text
stringlengths 22
301k
|
|---|
@RULES
_ENDRULE [base] <- \@ \@ _xWHITE @@
#@POST
# noop()
#@RULES
#_xNIL <- _xWILD [min=1 max=1 fail=(\@)] @@
@RULES
_soRULES [base] <- \@ RULES [t] @@
_soPOST [base] <- \@ POST [t] @@
_soCHECK [base] <- \@ CHECK [t] @@
_soPRE [base] <- \@ PRE [t] @@
_soNODES [base] <- \@ NODES [t] @@
_soMULTI [base] <- \@ MULTI [t] @@
_soPATH [base] <- \@ PATH [t] @@
_soCODE [base] <- \@ CODE [t] @@
_soSELECT [base] <- \@ SELECT [t] @@
_soRECURSE [base] <- \@ RECURSE [t] @@
_eoPOST [base] <- \@ \@ POST [t] @@
_eoCHECK [base] <- \@ \@ CHECK [t] @@
_eoPRE [base] <- \@ \@ PRE [t] @@
_eoRULES [base] <- \@ \@ RULES [t] @@
_eoRECURSE [base] <- \@ \@ RECURSE [t] @@
_eoSELECT [base] <- \@ \@ SELECT [t] @@
_eoNODES [base] <- \@ \@ NODES [t] @@
_eoMULTI [base] <- \@ \@ MULTI [t] @@
_eoPATH [base] <- \@ \@ PATH [t] @@
_eoCODE [base] <- \@ \@ CODE [t] @@
_ENDRULE [base] <- \@ \@ @@
@POST
rfanonlit(2)
single()
@RULES
_NONLIT [base] <- \_ _xALPHA @@
@RULES
_ARROW [base] <- \< \- @@
@POST
rfaname(1)
single()
@RULES
_LIT [base] <- _xALPHA @@
@POST
rfanum(1)
single()
@RULES
_NUM [base] <- _xNUM @@
|
# Fetch index entry for dictionary concept. Gets index concept that str would be added under.
L("return_con") = wordindex(L("str")); |
@DECL
Spaces(L("output"), L("num"), L("spaces")) {
"debug.txt" << "moose\n";
while (L("i") < L("num")) {
L("output") << L("spaces");
L("i")++;
}
}
@@DECL |
@DECL
VerticeXMLTree(
L("out"),
L("n"),
L("parent id")
)
{
L("childs") = pndown(L("n"));
L("leaf") = 0;
while (L("childs")) {
L("name") = pnname(L("childs"));
if (strlength(L("name")) > 1) {
L("id") = G("id")++;
L("tag") = strtrim(strpiece(L("name"),0,strlength(L("name"))-1));
L("out") << "<vertice";
VerticeGetAttributes(L("out"),L("childs"));
L("out") << ">\n";
L("out") << " <id>" << L("id") << "</id>\n";
L("out") << " <label>" << L("tag") << "</label>\n";
L("out") << "</vertice>\n";
if (L("parent id")) {
L("out") << "<edge>\n";
L("out") << " <source>" << L("parent id") << "</source>\n";
L("out") << " <target>" << L("id") << "</target>\n";
L("out") << "</edge>\n";
}
if (pndown(L("childs"))) {
VerticeXMLTree(L("out"),L("childs"),L("id"));
}
}
L("childs") = pnnext(L("childs"));
}
return L("leaf");
}
VerticeXMLRecurseAll(
L("out"),
L("n"), # Current node.
