue") << "\n";
# Change a value of a given attribute to another value.
attrchange(findroot(), "color", "red", "blue");
"output.txt" << "4) are apples blue now? " <<
attrwithval(G("apple"), "color", "blue") << "\n"; |
@DECL
SomeFunction(L("attr"), L("value")) {
if (!L("attr") || !L("value"))
return 0;
"output.txt" << L("attr") << " is " << L("value") << "\n";
return 1;
}
@@DECL |
@NODES _ROOT
@POST
noop();
@@POST
@RULES
_xNIL <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
\- [opt] ### (3)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (4)
_cardinalNumeral [one] ### (5)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (6)
\- [one] ### (7)
@@
_xNIL <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
\- [opt] ### (3)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (4)
_ordinalNumeral [one] ### (5)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (6)
\- [one] ### (7)
@@
@@RULES
@CHECK
if ( N("Numeral Value",1) != 100 && N("Numeral Value",1) != 200 && N("Numeral Value",1) != 300 && N("Numeral Value",1) != 400 && N("Numeral Value",1) != 500 && N("Numeral Value",1) != 600 && N("Numeral Value",1) != 700 && N("Numeral Value",1) != 800 && N("Numeral Value",1) != 900)
fail();
if ( N("Numeral Value",5) >99)
fail();
@@CHECK
@POST
S("Numeral Value") = N("Numeral Value",1) + N("Numeral Value",3);
single();
@@POST
@RULES
_cardinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
_cardinalNumeral [one] ### (5)
@@
_ordinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
_ordinalNumeral [one] ### (5)
@@
@@RULES
@CHECK
if ( N("Numeral Value",1) != 20 && N("Numeral Value",1) != 30 && N("Numeral Value",1) != 40 && N("Numeral Value",1) != 50 && N("Numeral Value",1) != 60 && N("Numeral Value",1) != 70 && N("Numeral Value",1) != 80 && N("Numeral Value",1) != 90)
fail();
if ( N("Numeral Value",5) >9)
fail();
@@CHECK
@POST
S("Numeral Value") = N("Numeral Value",1) + N("Numeral Value",5);
single();
@@POST
@RULES
_cardinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
\- [opt] ### (3)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (4)
_cardinalNumeral [one] ### (5)
@@
_ordinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
\- [opt] ### (3)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (4)
_ordinalNumeral [one] ### (5)
@@
@@RULES
@CHECK
if ( N("Numeral Value",1) >100 )
fail();
if ( N("Numeral Value",3) < 99 )
fail();
@@CHECK
@POST
G("num1") = N("Numeral Value",1) ;
G("num2") = N("Numeral Value",5) ;
S("Numeral Value") = G("num1") + G("num2") ;
single();
@@POST
@RULES
_cardinalNumeral <-
_cardinalNumeral [s one] ### (1)
_xWILD [s opt match=(_xWHITE "_whiteSpace")] ### (2)
_xWILD [s opt match=("and" "&")] ### (3)
_xWILD [s opt match=(_xWHITE "_whiteSpace")] ### (4)
_cardinalNumeral [s one] ### (5)
@@
_ordinalNumeral <-
_cardinalNumeral [s one] ### (1)
_xWILD [s opt match=(_xWHITE "_whiteSpace")] ### (2)
_xWILD [s opt match=("and" "&")] ### (3)
_xWILD [s opt match=(_xWHITE "_whiteSpace")] ### (4)
_ordinalNumeral [s one] ### (5)
@@
@@RULES
@CHECK
if ( N("Numeral Value",1) <1 || N("Numeral Value",1)>999)
fail();
if ( N("Numeral Value",3) != 100 && N("Numeral Value",3) != 1000 && N("Numeral Value",3) != 1000000 && N("Numeral Value",3) != 1000000000 && N("Numeral Value",3) != 1000000000000 )
fail();
@@CHECK
@POST
G("num1") = N("Numeral Value",1) ;
G("num2") = N("Numeral Value",3) ;
S("Numeral Value") = G("num1") * G("num2") ;
single();
@@POST
@RULES
_cardinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
_cardinalNumeral [one] ### (3)
@@
_ordinalNumeral <-
_cardinalNumeral [one] ### (1)
_xWILD [opt match=(_xWHITE "_whiteSpace")] ### (2)
_ordinalNumeral [one] ### (3)
@@
@@RULES
######################
# (rule 4a) #
# Decimal numbers #
######################
############################
# Rule 5 #
# Dotted telephone numbers #
############################
@CHECK
if (strlength(N("$text",5)) != 4 ||
strlength(N("$text",3)) != 3 ||
strlength(N("$text",1)) != 3
) {
fail() ;
}
else succeed() ;
@@CHECK
@POST
S("area code") = N("$text",1);
S("local number") = N("$text",3) + N("$text",5);
single() ;
@@POST
@RULES
_phoneNumber <-
_xNUM [one] ### (1)
\. [one] ### (2)
_xNUM [one] ### (3)
\. [one] ### (4)
_xNUM [one] ### (5)
@@
@@RULES
@POST
S("NumberOfFields") = 3;
single();
@@POST
@RULES
_dottedNumber <-
_xNUM [one] ### (1)
\. [one] ### (2)
_xNUM [one] ### (3)
\. [one] ### (2)
_xNUM [one] ### (3)
@@
@@RULES
@POST
S("NumberOfFields") = N("NumberOfFields",1)+1;
listadd(1,3,"true");
@@POST
@RULES
_dottedNumber <-
_dottedNumber [one]
\. [one]
_xNUM [one]
@@
@@RULES
@POST
S("Numeral Value") = num(N("$text",1)) ;
S("fractional value") = num(N("$text",3)) ;
single() ;
@@POST
@RULES
_decimalNumber <-
_xNUM [one] ### (1)
\. [one] ### (2)
_xNUM [one] ### (3)
@@
@@RULES
######################
# (rule 4b) #
# Numbers with commas#
######################
@POST
S("Numeral Value") = 1000*num(N("$text",1)) + num(N("$text",3)) ;
single() ;
@@POST
@RULES
_decimalNumber <-
_xNUM [one] ### (1)
_xWILD [one match=(",")] ### (2)
_xNUM [one] ### (3)
@@
@@RULES
@POST
S("Numeral Value") = 1000*N("Numeral Value",1) + num(N("$text",3));
single();
@@POST
@RULES
_decimalNumber <-
_decimalNumber ### (1)
\, [one] ### (2)
_xNUM [one] ### (3)
@@
@@RULES
|
@PATH _ROOT _headerZone
@RULES
_iClose <- \< \/ i \> @@
_iOpen <- \< i _xWILD [fail=(\>)] \> @@
_liClose <- \< \/ li \> @@
_liOpen <- \< li _xWILD [fail=(\>)] \> @@
_paraClose <- \< \/ p \> @@
_paraOpen <- \< p \> @@
|
@NODES _LINE
@POST
X("words") = X("words") + 1;
if (strlength(N("$text")) > 1) {
if (striscaps(N("$text"))) {
X("caps") = X("caps") + 1;
}
else if (strisupper(N("$text"))) {
X("upper") = X("upper") + 1;
}
}
if (N("common") || (N("lower") && spellword(N("$text")))) {
X("prose")++;
}
@RULES
_xNIL <-
_xALPHA ### (1)
@@
|
@NODES _LINE
@RULES
_brackets <-
\[ ### (1)
_xWILD [fail=(\])] ### (2)
\] ### (3)
@@
|
@NODES _ROOT
@RULES
_row <-
_trStart ### (1)
_xWILD [fail=(_trEnd)] ### (2)
_trEnd ### (3)
@@
|
@NODES _ROOT
@CHECK
if (G("people"))
succeed();
fail();
@RULES
_person <-
fred ### (1)
@@
@CHECK
if (G("people"))
fail();
succeed();
@RULES
_animal <-
horse ### (1)
@@
|
# Fetch the first node in given node's list.
L("return_con") = listnode(L("node")); |
@NODES _ROOT
@RULES
_BLANKLINE <-
_xWILD [min=0 max=0 matches=(\ \t \r)] ### (1)
\n ### (2)
@@
@POST
S("num") = G("line count")++;
single();
@RULES
_LINE <-
_xWILD [min=0 max=0 fails=(\r \n)] ### (1)
_xWILD [one match=(\n _xEND)] ### (2)
@@
|
@NODES _ROOT
@POST
excise(1,1);
@RULES
_xNIL <-
_zap ### (1)
@@
|
# Find 'mother' in the KB, then go up to the parent concept, then go back down to 'mother' concept, then find mother's first sibling, then find daughters previous sibling
@CODE
# Find 'mother' in the KB
G("First") = findhierconcept("mother", findroot());
# goto parent concept (female)
G("Second") = up(G("First"));
"output.txt" << conceptname(G("Second")) << "\n";
# go back down to 'mother'
G("First") = down(G("Second"));
"output.txt" << conceptname(G("First")) << "\n";
# find mother's first sibling, which should be 'daughter'
G("Second") = next(G("First"));
"output.txt" << conceptname(G("Second")) << "\n";
# find daughter's previous sibling, which should be mother
G("First") = prev(G("Second"));
"output.txt" << conceptname(G("First")) << "\n"; |
@NODES _ROOT
@RULES
_Email <-
_xWILD [match=(_xALPHA \. _xNUM \_)]
\@ [s]
_xWILD [match=(_xALPHA _xNUM \_)]
\.