L("level") # level of recursion
)
{
L("childs") = pndown(L("n"));
L("leaf") = 0;
while (L("childs")) {
L("name") = pnname(L("childs"));
if (strlength(L("name")) > 1) {
L("tag") = strpiece(L("name"),1,strlength(L("name"))-1);
if (pndown(L("childs"))) {
G("id")++;
L("out") << "\n" << VerticeSpacing(L("level")) << "<" << L("tag");
VerticeGetAttributes(L("out"),L("childs"));
L("out") << ">";
if (!VerticeXMLRecurseAll(L("out"),L("childs"),L("level")+1))
L("out") << "\n" << VerticeSpacing(L("level"));
L("out") << "</" << L("tag") << ">";
} else {
L("out") << pnvar(L("childs"),"$text");
L("leaf") = 1;
}
}
L("childs") = pnnext(L("childs"));
}
return L("leaf");
}
VerticeGetAttributes(L("out"),L("n")) {
VerticeAddAttribute(L("out"),L("n"),"stem");
VerticeAddAttribute(L("out"),L("n"),"voice");
VerticeAddAttribute(L("out"),L("n"),"tense");
VerticeAddAttribute(L("out"),L("n"),"aspect");
}
VerticeAddAttribute(L("out"),L("n"),L("name")) {
L("value") = pnvar(L("n"),L("name"));
if (L("value"))
L("out") << " " << L("name") << "=\"" << L("value") << "\"";
}
VerticeSpacing(L("num")) {
L("i") = 1;
L("VerticeSpacing") = " ";
while (L("i")++ < L("num")) {
L("VerticeSpacing") = L("VerticeSpacing") + " ";
}
return L("VerticeSpacing");
} |
@NODES _ROOT
@POST
N("wordcon") = dictfindword(strtolower(N("$text")));
if (N("wordcon"))
{
G("vals") = findvals(N("wordcon"),"pos");
G("first val") = "_" + strtoupper(getstrval(G("vals")));
N("pos num") = 0;
while (G("vals"))
{
G("val") = getstrval(G("vals"));
if (G("val") == "noun")
N("noun") = 1;
else if (G("val") == "verb")
N("verb") = 1;
else if (G("val") == "definite article")
N("definite article") = 1;
else if (G("val") == "indefinite article")
N("indefinite article") = 1;
else if (G("val") == "adjective")
N("adjective") = 1;
else if (G("val") == "adverb")
N("adverb") = 1;
else if (G("val") == "conj")
N("conj") = 1;
else if (G("val") == "pronoun")
N("pronoun") = 1;
else if (G("val") == "prep" || G("val") == "preposition")
N("prep") = 1;
else if (G("val") == "interj" || G("val") == "interjection")
N("interj") = 1;
else if (G("val") == "aux")
N("aux") = 1;
else if (G("val") == "pastpart")
N("pastpart") = 1;
G("vals") = nextval(G("vals"));
N("pos num")++;
}
if (N("pos num") > 1)
group(1,1,"_AMBIG");
else
group(1,1,strsubst(G("first val")," ","_"));
}
@RULES
_xNIL <-
_xALPHA [s] ### (1)
@@
|
# Fetch the last node in phrase.
L("return_con") = lastnode(L("phrase")); |
# Find attribute and value pair in the subhierarchy of a concept.
attrexists(L("hier"), L("attr_s"), L("val_s")); |
@NODES _ROOT
@RULES
_xNIL <-
_xWILD ### (1)
@@
|
@NODES _ROOT
@POST
L("abbrev") = getconcept(G("state"),"abbrev");
addstrval(L("abbrev"),"letters",N("text"));
G("abbrev") << strtolower(N("text")) << " s=state abbrev=1 state=" << QuoteIfNeeded(G("state name")) << "\n";
@RULES
_xNIL <-
_abbrev ### (1)
@@
|
@NODES _item
@POST
X("catid") = N("$text",2);
@RULES
_xNIL <-
_categoryidOpen [s] ### (1)
_xWILD [s fails=(_categoryidClose)] ### (2)
_categoryidClose ### (3)
@@
@POST
"names.txt" << N("$text",2) << "\n";
addCategory(N("$text",2),X("catid"));
X("name") = N("$text",2);
@RULES
_xNIL <-
_nameOpen [s] ### (1)
_xWILD [s fails=(_nameClose)] ### (2)
_nameClose ### (3)
@@
@PRE
<2,2> lengthr(2,100);
@POST
"tooltip.txt" << N("$text",2) << "\n";
@RULES
_xNIL <-
_tooltipOpen [s] ### (1)
_xWILD [s fails=(_tooltipClose)] ### (2)
_tooltipClose ### (3)
@@ |
@CODE
# If found new sentence breaks, do this pass.
if (!G("eos"))
exitpass();
@@CODE
#@PATH _ROOT _TEXTZONE _sent
@NODES _sent
@POST
splice(1,1);
@RULES
_xNIL <-
_EOS
@@
|
@CODE
closefile(G("file"));
@@CODE |
# Transfer confidence information from human candidate node to
# a human name that was found inside it.
@NODES _ROOT _contactZone _LINE _humanNameCandidate
@CHECK
if (X("ResumeOf") <= 0) fail();
@POST
N("ResumeOf") = X("ResumeOf"); # 12/25/99 AM.
# ninc(1, "ResumeOf")
noop();
@RULES
_xNIL <-
_humanName [s]
@@
|
@NODES _LINE
@PRE
<3,3> length(1);
<6,6> length(3);
<8,8> length(3);
<10,10> length(4);
@RULES
# Ex: (+1)\_
_phoneNumber <- _xWILD [min=1 max=1 s match=("_openPunct" "\(")] \+ [s] _xWILD [min=1 max=1 trig s layer=("_countryCode") match=("_monthNum" "1")] _xWILD [min=1 max=1 s match=("_closePunct" "\)")] _xWHITE [star s] _xNUM [s layer=("_areaCode")] _xWHITE [star s] _xNUM [s layer=("_prefix")] _xWHITE [star s] _xNUM [s layer=("_suffix")] @@
@PRE
<7,7> length(4);
@RULES
_phoneNumber <- _xNUM [s layer=("_countryCode")] \- [trig s] _xNUM [s layer=("_areaCode")] \- [s] _xNUM [s layer=("_prefix")] \- [s] _xNUM [s layer=("_suffix")] @@
@PRE
<2,2> length(3);
<5,5> length(3);
<9,9> length(4);
@RULES
_phoneNumber <- _xWILD [min=1 max=1 s match=("_openPunct" "\(")] _xNUM [s layer=("_areaCode")] _xWILD [min=1 max=1 s match=("_closePunct" "\)")] _xWHITE [star s] _xNUM [s layer=("_prefix")] _xWHITE [star s] \- [s] _xWHITE [star s] _xNUM [s layer=("_suffix")] @@
@PRE
<1,1> length(3);
<3,3> length(3);
<5,5> length(4);
@RULES
_phoneNumber <- _xNUM [s layer=("_areaCode")] _xPUNCT [trig s] _xNUM [s layer=("_prefix")] _xPUNCT [s] _xNUM [s layer=("_suffix")] @@
@PRE
<1,1> length(2);
<3,3> length(3);
<5,5> length(6);
@RULES
_phoneNumber <- _xWILD [min=1 max=1 s layer=("_countryCode") match=("_year" "61")] \- [s] 411 [trig s layer=("_areaCode")] \- [s] 348700 [s layer=("_suffix")] @@
@PRE
<1,1> length(3);
<3,3> length(3);
<5,5> length(4);
@RULES
_phoneNumber <- _xNUM [s layer=("_areaCode")] _xWHITE [star s] _xNUM [s layer=("_prefix")] \- [trig s] _xNUM [s layer=("_suffix")] @@
_phoneNumber <- _xNUM [s layer=("_areaCode")] _xWHITE [star s] _xNUM [s layer=("_prefix")] _xWHITE [star s] _xNUM [s layer=("_suffix")] @@
@PRE
<2,2> length(3);
<5,5> length(3);
<7,7> length(4);
@RULES
_phoneNumber <- _xWILD [min=1 max=1 s match=("_openPunct" "\(")] _xNUM [s layer=("_areaCode")] _xWILD [min=1 max=1 s match=("_closePunct" "\)")] _xWHITE [star s] _xNUM [s layer=("_prefix")] _xWHITE [star s] _xNUM [s layer=("_suffix")] @@
_phoneNumber <- _xWILD [min=1 max=1 s match=("_openPunct" "\(")] 408 [trig s] _xWILD [min=1 max=1 s match=("_closePunct" "\)")] _xWHITE [star s] 283 [s] \- [s] 0605 [s] @@
|
@NODES _LINE
@RULES
_experiencePart <- _DateRange [s] @@
# Ex: Senior\_Software\_Engineer
_experiencePart <- _xWILD [min=1 max=1 s match=("_Caps" "_jobTitle")] @@
|
@CODE
DisplayKB(G("parse"),2);
@@CODE |
@PATH _ROOT _RULES
@RECURSE littopair
@POST
rfalittopair(1)
single()
@RULES
_PAIR <- _LIT @@
@@RECURSE littopair
@POST
rfapairs(2)
single()
@RULES
_PAIRS [base] <- \[ _xWILD [match=(_LIT _PAIR \*) recurse=(littopair)] \] @@
|
# Find the kb root, then make a child concept under root named "child"
G("the root") = findroot();
# make the child concept
makeconcept(G("the root"),"child");