_xWILD [match=(_xALPHA \. _xNUM \_)]
@@
|
@CODE
G("sortedWordArr") = sortconsbyattr(G("wordArr"),"count",1,1);
G("wordArr") = 0;
if (G("wordArrLen")) {
L("i") = 0;
while (L("i") < G("wordArrLen")) {
L("word") = G("sortedWordArr")[L("i")];
"wordFile.txt" << conceptname(L("word")) << ", " << numval(L("word"), "count") << "\n";
++L("i");
}
}
@@CODE |
# Create the concept apple, then create two attributes of apple, then delete both of them. Finally, add both of the attributes back and then delete only one
if(findconcept(findroot(), "apples"))
rmconcept(findconcept(findroot(), "apples"));
G("apples") = makeconcept(findroot(), "apples");
addstrval(G("apples"), "color", "red");
addstrval(G("apples"), "taste", "good");
# remove all attributes
rmattrs(G("apples"));
addstrval(G("apples"), "color", "red");
addstrval(G("apples"), "taste", "good");
# only remove named attribute
rmattr(G("apples"),"color"); |
@NODES _split
@POST
L("text") = N("code", 1);
"debug.txt" << L("text") << "\n";
if (strcontains("-", L("text"))) {
L("range") = split(L("text"), "-");
L("class_start") = L("range")[0];
L("class_end") = L("range")[1];
L("class_end") = split(L("class_end"), ".");
L("class_end") = L("class_end")[0];
S("con") = AddUniqueCon(X("con"), N("term", 1));
addstrval(S("con"), "code", L("text"));
addstrval(S("con"), "start", L("class_start"));
addstrval(S("con"), "end", L("class_end"));
merge();
}
noop();
@RULES
_class <-
_entry ### (1)
@@
|
@PATH _ROOT
@POST
noop();
@RULES
_xNIL <- _xSTART
_xWILD [s min=0 max=25 match=(_BLANKLINE _horizRule)]
_xANY
@@
@RULES
_contactBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_contactHeader) trigger]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_educationBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_educationHeader) trigger]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
# This one finds the word or phrase alone on a line.
# But now I'm accepting it if it's one of very few words on a line,
# so this rule is unneeded.
# _educationBoundary <-
# \n [opt s]
# _xBLANK [star s]
# _xWILD [s one matches=(_educationWord _educationPhrase)]
# _xBLANK [star s]
# \n [opt s]
# @@
_experienceBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_experienceHeader) trigger]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
# _experienceBoundary <-
# \n [opt s]
# _xBLANK [star s]
# _xWILD [s one matches=(_experienceWord _experiencePhrase)]
# _xBLANK [star s]
# \n [opt s]
# @@
_objectiveBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_objectiveHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_presentationsBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_presentationsHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_publicationsBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_publicationsHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_skillsBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_skillsHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_referencesBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_referencesHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
_otherBoundary <-
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
_xWILD [tree one matches=(_otherHeader) trig]
_BLANKLINE [s star]
_horizRule [s opt]
_BLANKLINE [s star]
@@
# _otherBoundary <-
# \n [opt s]
# _xBLANK [star s]
# _xWILD [s one matches=(_otherWord _otherHeaderPhrase)]
# _xBLANK [star s]
# \n [opt s]
# @@
# Finding headers based on having blank lines around.
# Should check first word capitalized.
#@CHECK
# Nlt(4, "nblobs", 5)
# Neq(4, "addressParts", 0)
#@RULES
#_otherBoundary <-
# _BLANKLINE [s star]
# _horizRule [s opt]
# _BLANKLINE [s plus]
# _xWILD [s one fails=(_horizRule _BLANKLINE) trig]
# _BLANKLINE [s plus]
# _horizRule [s opt]
# _BLANKLINE [s star]
# @@
# Let a horiz line be an empty section.
#@RULES
#_otherBoundary <-
# _BLANKLINE [s star]
# _horizRule [s]
# _BLANKLINE [s star]
# @@
|
@CODE
G("langs") = getconcept(findroot(),"langs");
@@CODE |
@PATH _ROOT _LINE
@POST
"negex.dict" << N("$text",1) << " negex=" << X("code",2) << "\n";
"negtest.txt" << N("$text",1) << "\n";
@RULES
_xNIL <-
_phrase ### (1)
@@
|
@CODE
L("hello") = 0;
@@CODE
#@PATH _ROOT _TEXTZONE _sent _clause _seg
@NODES _seg
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
if (!N("noun",4))
fail();
@POST
L("tmp4") = N(4);
L("tmp3") = N(3);
if (N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
}
else
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
}
group(4,4,"_noun");
pncopyvars(L("tmp4"),X());
clearpos(X(),1,1); # Zero out token info.
xrename("_np");
@RULES
_xNIL <-
_xSTART
_pro [s]
_xALPHA
_xALPHA
_xEND
@@
# dqan
@CHECK
L("x4") = pnparent(X()); # 07/13/12 AM.
if (N("verb",6)
&& pnvar(L("x4"),"need verb")) # 02/12/05 AM.
fail();
@POST
L("tmp6") = N(6);
if (N(7))
{
if (N("adj",6) || !N("noun",6))
{
group(6,6,"_adj");
pncopyvars(L("tmp6"),N(6));
fixadj(N(6));
}
else
{
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
}
}
else
{
group(6,6,"_noun");
pncopyvars(L("tmp6"),N(6));
}
xrename("_np");
clearpos(X(),1,1);
@RULES
_xNIL <-
_xSTART
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [star]
_xALPHA
_noun [star]
_xEND
@@
# dqan
@POST
fixnpnonhead(4);
fixnpnonhead(5);
xrename("_np");
L("head") = lasteltnode(7);
pncopyvars(L("head"),X());
clearpos(X(),1,1);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_xALPHA
_xALPHA
_adj [star]
_noun [plus]
_xEND
@@
# dqan
@POST
fixnpnonhead(4);
xrename("_np");
L("head") = lasteltnode(6);
pncopyvars(L("head"),X());
clearpos(X(),1,1);
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_xALPHA
_adj [star]
_noun [plus]
_xEND
@@
|
# create the concepts named words and noun as children to the root of the KB, then make the concept book a child of words
G("words") = makeconcept(findroot(), "words");
G("noun") = makeconcept(findroot(),"noun");
G("noun_book") = makeconcept(G("words"),"book"); |
# Fetch the string value name of a concept
@CODE
"output.txt" << conceptname(makeconcept(findroot(),"test")); |
@DECL
SaveName(L("name")) {
L("word") = strtolower(L("name"));
L("gender") = "male";
if (strcontains("female",strtolower(G("$inputhead")))) {
L("gender") = "female";
}
G("file") << L("word") << " name=first gender=" << L("gender") << "\n";
}
@@DECL |
@NODES _ROOT
@RULES
_rest <-
_xWILD [fail=(_td)]
@@
|
# Sort concept's subhierarchy alphabetically.
sorthier(L("con")); |
## CODE REGION
@CODE
# NOTE: Because the analyzer could potentially find multiple contact
# zones, gathering data globally. (Could gather data in each
# contact zone node also, if beneficial.)
# Count the number of ASSIGNED phone numbers found in contact zones.
G("Assigned Phones") = 0; # No assigned phones yet.
# Count the TOTAL number of phone numbers found in contact zones.
G("Total Phones") = 0;
# Count the total number of names found in contact zone.
G("Total Names") = 0;
@@CODE
# Look only in contact zones.
@MULTI _contactZone
@POST
++G("Assigned Phones");
@RULES
_xNIL <- _xWILD [one match=(
_phoneHomePhrase
_phoneWorkPhrase
_phoneFaxPhrase
_phonePagerPhrase
_phoneCellPhrase
_phoneHomeFaxPhrase
)] @@
@POST
++G("Total Phones");
@RULES
_xNIL <- _phoneNumber @@
@POST
++G("Total Names");
@RULES
_xNIL <- _humanName @@
|
@NODES _attrBlock
@POST
pnrename(N(1),"_word");
pnrename(N(2),"_zero");
pnrename(N(3),"_attr");
pnrename(N(4),"_type");
pnrename(N(5),"_value");
@RULES
_xNIL <-
_LINE ### (1)
_LINE ### (2)
_LINE ### (3)
_LINE ### (4)
_LINE ### (5)
@@
|
@NODES _ROOT
@POST
S("header") = N("header", 1);
single();
@RULES
_headerZone <-
_header ### (1)
_xWILD [fail=(_header _xEND)] ### (2)
@@
|
@DECL
#################
## FN: POSTORDER
## SUBJ: Postorder action for kb traversal.
#################
postorder(L("con"))
{
if (!L("con"))
return; # Error.
"postorder.txt" << "postorder: " << conceptname(L("con")) << "\n";
}
#################
## FN: PREORDER
## SUBJ: Preorder action for kb traversal.
#################
preorder(L("con"),L("depth"))
{
if (!L("con"))
return; # Error.
indent(L("depth"),"preorder.txt");
"preorder.txt" << conceptname(L("con")) << "\n";
}
#################
## FN: INDENT
## SUBJ: Output tabs.
#################
indent(L("num"),L("out"))
{
while (L("num") > 0)
{
L("out") << "\t";
--L("num");
}
}
#################
## FN: KBTRAVERSE
## SUBJ: Traverse a subhierarchy in knowledge base.
#################
kbtraverse(L("root"))
{
L("con") = L("root");
L("depth") = 0;
L("path") = 0; # Store the current path in the parse tree.
while (L("con")) # Traverse the parse tree.