# do something else ...
# later on, get the string variant
G("the child") = findconcept(G("the root"),"child");
# fetch the numeric variant
G("same concept") = findconcept(G("the root"), 1); |
@NODES _ROOT
@POST
if ()
single();
@RULES
_xNIL <-
_labelEntry ### (1)
@@
|
@NODES _LINE
@POST
singler(3,3)
@RULES
_CompleteSchoolName [base] <-
_xWILD [s one match=( _xSTART The _xPUNCT )]
_xWHITE [star s]
_SchoolNamePhrase [t]
@@
# Joe (SchoolNamePhrase University) -> ModSchoolPhrase(Joe University)
@RULES
_SchoolNamePhrase [base] <-
_xWILD [one s fails=( _xPUNCT _xWHITE _xNUM
_degree _major _minor _gpa
_DateRange _SingleDate)] # 10/09/99 PS
_xWHITE [one s] # 10/07/99 AM.
_SchoolNamePhrase [t]
@@
|
@DECL
AddWord(L("person"),L("word")) {
"word.txt" << L("person") << " " << L("word") << "\n";
DictionaryWord(G("word"),"person",L("person"),"str");
}
@@DECL |
@NODES _ROOT
@RULES
_paragraph <-
_xWILD [plus match=(_LINE)] ### (2)
@@
|
@NODES _ROOT
@POST
excise(1,4);
noop();
@RULES
_xNIL <-
_xSTART
_xWILD [fail=(_xALHPA _xPUNCT)]
_xWILD [one matches=(_xALHPA _xPUNCT)]
_xWILD [fail=(\n \r)] ### (1)
_xWILD [one matches=(\n \r)] ### (3)
@@
@POST
singler(3,3);
@RULES
_term <-
_xNUM
_xWILD [fail=(_xALPHA)]
_xWILD [fail=(\n \r \")] ### (1)
_xWILD [one matches=(\" \n \r _xEND)] ### (3)
@@ |
@NODES _ROOT
@RULES
_paragraph [unsealed] <-
_sentence [s plus] ### (1)
_xWILD [s matches=(_paragraphSeparator _xEND)] ### (2)
@@ |
# Match a _verb node optionally preceded by _modal, _have, or _be nodes, reducing to _vgroup
@RULES
_vgroup <- _modal [option]
_have [option] _be [option]
_verb @@ |
@NODES _ROOT
@PRE
<1,1> var("bases");