{
preorder(L("con"),L("depth"));
L("path")[L("depth")] = L("con"); # Update path in the parse tree.
L("childs") = down(L("con"));
if (L("con") != L("root")) # FIX. # 10/03/13 AM.
L("next") = next(L("con"));
else # FIX. # 10/03/13 AM.
L("next") = 0; # FIX. // 10/03/13 AM.
if (L("childs")) # If node has children, traverse them.
{
L("con") = L("childs");
++L("depth");
}
else if (L("next")) # If node has right sibling, get it.
{
postorder(L("con"));
L("con") = L("next");
}
else # Try continuing from a parent.
{
postorder(L("con"));
L("done") = 0;
L("con") = 0;
while (!L("done")) # Find a continuation or quit.
{
if (--L("depth") > 0)
{
L("con") = L("path")[L("depth")]; # Move up to parent.
if (L("con"))
postorder(L("con"));
if (L("con") = next(L("con"))) # Try to move right.
L("done") = 1;
}
else
L("done") = 1;
}
}
}
# Traverse the given root
postorder(L("root"));
}
@CODE
L("hello") = 1;
#kbtraverse(findroot()); # Traverse the kb.
@@CODE
|
@PATH _ROOT _group _subgroup _LINE
@PRE
<1,1> var("SP");
<2,2> var("NOSP");
@POST
L("text") = str(N("$text",1)) + str(N("$text",2));
if (strlength(L("text")) <= 5) {
if (IsHex(L("text")))
single();
}
@RULES
_hex <-
_xWILD [one] ### (1)
_xWILD [min=1 max=4] ### (2)
@@ |
@CODE
L("hello") = 0;
@@CODE
@NODES _TEXTZONE
@POST
L("start") = N(2);
L("end") = lasteltnode(2);
pnreplacevalrange("mypos","NP",L("start"),L("end"));
singler(2,3);
@RULES
_company [layer=_noun] <-
_xWILD [one match=(_dbldash \, _conj)]
_xWILD [s plus match=(_xCAP) except=(_companyDESIG)]
_companyDESIG [s]
@@
# Idiom...
# look forward to
@POST
N("particle",3) = 1; # Flag particle for use someday.
L("tmp1") = N(1);
L("tmp2") = N(2);
N("verb",1) = 1;
N("noun",1) = N("adj",1) = N("adv",1) = 0;
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
N("adv",2) = 1;
N("adj",2) = 0;
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [one match=(look looks looking looked)]
forward
to [s]
@@
# 10 day old
# num noun old
@CHECK
S("number") = number(N(2));
if (S("number") != "singular")
fail();
@POST
pncopyvars(3);
single();
@RULES
_ambig <-
_xWILD [one match=(_quan _num _xNUM)]
_noun
old
@@
# days old
# noun old
@CHECK
S("number") = number(N(1));
if (S("number") != "plural")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
group(2,2,"_adjc");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"JJ");
@RULES
_xNIL <-
_noun
old
@@
# prep about
@PRE
<2,2> varz("mypos");
@POST
N("mypos",2) = "RB"; # about/RB
@RULES
_xNIL <-
_prep
about [s]
@@
# about $
# about num
# around num #
@POST
pnrename(N(1),"_adv");
# if (G("conform treebank"))
# chpos(N(1),"IN"); # treebank: about
chpos(N(1),"RB"); # about/RB
@RULES
_xNIL <-
_xWILD [s one match=(around about)]
_xWILD [one lookahead match=(_quan _num _xNUM)]
@@
_xNIL <-
_xWILD [s one match=(around about)]
_xWILD [s one lookahead match=(_num _money _percent)]
@@
# about noun
# around noun
@PRE
<2,2> var("num");
@POST
pnrename(N(1),"_adv");
if (G("conform treebank"))
chpos(N(1),"IN"); # treebank: about
else
chpos(N(1),"RB"); # about/RB
@RULES
_xNIL <-
_xWILD [s one match=(around about)]
_noun [lookahead]
@@
# none of
@POST
L("tmp1") = N(1);
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"NN");
@RULES
_xNIL <-
none
of [s lookahead]
@@
# noun num
@POST
if (pnname(N(1)) == "_np")
{
N("bracket",1) = 0; # CLEAR.
group(1,2,"_money");
nountonp(1,1);
N("money",1) = 1;
}
else
{
group(1,2,"_money");
group(1,1,"_noun");
N("money",1) = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [s one match=(thousand million mln billion trillion)]
@@
# Normalize tagging.
# chief executive officer
# chief operating officer
@POST
pncopyvars(3);
single();
@RULES
_noun <-
chief [s]
_xWILD [s one match=(executive operating financial)]
_xWILD [s one match=(officer officers)]
@@
# got
@POST
N("-edn",3) = 0;
N("-en",3) = 1;
@RULES
_xNIL <-
_have [s]
_adv [star]
got [s]
@@
@POST
N("-edn") = 0;
N("-ed") = 1;
@RULES
_xNIL <-
got [s]
@@
# for now
# now
# Idiom, fairly arbitrary tagging.
@POST
if (nonliteral(N(2)))
pnrename(N(2),"_noun");
chpos(N(2),"RB");
group(1,2,"_advl");
@RULES
_xNIL <-
for [s]
now [s]
@@
# Need a designator for default pos, original pos, etc.
# in Eastern
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_prep
_xWILD [one match=(northern southern eastern western)]
@@
# still hope
# Sometimes, only semantics will solve a thing.
# Or at least collocation knowledge...
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
group(1,1,"_adv");
pncopyvars(L("tmp1"),N(1));
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("txt") = strtolower(N("$text",2));
if (strendswith(L("txt"),"s"))
fixverb(N(2),"active","VBZ");
else
fixverb(N(2),"active",0); # Don't know yet...
@RULES
_xNIL <-
still
_xWILD [one match=(
hope hopes
)] @@
# come true.
# Seems highly idiomatic to me.
# come seems a copula, linking verb here.
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [s one match=(come comes coming came)]
true
@@
# much of
# "much of the food".
# Seems more like a quantifier/predeterminer, so silly to
# fret about whether it's a jj or rb here, or nn.
@POST
L("tmp1") = N(1);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"JJ");
group(1,2,"_advl");
@RULES
_xNIL <-
much
of [s]
@@
# nothing short of -> adv
# Note: Idiom.
# alpha of
# part of
# Note: Some idiomatic nouns.
@POST
fixnphead(1);
@RULES
_xNIL <-
_xWILD [one match=(part)]
of [s lookahead]
@@
# alpha of
# short of
# Note: Some idiomatic adjs.
@POST
alphatoadj(1);
group(1,1,"_adjc");
@RULES
_xNIL <-
_xWILD [one match=(short)]
of [s lookahead]
@@
# alpha of
# full of
# empty of
# Note: Some adj subcategories...
@POST
alphatoadj(1);
@RULES
_xNIL <-
_xWILD [one match=(full empty)]
of [s lookahead]
@@
# alpha of
@CHECK
if (!N("verb",1))
fail(); # Verb + one other.
L("txt") = nodesem(N(1));
if (finddictattr(L("txt"),"verb-of"))
fail();
@POST
N("verb",1) = 0;
--N("pos num",1);
alphaunambigred(1);
@RULES
_xNIL <-
_xALPHA
of [s lookahead]
@@
# well prep
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
group(2,3,"_prep");
@RULES
_xNIL <-
_xWILD [one match=(_qEOS _xSTART \, )]
well
_prep
@@
# both X and Y
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"CC");
@RULES
_xNIL <- # 22
both [s]
_xWILD [one]
and [s]
@@
# what alpha is
@CHECK
if (!N("verb",2))
fail();
S("number") = number(N(2));
if (S("number") != "plural"
|| !N("-s",3))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
clearpos(N(2),1,0);
fixvg(N(2),"active","VBP");
@RULES
_xNIL <-
what [s]
_xALPHA
_be [s lookahead]
@@
# what
# If not claimed elsewhere, give it a default here.
#@CHECK
# if (N("mypos"))
# fail();
#@POST
# chpos(N(1),"WP");
#@RULES
#_xNIL <-
# what [s]
# @@
# Many adverbs are special.
# new alpha
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
fixadj(N(1));
if (N("adj",2) && N("noun",2))
; # Don't know which.
else if (N("adj",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
new
_xALPHA
@@
# det while
@POST
chpos(N(3),"NN");
pnrename(N(3),"_noun");
@RULES
_xNIL <-
_xWILD [s one match=(a an the)]
_xWILD [star match=(_adv)]
while [s]
@@
# while not alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
while [s]
not [s]
_xALPHA
@@
# while alpha
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("noun",2) || !N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "inf"
&& S("vc") != "-s")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
@RULES
_xNIL <-
while [s]
_xALPHA
@@
# while
@CHECK
if (N("mypos",1) == "IN")
fail();
@POST
chpos(N(1),"IN");
@RULES
_xNIL <-
while [s]
_xWILD [s one fail=(away)]
@@
# alpha
@CHECK
# noun vs adj.
if (N("pos num",2) != 2)
fail();
if (!N("noun",2) || !N("adj",2))
fail();
@POST
# Could be wrong, but choose adj for now.
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
@RULES
_xNIL <-
_xWILD [one match=(_xSTART)]
_xALPHA
@@
# Idiom.
@CHECK
if (!N("noun",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
vice [s]
_xALPHA
_xWILD [one lookahead match=(_prep \, _qEOS _xEND)]
@@
# idiom: after all
@POST
group(1,2,"_advl");
@RULES
_xNIL <-
after [s]
all [s]
_xWILD [one lookahead fail=(_det _quan _num _xNUM _adj _noun)]
@@
# by then
# by fnword
# I disagree with pos assignment of RB here. Should be NN.
@CHECK
if (N("by-then",2))
fail();
@POST
L("tmp2") = N(2);
pnrename(N(2),"_noun");
chpos(N(2),"RB"); # Debatable.
nountonp(2,1);
N("by-then",2) = 1;
@RULES
_xNIL <-
by [s]
_xWILD [s one match=(now then)]
@@
# idiom
# those of them
@POST
nountonp(3,1);
nountonp(1,1);
group(1,3,"_np");
@RULES
_xNIL <-
those [s]
of [s]
_proObj [s]
@@
# those who
@POST
nountonp(1,1);
@RULES
_xNIL <-
those [s]
_xWILD [s one lookahead match=(who which that whom _prep)]
@@
# because of
# because
# Note: pretagged has inconsistencies
# "because of" = sometimes 1st is RB, sometimes IN.
# DON'T CARE!
@CHECK
if (N("mypos",1))
fail();
@POST
if (N(2))
{
L("tmp2") = N(2);
chpos(N(1),"IN");
group(1,2,"_prep");
pncopyvars(L("tmp2"),N(2));
N("because-of",1) = 1;
clearpos(N(1),1,0);
}
else
chpos(N(1),"IN");
@RULES
_xNIL <-
because [s]
of [s opt]
@@
# even though
@CHECK
if (N("mypos",1))
fail();
@POST
L("tmp2") = N(2);
chpos(N(1),"RB");
group(1,1,"_adv");
group(1,2,"_fnword");
pncopyvars(L("tmp2"),N(1));
clearpos(N(1),1,0);
@RULES
_xNIL <-
even [s]
though [s] # FIX. # 04/15/10 AM.
@@
# there 's
@POST
pnrename(N(2),"_verb");
chpos(N(2),"VBZ"); # 's/VBZ # 's -> verb
N("copula",2) = 1;
@RULES
_xNIL <-
_xWILD [s one match=(there _nounpro)]
_aposS
@@
@POST
if (!N("mypos",1))
chpos(N(1),"WP");
if (pnname(N(1)) == "_fnword")
{
nountonp(1,1);
N("fnword",1) = "whword";
}
L("tmp2") = N(2);
pnrename(N(2),"_be");
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"VBZ");
N("apos-s",2) = 1;
@RULES
_xNIL <-
_whword [s]
_aposS
@@
# nounpro adj
# Eg: anything (that is) green
@POST
N("ellipted-that-is",2) = 1;
@RULES
_xNIL <-
_nounpro [s]
_adj [lookahead]
@@
# that 's
@POST
# L("tmp1") = N(1);
# group(1,1,"_noun");
# pncopyvars(L("tmp1"),N(1));
# chpos(N(1),"WDT");
pnrename(N(2),"_verb");
chpos(N(2),"VBZ");
N("verb",2) = 1;
@RULES
_xNIL <-
that [s]
_aposS
@@
# pro 's
@POST
pnrename(N(2),"_verb");
chpos(N(2),"VBZ"); # 's/VBZ
N("verb",2) = 1;
N("copula",2) = 1;
@RULES
_xNIL <-
_pro
_aposS
@@
# noun 's and
@POST
L("tmp1") = N(1);
chpos(N(2),"POS"); # 's/POS
group(1,2,"_adj");
N("apos-s",1) = 1;
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_noun
_aposS
_xWILD [one match=(_conj \) \} \] \, _dbldash _xEND)]
@@
# 's and
@POST
chpos(N(1),"POS"); # 's/POS
@RULES
_xNIL <-
_aposS
_xWILD [one match=(_conj \) \} \] \, _dbldash _xEND)]
@@
# than predicted
# than alpha
@CHECK
S("vc") = vconj(N(2));
if (S("vc") != "-en"
&& S("vc") != "-edn"
&& S("vc") != "-ing")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
if (S("vc") == "-ing")
{
L("voice") = "active";
chpos(N(2),"VBG");
}
else if (S("vc") == "-en") # 02/19/05 AM.
{
L("voice") = "passive";
chpos(N(2),"VBN");
}
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = L("voice");
N("first verb",2) = N("verb node",2) = L("v");
clearpos(N(2),1,0);
@RULES
_xNIL <-
than [s]
_xALPHA
@@
# there is
# existential there.
@CHECK
if (pnname(N(2)) == "_np")
S("n") = pndown(N(2)); # Don't put mypos on an _np.
else
S("n") = N(2);
if (pnvar(S("n"),"mypos"))
fail();
@POST
pnreplaceval(S("n"),"mypos","EX");
@RULES
_xNIL <-
_xWILD [one match=(_xSTART \, _qEOS)]
there [s]
_be [s]
@@
# once used to
@CHECK
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
# fixverb(N(2),"passive","VBP");
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
_xALPHA
_xWILD [s one lookahead match=(to)]
@@
# a more
# det more
# more
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"RBR");
pnrename(N(2),"_adj"); # Forms np more easily...
@RULES
_xNIL <-
a [s]
more [s]
@@
# more like
@CHECK
if (N("mypos",1) == "RBR")
fail(); # Loop guard.
@POST
pnrename(N(2),"_prep");
pnrename(N(1),"_adv");
chpos(N(1),"RBR");
@RULES
_xNIL <-
_xWILD [s one match=(more less)]
like [s]
@@
# more verb than
@POST
chpos(N(1),"RBR"); # more/RBR
chpos(N(3),"IN"); # than/IN
pnrename(N(1),"_adv");
if (vconjq(N(2),"-en"))
{
pnrename(N(2),"_adj"); # verb -> adj
N("mypos",2) = "VBN"; # Conform Treebank.
}
else if (vconjq(N(2),"-ing"))
{
pnrename(N(2),"_adj"); # verb -> adj
N("mypos",2) = "VBG"; # Conform Treebank.
}
@RULES
_xNIL <-
more [s]
_verb
than [s]
@@
# more alpha than
@POST
chpos(N(3),"IN"); # than/IN
pnrename(N(1),"_adv");
L("tmp2") = N(2);
if (N("adj",2))
{
chpos(N(1),"RBR"); # more/RBR
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
else if (N("adv",2))
{
chpos(N(1),"RBR"); # more/RBR
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
fixadv(N(2));
}
else if (N("verb",2) && !vconjq(N(2),"inf"))
{
chpos(N(1),"JJR"); # more/JJR
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
}
else if (N("noun",2))
{
chpos(N(1),"JJR"); # more/JJR
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
}
@RULES
_xNIL <-
more [s]
_xALPHA
than [s]
@@
# verb alpha
# Note: Some past participle hard-wire. [JJ-VBN JJ|VBN]
@PRE
<4,4> varne("stem","by");
@CHECK
if (!copulaq(N(1)))
fail();
@POST
alphatoadj(3);
group(3,3,"_adjc");
@RULES
_xNIL <-
_verb
_adv [star]
_xWILD [one s match=(
crushed
dumbfounded
interested
relieved
)]
_xWILD [one lookahead match=( _prep \, _qEOS _xEND )]
@@
# be alpha
# be not alpha
@CHECK
if (!N("adj",3) || !N("verb",3))
fail();
if (N("adv",3) || N("noun",3))
fail();
if (!vconjq(N(3),"inf"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
fixadj(N(3));
@RULES
_xNIL <-
_be [s]
not [s opt]
_xALPHA
@@
# be alpha
@CHECK
if (!N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "-ing"
&& S("vc") != "-en"
&& S("vc") != "-edn")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_be [s]
_xALPHA
@@
# fnword not alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"-ing"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
fixverb(N(3),"active","VBG");
@RULES
_xNIL <-
_xWILD [one match=(_fnword)]
not [s]
_xALPHA
@@
# how long
# how alpha
@CHECK
if (!N("adj",2) || !N("verb",2))
fail();
if (!vconjq(N(2),"inf"))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
how [s]
_xALPHA
@@
# how about
@POST
if (!N("mypos",2))
chpos(N(2),"IN");
@RULES
_xNIL <-
how [s]
about [s lookahead]
@@
# how much
# how many
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
group(1,2,"_howmuch");
group(1,1,"_fnword");
@RULES
_xNIL <-
how [s]
_xWILD [one match=(much many)]
@@
# in return
# idiom
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
in [s]
return
@@
# This trend
# this alpha
@CHECK
if (!N("noun",3))
fail();
if (!singular(N(3)))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
@RULES
_xNIL <-
_xWILD [one match=(_qEOS _xSTART)]
this [s ren=_det]
_xALPHA
@@
# most verb
@POST
chpos(N(1),"JJS"); # most/JJS
pnrename(N(1),"_noun");
@RULES
_xNIL <-
most [s]
_xWILD [one lookahead match=(_verb _modal _vg)]
@@
# idiom: the most part.
# the most part
# det most noun
@POST
pnrename(N(2),"_adj");
chpos(N(2),"JJS");
L("tmp3") = N(3);
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
chpos(N(3),"NN");
@RULES
_xNIL <-
the [s]
most [s]