@POST
# Check for section or subsection title.
"extract_codes.txt" << "Beginning code extraction step." << ":\n";
L("section_title") = N("section_title", 1);
L("subsection_title") = N("subsection", 1);
G("line_count") = G("line_count") + 1;
L("section_name") = pnname(N(1)) + "_" + str(G("line_count"));
"extract_codes.txt" << "section_name=" << L("section_name") << ":\n";
L("section") = AddUniqueCon(G("note_words"), L("section_name"));
if (L("section_title")) {
addstrval(L("section"),"section",L("section_title"));
"extract_codes.txt" << "SECTION " << L("section_title") << ":\n";
}
if (L("subsection")) {
addstrval(L("section"),"subsection",L("subsection_title"));
"extract_codes.txt" << "SUBSECTION " << L("subsection") << ":\n";
}
# L("bases") = addattr(L("section"), "bases");
L("all_bases") = pnvar(N(1), "bases");
L("bases_len") = arraylength(L("all_bases"));
"extract_codes.txt" << "\t" << "Adding " << L("bases_len") << " total bases." << "\n";
# addstrval(L("section"),"bases", L("all_bases"));
L("i") = 0;
while(L("i") < L("bases_len")) {
"extract_codes.txt" << "\t" << "Adding " << L("all_bases")[L("i")] << "\n";
AddUniqueStr(L("section"),"bases", L("all_bases")[L("i")]);
"extract_codes.txt" << "\t" << conceptname(G("eui_to_base_term")) << "\n";
L("eui_con") = getconcept(G("eui_to_base_term"), L("all_bases")[L("i")])
L("eui_term") = strval(L("eui_con"), "term");
AddUniqueStr(L("section"), "terms", L("eui_term"));
L("i")++;
}
# addstrval(L("section"),"bases", N("bases", 1));
"extract_codes.txt" << "\t" << N("bases", 1) << "\n";
# extractCodeRanks(N("bases", 1), "diagnoses");
# extractCodeRanks(N("bases", 1), "procedures");
@RULES
_xNIL <-
_xWILD [one matches=( _section _subsection _sentence _looseText _item)]
@@ |
@NODES _ROOT
@PRE
<1,1> var("up");
@POST
splice(1,1);
@RULES
_xNIL <-
_LINE ### (1)
@@
|
@NODES _LINE
@PRE
<1,1> cap()
@RULES
# Ex: Sweden
_countryWord <- _xWILD [min=1 max=1 s match=(Sweden USA Canada Mexico France Australia ROC USSR PRC UAR UK Spain Italy Finland Turkey Israel Egypt Yugoslavia Japan Norway Denmark Holland Germany Austria Hungary Albania Portugal Poland Greece Czechoslovakia India Pakistan Rumania Romania Bulgaria US Latvia Estonia Switzerland Morocco Algeria Yemen Ethiopia Libya Palestine Iraq Iran Lebanon Venezuela Russia China Peru Chile Korea England Belize Paraguay Uruguay Argentina Brazil Ecuador Gabon Nigeria Sudan Indonesia Thailand Malaysia Singapore Laos Taiwan Ukraine)] @@
|
@NODES _ROOT
@POST
singler(1,2);
@RULES
_IGNORE <-
_xSTART ### (1)
_xWILD ### (2)
_beginAbs [lookahead] ### (3)
@@
@POST
singler(2,2);
@RULES
_IGNORE <-
_endAbs ### (1)
_xWILD ### (2)
_beginFigure ### (3)
@@
@POST
singler(2,2);
@RULES
_IGNORE <-
_endAbs ### (1)
_xWILD ### (2)
_section ### (3)
@@
|
@NODES _LINE
@PRE
<3,3> numrange(1, 31)
@RULES
# ambiguity, month-day-year reading is preferred.
_date [] <-
_monthNum [s layer = (_month) ]
\/ [s]
_xNUM [s layer = (_day) ]
\/ [s]
_year [s]
@@
_date [] <-
_monthNum [s layer = (_month) ]
\- [s]
_xNUM [s layer = (_day) ]
\- [s]
_year [s]
@@
_date [] <-
_monthNum [s layer = (_month) ]
\/ [s]
_xNUM [s layer = (_day) ]
\/ [s]
_year [s]
@@
@PRE
<1,1> numrange(1, 31)
@RULES
# ambiguity, month-day-year reading is preferred.
# here, there's no ambiguity,
_date [] <-
_xNUM [s layer = (_day) ]
\/ [s]
_monthNum [s layer = (_month) ]
\/ [s]
_year [s]
@@
_date [] <-
_xNUM [s layer = (_day) ]
\- [s]
_monthNum [s layer = (_month) ]
\- [s]
_year [s]
@@
_date [] <-
_xNUM [s layer = (_day) ]
\/ [s]
_monthNum [s layer = (_month) ]
\/ [s]
_year [s]
@@
@PRE
<1,1> numrange(1, 31)