part
@@
# idiom: most of.
# most of
@POST
chpos(N(1),"JJS");
@RULES
_xNIL <-
most [s]
of [s lookahead]
@@
# most alpha
@CHECK
if (!N("noun",2))
fail();
L("txt") = nodetext(N(2));
if (!strendswith(L("txt"),"s"))
fail();
@POST
pnrename(N(1),"_adj");
chpos(N(1),"JJS");
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"NNS");
@RULES
_xNIL <-
most [s]
_xALPHA
@@
# most alpha
@CHECK
if (N("mypos",1))
fail();
if (!N("noun",2) || !N("verb",2))
fail();
if (N("pos num",2) != 2)
fail();
@POST
chpos(N(1),"JJS");
pnrename(N(1),"_adj");
@RULES
_xNIL <-
most [s]
_xALPHA [lookahead]
@@
# very noun
# Note: Change the default pos.
# Ex: "the very thing..."
@POST
pnrename(N(1),"_adj"); # premod -> adj
chpos(N(1),"JJ"); # very/JJ
@RULES
_xNIL <-
very [s]
_xWILD [one lookahead match=(_noun _np)]
@@
# very disappointing
# premod alpha
@CHECK
if (!N("adj",2) && !N("verb",2))
fail();
S("vc") = vconj(N(2));
if (S("vc") != "-ing"
&& S("vc") != "-en"
&& S("vc") != "-edn")
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
group(1,2,"_adj");
# pncopyvars(L("tmp2"),N(1));
@RULES
_xNIL <-
_premod
_xALPHA
@@
# premod alpha
@CHECK
if (!N("adj",2))
fail();
@POST
alphatoadj(2);
@RULES
_xNIL <-
_premod
_xALPHA
@@
# ago
@POST
pncopyvars(1);
single();
@RULES
_noun <-
_xALPHA
ago [s]
@@
# ," alpha
@CHECK
if (!N("verb",3))
fail();
if (!vconjq(N(3),"-ed"))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
L("v") = N(3);
group(3,3,"_vg");
mhbv(N(3),L("neg"),0,0,0,0,L("v"));
N("voice",3) = "active";
@RULES
_xNIL <-
\, [s]
\" [s]
_xALPHA
@@
# to alpha alpha verb
# Note: Ellipted that.
@PRE
<2,2> var("verb");
<3,3> var("noun");
@CHECK
if (!vconjq(N(2),"inf"))
fail();
@POST
fixnphead(3);
alphatovg(2,"active","VB");
@RULES
_xNIL <-
to [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_verb _vg _modal)]
@@
# to alpha alpha .
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf"))
fail();
@POST
L("tmp2") = N(2);
L("tmp3") = N(3);
if (N("noun",3) &&
(N("mass",3) || plural(N(3)) || N("cap",3)) )
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
fixnoun(N(3));
nountonp(3,1);
}
else if (copulaq(N(2)) && N("adj",3))
{
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
}
else if (N("adv",3))
{
group(3,3,"_adv");
pncopyvars(L("tmp3"),N(3));
}
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
chpos(N(2),"VB");
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("voice",2) = "active";
N("verb node",2) = L("v");
clearpos(N(2),1,0);
@RULES
_xNIL <-
to [s]
_xALPHA
_xALPHA
_xWILD [one match=(_qEOS _prep _xEND)]
@@
# Weak rule.
# alpha alpha ,
@CHECK
if (N("pos50 a-a-comma",2))
fail(); # Loop guard.
@POST
N("pos50 a-a-comma",2) = 1;
L("tmp3") = N(3);
if (!N("verb",3))
{
if (N("noun",3))
{
group(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
}
}
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=( \, )]
@@
# alpha fnword
@CHECK
if (!N("noun",2))
fail();
if (N("verb",2))
{
S("vc") = vconj(N(2));
if (S("vc") != "-s"
&& S("vc") != "inf")
fail(); # Too ambiguous.
}
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _qEOS)]
_xALPHA
_xWILD [one lookahead match=(_fnword _prep)]
@@
# too far
@CHECK
if (N("mypos",2))
fail();
@POST
# Conformity. (
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
@RULES
_xNIL <-
_xWILD [s one match=(too very extremely really)]
far
@@
# adv alpha
# too alpha
@CHECK
if (!N("adj",2) && !N("adv",2))
fail();
@POST
L("tmp2") = N(2);
N("noun",2) = N("verb",2) = 0; # Zero these out.
if (!N("adj",2))
{
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
}
else if (!N("adv",2))
{
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
fixadj(N(2));
}
@RULES
_xNIL <-
_xWILD [s one match=(too)]
_xALPHA
@@
# Catching a bad sentence non-break.
@POST
N("eos",2) = 1; # END OF SENTENCE.
++G("eos"); # Count new sentence breaks.
@RULES
_xNIL <-
_prep
_citystate [s]
_xWILD [s one lookahead match=(_xCAP)]
@@
# as adj as
# as adv as
# as verb as
@POST
chpos(N(3),"IN");
chpos(N(1),"RB");
single();
@RULES
_fnword <-
as [s]
_xWILD [one match=(_adj _adv _verb)]
as [s]
@@
# TOO BROAD.
# as adj
#@CHECK
# if (N("mypos",1))
# fail();
#@POST
# chpos(N(1),"RB");
# pnrename(N(1),"_adv");
#@RULES
#_xNIL <-
# as [s]
# _xWILD [one lookahead match=(_adj)]
# @@
# as alpha as
@POST
L("tmp2") = N(2);
chpos(N(3),"IN");
chpos(N(1),"RB");
if (N("adj",2) && !N("adv",2))
group(2,2,"_adj");
else if (N("adv",2) && !N("adj",2))
group(2,2,"_adv");
else if (N("adj",2) && N("adv",2))
group(2,2,"_adj"); # Should check cases...
else if (N("verb",2))
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
S("as-x-as") = 1;
single();
@RULES
_fnword <-
as [s]
_xALPHA
as [s]
@@
# noun - alpha
@CHECK
if (!N("verb",3) || !N("adj",3))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_adj");
pncopyvars(L("tmp3"),N(3));
# clearpos(N(3),0,0);
chpos(N(3),"JJ");
group(1,3,"_adj");
clearpos(N(1),1,0);
@RULES
_xNIL <-
_noun
\-
_xALPHA
@@
# year was
@CHECK
S("num") = num(N("$text",1));
if (S("num") < 1990
|| S("num") > 2010)
fail();
@POST
group(1,1,"_noun");
nountonp(1,1);
N("sem",1) = "date";
++X("date ref");
@RULES
_xNIL <-
_xNUM
_xWILD [s one lookahead match=(is was has does)]
@@
# noun noun
# Domain semantics.
@CHECK
if (!N("cap",2))
fail();
if (N("ne type",2))
fail();
N("ne",2) = 1;
N("ne type conf",2) = 70; # A guess.
if (!N("stem",2))
N("stem",2) = N("$text",2);
if (N("sem",2) == "date")
N("ne type",2) = "date";
# else if (N("ne type",1) == "location")
# N("ne type",2) = "organization"; # Try this.
else if (N("ne type",1) == "person")
{
registerpersnames(N("$text",2));
N("ne type",2) = "person";
}
else if (N("ne type",1))
{
N("ne type",2) = N("ne type",1);
N("ne type conf",2) = 50;
}
fail(); # So that other rules in this pass will still match.
@RULES
_xNIL <-
_noun
_noun
_xWILD [one lookahead fail=(_xALPHA)]
@@
# Domain semantics.
@CHECK
if (!N("cap",1))
fail();
if (N("ne type",1))
fail();
@POST
N("ne type",1) = N("sem",1) = "location";
N("ne type conf",1) = 90;
N("ne",1) = 1;
@RULES
_xNIL <-
_noun
_xWILD [one s match=(
province
)]
@@
# pro say
@POST
fixnphead(2);
L("t1") = strtolower(N("$text",1));
if (L("t1") == "final")
alphatoadj(1);
@RULES
_xNIL <-
_xWILD [s plus match=(my your his her our your their final)]
_xWILD [one match=(say says)]
@@
# say
# says
# Note: Otherwise, it's a verb.
@POST
alphatoverb(1,0,0);
@RULES
_xNIL <-
_xWILD [one match=(say says)]
@@
# said X
@POST
alphatoverb(1,"active","VBP");
if (!N("ne",2) && !N("ne type",2))
{
N("ne type",2) = N("sem",2) = "person";
N("ne type conf",2) = 50;
N("stem",2) = N("$text",2);
N("inverted-verb-noun",1) = N("inverted-verb-noun",2) = 1;
}
@RULES
_xNIL <-
_xWILD [one match=(say says saying said
quip quips quipped quipping
recalls recalled
)]
_noun
@@
# X said in its ...
# (not a person, but an org, perhaps news org)
@CHECK
if (N("stem",2) != "say")
fail();
if (N("ne type",1))
fail();
if (N("sem",1) != "name"
&& !N("cap",1))
fail();
@POST
N("ne type",1) = N("sem",1) = "organization";
N("ne type conf",1) = 95; # Confidence that it's an org.
N("stem",1) = N("$text",1);
if (pnname(N(2)) != "_verb")
{
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
clearpos(N(2),1,0);
fixvg(N(2),"active",0);
}
@RULES
_xNIL <-
_noun
_xWILD [one match=(_verb say says saying said
quip quips quipping quipped recalls recalled)]
_prep
its [s]
@@
# X said
@CHECK
if (N("stem",2) != "say")
fail();
if (N("ne type",1))
fail();
if (N("sem",1) != "name"
&& !N("cap",1))
fail();
@POST
N("ne type",1) = N("sem",1) = "person";
N("ne type conf",1) = 60;
N("stem",1) = N("$text",1);
if (pnname(N(2)) != "_verb")
{
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
L("v") = N(2);
group(2,2,"_vg");
mhbv(N(2),L("neg"),0,0,0,0,L("v"));
pncopyvars(L("tmp2"),N(2));
N("verb node",2) = L("v");
clearpos(N(2),1,0);
fixvg(N(2),"active",0);
N("glom",2) = "left"; # 04/21/07 AM.
}
@RULES
_xNIL <-
_noun
_xWILD [one match=(_verb say says saying said
quip quips quipping quipped
recalls recalled
read )]
@@
# alpha being
# "being" is important, needs careful handling.
# Often ambiguous: "The human being changed."
@CHECK
if (!plural(N(1)))
fail();
@POST
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
N("sem",3) = N("stem",3) = "be";
chpos(N(3),"VBG");
@RULES
_xNIL <-
_xWILD [one match=(_xALPHA _noun)]
_xWILD [star match=(_adv _advl)]
being
@@
# carrier X
@CHECK
if (N("stem",1) != "carrier")
fail();
if (!N("cap",2) && N("sem",2) != "name")
fail();
if (N("ne type",2))
fail();
@POST
N("ne type",2) = "naval vessel";
N("ne type conf",2) = 75;
@RULES
_xNIL <-
_noun
_noun
@@
# name of org
# noun of noun
@CHECK
if (N("ne type",1))
fail();
if (N("ne type",3) != "organization")
fail();
if (!N("cap",1) && N("sem",1) != "name")
fail();
@POST
N("ne type",1) = "person";
N("ne type conf",1) = 75;
if (!N("stem",1))
N("stem",1) = N("$text",1);
@RULES
_xNIL <-
_noun
of [s]
_noun
@@
# alpha .
# Transitive verbs with one other alternative.
@CHECK
if (N("pos num",1) != 2)
fail();
if (!N("verb",1))
fail();
if (!finddictattr(N("$text",1),"trans"))
fail();
@POST
L("tmp1") = N(1);
if (N("noun",1))
{
group(1,1,"_noun");
pncopyvars(L("tmp1"),N(1));
fixnoun(N(1));
}
else
"trans.txt" << "Other alt: " << N("$text",1) << "\n";
@RULES
_xNIL <-
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND)]
@@
# what
# If not claimed elsewhere, give it a default here.
@CHECK
if (N("mypos"))
fail();
@POST
chpos(N(1),"WP");
@RULES
_xNIL <-
what [s]
@@
# Parentheticals. #
@POST
excise(3,3);
excise(1,1);
@RULES
_xNIL <-
\{
_xANY
\}
@@
# Fix up. #
# 'd
# ' d
@POST
L("tmp1") = N(1);
if (vconjq(N(3),"inf"))
{
group(1,1,"_modal");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"MD");
N("sem",1) = N("stem",1) = "will";
N("text",1) = "would";
}
else if (vconjq(N(3),"-en"))
{
group(1,1,"_have");
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"VBD");
N("sem",1) = N("stem",1) = "had";
N("text",1) = "have";
N("verb",1) = 1;
N("pos num",1) = 1;
}
@RULES
_xNIL <-
_aposD
_xWILD [star match=(_adv _advl)]
_xWILD [one lookahead match=(_verb)]
@@
# aposD alpha
@PRE
<3,3> var("verb");
@CHECK
if (vconjq(N(3),"inf"))
{
S("inf") = 1;
succeed();
}
if (vconjq(N(3),"-en"))
{
S("-en") = 1;
succeed();
}
fail(); # Hopefully another part of speech...
@POST
L("tmp1") = N(1);
if (S("inf"))
{
alphatoverb(3,"active","VB");
group(1,1,"_modal");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"MD");
N("sem",1) = N("stem",1) = "will";
N("text",1) = "would";
}
else # S("-en")
{
alphatoverb(3,"passive","VBN");
group(1,1,"_have");
group(1,1,"_verb");
pncopyvars(L("tmp1"),N(1));
chpos(N(1),"VBD");
N("sem",1) = N("stem",1) = "had";
N("text",1) = "have";
N("verb",1) = 1;
N("pos num",1) = 1;
}
@RULES
_xNIL <-
_aposD
_xWILD [star match=(_adv _advl)]
_xALPHA
@@
# her
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"PP$");
@RULES
_xNIL <-
her [s]
_xWILD [one lookahead match=(_adj _noun)]
@@
# 's own
# own
@CHECK
if (N("mypos",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"JJ");
@RULES
_xNIL <-
_aposS
own
@@
# due num
# (which is) due (in the year)
@POST
L("tmp1") = N(1);
L("tmp2") = N(2);
group(2,2,"_num");
pncopyvars(L("tmp2"),N(2));
nountonp(2,1);
group(1,1,"_adj");
pncopyvars(L("tmp1"),N(1));
group(1,2,"_advl"); # Kind of an adverbial idiom.
@RULES
_xNIL <-
due
_xNUM
@@
# due direction
@POST
alphatoadv(1);
@RULES
_xNIL <-
due
_xWILD [s one lookahead match=(north south east west)]
@@
# from alpha to alpha
@CHECK
if (!N("noun",2) || !N("noun",4))
fail();
@POST
alphatonoun(2);
alphatonoun(4);
@RULES
_xNIL <-
from [s]
_xALPHA
to [s]
_xALPHA
@@
# equal to
@POST
alphatoadj(1);
@RULES
_xNIL <-
equal
to [s lookahead]
@@
# all alpha
@CHECK
if (!N("verb",2))
fail();
if (!vconjq(N(2),"inf"))
fail();
@POST
alphatoverb(2,"active","VBP");
@RULES
_xNIL <-
all [s]
_xALPHA
@@
# det first
@POST
L("tmp2") = N(2);
group(2,2,"_quan");
pncopyvars(L("tmp2"),N(2));
chpos(N(2),"JJ");
@RULES
_xNIL <-
_xWILD [one match=(_det _pro)]
first
_xWILD [one lookahead match=(_quan _num _xNUM _adj _noun)]
@@
# each increased by 10.
@CHECK
if (!N("verb",2))
fail();
@POST
alphatoverb(2,0,"VBP");
@RULES
_xNIL <-
each [s]
_xALPHA
_xWILD [s one match=(by to from)]
_xWILD [one match=(_noun _np)]
@@
# Semantics: scalar-verb followed by number.
@PRE
<1,1> var("verb");
<2,2> var("num");
@CHECK
if (!nodefeat(N(1),"vscalenum"))
fail();
@POST
alphatoverb(1,0,0);
N("done",1) = 1;
@RULES
_xNIL <-
_xALPHA
_noun [lookahead]
@@
# alpha to
# close to
# Note: adjective phrase idioms.
@POST
alphatoadj(1);
group(1,1,"_adjc");
@RULES
_xNIL <-
_xWILD [one match=(close)]
to [s lookahead]
@@
# who verb
@POST
fixverb(N(2),"active","VBP");
@RULES
_xNIL <-
who [s]
_verb
@@
# far higher
# much higher.
# alpha adj
@POST
alphatoadv(1);
group(1,2,"_adj");
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_xWILD [one match=(far much)]
_adj # FIX. # 04/15/10 AM.
@@
# no sooner
@POST
chpos(N(1),"RB");
chpos(N(2),"RB");
@RULES
_xNIL <-
no [s]
_xWILD [s one lookahead match=(sooner later fewer more less earlier later
closer farther further nearer
_adv
)]
@@
# sooner or later
@POST
chpos(N(1),"RB");
chpos(N(3),"RB");
@RULES
_xNIL <-
sooner [s]
or [s]
later [s]
@@
# based in geoloc
@PRE
<4,4> vareq("sem","geoloc");
@POST
alphatovg(2,"passive","VBN");
nountonp(4,1);
group(3,4,"_advl");
group(1,4,"_advl");
@RULES
_xNIL <-
\, [opt]
based
_prep
_noun
\, [opt]
@@
# the other
# Note: det, like "another".
# Note: But conform to Treebank. #
@POST
group(1,2,"_det");
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_det
other [s]
_xWILD [one lookahead match=(_quan _num _xNUM _adj _noun)]
@@
# det adj verb alpha
# det adv verb alpha
@PRE
<4,4> var("noun");
@CHECK
if (!vconjq(N(3),"-en") && !vconjq(N(3),"-ing"))
fail();
@POST
pnrename(N(3),"_adj"); # verb -> adj
if (!N("adj",3))
{
if (vconjq(N(3),"-en"))
N("mypos",3) = "VBN"; # Conform Treebank.
else
N("mypos",3) = "VBG"; # Conform Treebank.
}
N("dom200 verb-to-adj",3) = 1;
@RULES
_xNIL <-
_det [opt]
_xWILD [plus match=(_adj _adv _premod)]
_verb
_xWILD [one lookahead match=(_noun _xALPHA)]
@@
# det other
# the other
@POST
group(1,2,"_np");
N("ignorepos",1) = 1;
@RULES
_xNIL <-
_det
other [s]
_xWILD [one lookahead fail=(_xALPHA _quan _num _xNUM _adj _noun)]
@@
# to subject
@POST
alphatoverb(3,"active","VB");
@RULES
_xNIL <-
to [s]
_xWILD [star match=(_adv _advl)]
subject
@@
# subject to
@POST
alphatoadj(1);
@RULES
_xNIL <-
subject
_xWILD [star lookahead match=(_adv _advl)]
to [s]
@@
# plus
@POST
if (N("conj",3))
{
--N("pos num",3);
N("conj",3) = 0;
alphaunambigred(3);
}
@RULES
_xNIL <-
_det
_adj [star]
plus
@@
# plus
@POST
alphatoadj(2);
@RULES
_xNIL <-
\-
plus
@@
# plus
@POST
L("tmp1") = N(1);
group(1,1,"_conj");
pncopyvars(L("tmp1"),N(1));
N("mypos",1) = "CC"; # plus/CC # 04/22/07 AM.
@RULES
_xNIL <-
plus
@@
# up to
@CHECK
if (N("mypos",1))
fail();
@POST
chpos(N(1),"IN"); # up/IN
@RULES
_xNIL <-
up [s]
to [s lookahead]
@@
# Note: Fix up a noun meaning for VING/VBG that need it.
# prep alpha verb verb
@PRE
<2,2> var("adj");
@CHECK
if (!vconjq(N(3),"-ing"))
fail();
L("stem") = nodestem(N(2));
L("core") = finddictsattr(L("stem"),"core");
if (L("core") != "adj")
fail();
@POST
pnrename(N(3),"_noun"); # verb -> noun
fixnoun(N(3));
fixnpnonhead(2);
@RULES
_xNIL <-
_prep
_xALPHA
_verb
_xWILD [s one lookahead match=(_be _have _modal)]
@@
# 5 times stronger
@PRE
<2,2> var("noun");
<3,3> var("adj");
@CHECK
# Need comparative adj.
L("w") = strtolower(N("$text",3));
L("c") = adjconj(L("w"));
if (L("c") != "JJR")
fail();
@POST
alphatoadj(3);
group(3,3,"_adjc");
fixnphead(2);
N("sem",2) = "scalar";
@RULES
_xNIL <-
_xWILD [one match=(_num _xNUM)]
_xALPHA
_xALPHA
@@
# all together
# together
@POST
alphatoadv(2); # together/RB
if (N(1))
{
chpos(N(1),"RB"); # all/RB
pnrename(N(1),"_adv"); # [all] det -> adv
}
@RULES
_xNIL <-
all [s opt]
together
@@
# outside det
@POST
pnrename(N(1),"_prep");
N("mypos",1) = "IN";
@RULES
_xNIL <-
outside [s]
_xWILD [one lookahead match=(_det _quan _num _xNUM)]
@@
# det adj prep
@PRE
<2,2> var("num");
<2,2> varz("mypos");
@POST
N("mypos",2) = "NN";
pnrename(N(2),"_noun"); # adj -> noun
@RULES
_xNIL <-
_det
_adj
_xWILD [one lookahead match=(_prep)]
@@
# giant cap
@POST
N("mypos",1) = "NN";
@RULES
_xNIL <-
giant [s]
_xCAP [s lookahead]
@@
# number of
# Note: literal word 'number'
@POST
fixnphead(1);
@RULES
_xNIL <-
number
_xWILD [one lookahead match=(_prep _verb)]
@@
# alpha num
# late num
# early num
@CHECK
S("num") = num(N("$text",2));
if (S("num") < 1500 || S("num") > 2099)
fail();
@POST
alphatoadj(1);
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
N("num",2) = S("num");
N("sem",2) = "date";
chpos(N(2),"CD"); # Conform Treebank.
++X("date ref");
++X("date=past"); # Assume a past date.
@RULES
_xNIL <-
_xWILD [one match=(early late)]
_xNUM
@@
# much more
# much less
@POST
alphatoadv(1);
@RULES
_xNIL <-
much
_xWILD [s one lookahead match=(more less)]
@@
# had better be
# Note: idiom.
@POST
chpos(N(1),"VBD"); # had/VBD.
chpos(N(2),"RBR"); # better/RBR
chpos(N(3),"VB"); # be/VB
group(1,3,"_be"); # Effectively.
group(1,1,"_verb");
N("voice",1) = "active";
N("copula",1) = 1;
N("ignorepos",1) = 1;
@RULES
_xNIL <-
had [s]
better
be [s]
@@
# that modal
@POST
chpos(N(2),"DT"); # that/DT
pnrename(N(2),"_noun");
@RULES
_xNIL <-
_xWILD [one fail=(_noun _np _xALPHA)]
that [s]
_modal [lookahead]
@@
# last spring
# last winter
# last fall
# last summer
@POST
fixnphead(2);
alphatoadj(1);
@RULES
_xNIL <-
last
_xWILD [one match=(fall winter spring summer autumn)]
@@
# conj opened prep noun
# Note: "opened" has a preference for active voice...
@PRE
<4,4> vareq("sem","date");
@POST
alphatoverb(2,"active","VBD");
@RULES
_xNIL <-
_conj
opened
_prep [lookahead]
_noun
@@
# neither noun $
@POST
chpos(N(1),"DT"); # neither/DT
pnrename(N(1),"_det");
@RULES
_xNIL <-
neither [s]
_xWILD [one lookahead match=(_noun _np)]
_xWILD [one match=(_qEOS _dbldash _xEND)]
@@
# prep less than alpha alpha
@PRE
<5,5> var("noun");
@CHECK
if (!N("noun",4) && !N("adj",4))
fail();
@POST
fixnphead(5);
fixnpnonhead(4);
@RULES
_xNIL <-
_prep
_xWILD [s one match=(more less)]
than [s]
_xALPHA
_xALPHA
_xWILD [one match=(_qEOS _dbldash _xEND)]
@@
# modal still alpha adj
@PRE
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"inf"))
fail();
@POST
alphatoverb(3,"active","VB");
alphatoadv(2);
@RULES
_xNIL <-
_modal
still
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_adj _noun _np _prep _conj)]
@@
# still adj
@POST
alphatoadv(1);
@RULES
_xNIL <-
still
_adj [lookahead]
@@
# to alpha
# Note: to ving -- assume it's a verb.
@PRE
<2,2> var("verb");
@CHECK
if (!vconjq(N(2),"-ing"))
fail();
@POST
alphatovg(2,"active","VBG");
@RULES
_xNIL <-
to [s]
_xALPHA
_xWILD [one lookahead match=(_det _quan _num _xNUM
_adj _noun _np)]
@@
# until pro
@POST
pnrename(N(1),"_fnword");
chpos(N(1),"IN"); # until/IN
@RULES
_xNIL <-
until [s]
_proSubj [s one lookahead]
@@
# numbered about
@PRE
<1,1> vareq("stem","number");
@POST
chpos(N(3),"RB"); # about/RB
@RULES
_xNIL <-
_verb
_adv [star]
about [s]
@@
# Apposition across "sentences".
# noun : alpha conj
@PRE
<3,3> var("noun");
@POST
fixnphead(3);
@RULES
_xNIL <-
_noun
\: [s]
_xALPHA
_conj [lookahead]
@@
# fnword det adj verb
# Note: Fixing plurality.
@PRE
<4,4> var("cap");
@CHECK
if (numberv(N(5)) != "plural")
fail();
@POST
chpos(N(4),"NPS");
@RULES
_xNIL <-
_fnword
_det [star lookahead]
_xWILD [star match=(_quan _num _xNUM)]
_adj
_verb
@@
# give way
# Note: idiom
@POST
fixnphead(2);
alphatoverb(1,"active",0);
@RULES
_xNIL <-
give
way
to [s]
@@
# but $
@POST
pnrename(N(1),"_adv");
chpos(N(1),"RB"); # but/RB
@RULES
_xNIL <-
but [s]
_xWILD [one match=(_qEOS _xEND )]
@@
# due date
@PRE
<2,2> vareq("sem","date");
@POST
N("mypos",1) = "JJ"; # due/JJ
group(1,1,"_adjc");
@RULES
_xNIL <-
due [s]
_xWILD [one lookahead match=(_noun _np)]
@@
# due [DEFAULT
@POST
alphatoadj(1);
@RULES
_xNIL <-
due
@@
# age num
@POST
L("t") = strtolower(N("$text",2));
if (L("t") == "age")
chpos(N(2),"NN"); # age/NN
else
chpos(N(2),"JJ"); # aged/JJ
group(2,3,"_adjc");
N("sem",2) = "specific_age";
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xWILD [one match=(age aged)]
_xWILD [one match=(_quan _num _xNUM)]
@@
# set noun
# set date.
@PRE
<2,2> vareq("sem","date");
@POST
alphatoverb(1,"passive","VBN");
@RULES
_xNIL <-
set [trigger]
_xWILD [one lookahead match=(_noun _np)]
@@
# verb home
@POST
alphatoadv(2);
@RULES
_xNIL <-
_verb
home
@@
# Note: Idioms like "word for word", "step by "step".
@PRE
<1,1> var("noun");
<3,3> var("noun");
@CHECK
L("t1") = strtolower(N("$text",1));
L("t3") = strtolower(N("$text",3));
if (L("t1") != L("t3"))
fail();
@POST
fixnphead(3);
fixnphead(1);
group(1,3,"_noun");
@RULES
_xNIL <-
_xALPHA
_prep
_xALPHA
@@
# her det
@PRE
<1,1> varz("mypos");
@POST
chpos(N(1),"PP"); # her/PP
nountonp(1,1);
@RULES
_xNIL <-
her [s]
_det [lookahead]
@@
# alpha ones
@PRE
<1,1> var("adj");
@POST
alphatoadj(1);
@RULES
_xNIL <-
_xALPHA
ones [s lookahead]
@@
# alpha alpha
# verb adv
# Note: Common idioms.
@POST
alphatoadv(2);
alphatoverb(1,0,0);
@RULES
_xNIL <- give away @@
# where alpha
@PRE
<2,2> var("verb");
@POST
N("verb",2) = 0;
--N("pos num",2);
alphaunambigred(2);
@RULES
_xNIL <-
where [s]
_xALPHA
@@
# alpha still alpha
# Note: Semantics needed really, cuz 'still' is ambiguous.
@PRE
<1,1> var("noun");
<3,3> var("verb");
@POST
alphatoadv(2);
@RULES
_xNIL <-
_xWILD [one match=(_noun _xALPHA)]
still [trigger]
_xWILD [one lookahead match=(_xALPHA _verb _modal _vg)]
@@
# conj fast verb
@POST
alphatoadv(2);
@RULES
_xNIL <-
_xWILD [one match=(_conj _prep)]
fast
_xWILD [one lookahead match=(_verb _vg _modal)]
@@
# noun alpha
# date alpha
@PRE
<1,1> vareq("sem","date");
@CHECK
if (!N("adj",2) && !N("adv",2))
fail();
if (nodefeat(N(2),"-er"))
succeed();
L("t") = nodetext(N(2));
if (adjconj(L("t")) == "JJR")
succeed();
if (advconj(L("t")) == "RBR")
succeed();
fail();
@POST
if (N("adv",2))
alphatoadv(2);
else
alphatoadj(2);
@RULES
_xNIL <-
_xWILD [one match=(_noun _np)]
_xALPHA
@@
# pro noun prep
# Note: Maybe a missing verb pos.
@POST
pnrename(N(2),"_verb"); # noun -> verb
if (singular(N(2)))
fixverb(N(2),"active","VBP");
else
fixverb(N(2),"active","VBZ");
chpos(N(3),"RP");
"err.txt" << "Missing verb pos for '" << N("$text",2) << "' ?"
<< "\n";
@RULES
_xNIL <-
_proSubj [s]
_noun
_xWILD [s one match=(up down out)]
@@
# other alpha*
@CHECK
if (N("noun",2) && plural(N(2)))
S("o") = 2;
else if (N("noun",3) && plural(N(3)))
S("o") = 3;
else if (N(4))
{
if (N("noun",4) && plural(N(4)))
S("o") = 4;
}
if (!S("o"))
fail();
@POST
fixnphead(S("o"));
@RULES
_xNIL <-
other [s]
_xALPHA
_xALPHA
_xALPHA [opt]
@@
# half of
@POST
chpos(N(1),"NN");
if (literal(N(1)))
alphatonoun(1);
else
pnrename(N(1),"_noun"); # half -> noun
@RULES
_xNIL <-
half [s]
of [s lookahead]
@@
# Zap double quotes around any single item.
# " any "
@POST
L("s") = eltnode(2);
L("e") = lasteltnode(2);
if (L("s") == L("e"))
{
excise(3,3);
excise(1,1);
}
# Else pass this pair by.
@RULES
_xNIL <-
_dblquote
_xWILD [plus]
_dblquote
@@
# Some unknown-word logic.
@PRE
<1,1> var("unknown");
<1,1> vareq("pos num",0);
@CHECK
L("t") = nodetext(N(1));
if (!strendswith(L("t"),"ings"))
fail();
@POST
alphatonoun(1);
N("number",1) = "plural";
N("pos num",1) = 1;
N("noun") = 1;
@RULES
_xNIL <-
_xALPHA
@@
# alpha pro
@PRE
<1,1> var("verb");
<2,2> var("proobj");
@POST
alphatoverb(1,0,0);
@RULES
_xNIL <-
_xALPHA
_pro [lookahead]
@@
# worthy of
@POST
alphatoadj(1); # worthy/JJ.
@RULES
_xNIL <-
worthy
of [s lookahead]
@@
# by using
@PRE
<1,1> vareq("stem","by");
@POST
alphatovg(2,"active","VBG");
@RULES
_xNIL <-
_prep
using
@@
# no longer
# Note: IDIOM.
@POST
chpos(N(1),"RB"); # no/RB
chpos(N(2),"RB"); # longer/RB
group(1,2,"_adv");
N("no-longer",1) = 1;
@RULES
_xNIL <-
no [s]
longer [s]
@@
# based prep
# Note: Need features of verb + particle dealing with passive.
# base X on should be on/RP.
# past tense should be the same, perhaps, but conforming to
# Treebank.
@POST
alphatoverb(1,"passive","VBN"); # based/VBN.
if (N("stem",2) != "to")
chpos(N(2),"IN"); # pre/IN. # Conform Treebank.
@RULES
_xNIL <-
based
_prep
@@
# first of all
@POST
alphatoadv(1);
group(1,3,"_adv");
@RULES
_xNIL <-
first
of [s]
all [s]
@@
# tender offer
# Note: Idiomatic business language.
@POST
N("pos_np",1) = "NN";
@RULES
_xNIL <-
tender [s]
_xWILD [s one lookahead match=(offer offers)]
@@
# alpha activity
# Note: Want to capture the adj/verb sense of alpha,
# if possible. [Semantics]
# Eg: building activity (not the activity of builsings)
@PRE
<1,1> var("adj");
<1,1> var("verb");
@CHECK
if (!vconjq(N(1),"-ing"))
fail();
@POST
N("pos_np",1) = "VBG";
@RULES
_xNIL <-
_xALPHA
_xWILD [s one match=(activity activities)]
@@
# in charge of
# Note: Idiom.
@POST
chpos(N(1),"IN"); # in/IN.
alphatonoun(2);
@RULES
_xNIL <-
in [s]
charge # FIX. # 04/15/10 AM.
of [s lookahead]
@@
|
@NODES _humanName
# Just some samples here. These could be done by a RUG hierarchy!!
# If name is unlabeled, meaning first middle, last haven't been set.
@PRE
<2,2> lengthr(2,20) # Title must be length > 1.
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("prefixName") = N("$text", 2);
X("firstName") = N("$text", 5);
X("middleName") = N("$text", 8);
X("lastName") = N("$text", 11);
X("suffixName") = N("$text", 14);
# noop()
@RULES
# Dr. John A. Blipper [, Jr.]
_xNIL <-
_xSTART
_xALPHA [s layer=(_prefixName)]
\. [s]
_xWHITE [s star]
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_middleName)]
\. [s opt]
_xWHITE [s star]
_xWILD [s one match=(_surName _xALPHA) layer=(_lastName)]
# _xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@PRE
<2,2> lengthr(2,20) # Title must be length > 1.
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("prefixName") = N("$text", 2);
X("firstName") = N("$text", 5);
X("lastName") = N("$text", 8);
X("suffixName") = N("$text", 11);
# noop()
@RULES
# Dr. John Blipper [, Jr.]
_xNIL <-
_xSTART
_xALPHA [s layer=(_prefixName)]
\. [s]
_xWHITE [s star]
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xWILD [s one match=(_surName _xALPHA) layer=(_lastName)]
# _xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("firstName") = N("$text", 2);
X("middleName") = N("$text", 5);
X("lastName") = N("$text", 8);
X("suffixName") = N("$text", 11);
# noop()
@RULES
_xNIL <-
_xSTART
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xALPHA [s layer=(_middleName)]
\. [s opt]
_xWHITE [s star]
_xWILD [s one match=(_surName _xALPHA) layer=(_lastName)]
# _xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
@CHECK
if (!X("unlabeled")) fail();
@POST
X("unlabeled") = 0;
X("firstName") = N("$text", 2);
X("lastName") = N("$text", 5);
X("suffixName") = N("$text", 8);
# noop();
@RULES
_xNIL <-
_xSTART
_xALPHA [s layer=(_firstName)]
\. [s opt]
_xWHITE [s star]
_xWILD [s one match=(_surName _xALPHA) layer=(_lastName)]
# _xALPHA [s layer=(_lastName)]
\, [s opt]
_xWHITE [s star]
_xALPHA [s opt layer=(_suffixName)]
_xEND
@@
|
@CODE
G("output") = G("$inputhead") + ".dict";
@@CODE |
@CODE
G("parse") = getconcept(findroot(),"parse");
G("verbs") = getconcept(findroot(),"verbs");
G("caps") = getconcept(findroot(),"caps");
G("names") = getconcept(findroot(),"names");
@@CODE |
@CODE
L("hello") = 0;
if (!G("dbl quotes"))
exitpass(); # No double quotes found.
@@CODE
@NODES _TEXTZONE
@POST
if (G("inquotes"))
G("inquotes") = 0;
else
G("inquotes") = 1;
@RULES
_xNIL <-
_dblquote
@@
@CHECK
if (!G("inquotes"))
fail();
@POST
N("quoted") = 1;
N("cap") = 1;
++X("quoted caps");
@RULES
_xNIL <-
_xCAP [s]
@@
# Should track title capitalization...
@CHECK
if (!G("inquotes"))
fail();
@POST
N("quoted") = 1;
++X("quoted prose");
@RULES
_xNIL <-
_xALPHA [s]
@@
@CHECK
if (!G("inquotes"))
fail();
@POST
N("quoted") = 1;
@RULES
_xNIL <-
_xANY
@@
|
Subsets and Splits