@RULES
# Ex: 30 July(,.)
_date [] <-
_xNUM [s layer = (_day) ]
_xWHITE [s]
_xWILD [s matches = (_monthWord) ]
\. [s opt]
\, [s opt]
_xWHITE [s]
_year [s]
@@
@PRE
<4,4> numrange(1, 31)
@RULES
# Ex: July(.) 30,
_date [] <-
_xWILD [s matches = (_monthWord) ]
\. [s opt]
_xWHITE [s]
_xNUM [s layer = (_day) ]
\, [s opt]
_xWHITE [s]
_year [s]
@@ |
@NODES _ROOT
@PRE
<1,1> uppercase();
@RULES
_identifier <-
_xALPHA ### (1)
\: [one] ### (2)
_xALPHA
_xEND
@@
|
@PATH _ROOT _paragraph _sentence
@PRE
<1,1> var("human");
@POST
S("con") = getconcept(G("people"),strtolower(phrasetext()));
single();
@RULES
_person <-
_titleCaps ### (1)
@@
|
@CODE
L("hello") = 0;
@@CODE
@NODES _sent
# 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 _xEND)]
@@
# Not sure why alpha in first position in these rules....
@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 (N(4))
{
L("firstj") = N(4);
L("lastj") = lasteltnode(4);
}
groupnp();
@RULES
_xNIL <-
_xALPHA
_xWILD [plus match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS)]
@@
_xNIL <-
_xALPHA
_xWILD [star match=(_det _pro)]
_xWILD [plus match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS)]
@@
_xNIL <-
_xWILD [one fail=(_xALPHA _det _quan _num _xNUM _adj _noun)]
_xWILD [s star match=(_det _proPoss)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_adj _xALPHA _aposS)]
@@
@POST
L("tmp4") = lasteltnode(4);
pncopyvars(L("tmp4"));
sclearpos(1,0);
S("bracket",2) = 1;
S("ne") = 0;
singler(2,4);
@RULES
_np <-
_xSTART
_xWILD [star match=(_det _quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_noun _adj _xALPHA _aposS)]
@@
@POST
L("n") = lasteltnode(4);
pncopyvars(L("n"));
sclearpos(1,0); # Zero out token info.
singler(1,4);
@RULES
_np <-
_proPoss [s]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xWILD [one lookahead fail=(_noun _xALPHA)]
@@
@POST
pncopyvars(3);
sclearpos(1,0); # Zero out token info.
singler(2,3);
@RULES
_np <-
_xWILD [one fail=(_xALPHA _noun _det _quan _num _xNUM _aposS)]
_adj [plus]
_noun
_xWILD [one lookahead fail=(_noun _xALPHA)]
@@
# nounless.
@POST
L("tmp2") = lasteltnode(2);
pncopyvars(L("tmp2"));
sclearpos(1,0);
singler(1,2);
@RULES
_np <-
_xWILD [plus match=(_det _quan _num _xNUM)]
_adj [plus]
_xWILD [one lookahead fail=(_noun _adj _xALPHA _dbldash)]
@@
# vg part of
# idiomatic
@POST
L("tmp2") = N(2);
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
nountonp(2,1);
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
@RULES
_xNIL <-
_vg
_xWILD [s one match=(part parts)]
of [s lookahead]
@@
# det noun noun
@POST
pncopyvars(3);
sclearpos(1,0); # Zero out token info.
singler(1,3);
@RULES
_np <-
_det
_noun
_noun
_xWILD [lookahead one fail=(_noun _conj)]
@@
# prep noun alpha .
@CHECK
if (!N("noun",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); # Zero out token info.
@RULES
_xNIL <-
_prep
_noun
_xALPHA
_xWILD [one lookahead match=(_qEOS _xEND)]
@@
# det alpha
@CHECK
if (!N("noun",2) && !N("adj",2))
fail();
if (N("noun",2) && N("adj",2)) # Can't decide yet.
fail();
@POST
L("tmp2") = N(2);
if (N("noun",2))
group(2,2,"_noun");
else if (N("adj",2))
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
@RULES
_xNIL <-
_det
_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)]
@@
# pro
@POST
if (pnname(N(3)) == "_vg")
if (!N("voice",3))
N("voice",3) = "active";
if (pnname(N(2)) == "_pro")
{
nountonp(2,1);
}
@RULES
_xNIL <-
_xWILD [one match=(_xSTART _adv _advl _conj _fnword)]
_pro [s]
_xWILD [one lookahead match=(_vg)]
@@
# prep adj alpha
# dqa alpha
@CHECK
if (!N("noun",5) && !N("adj",5))
fail();
dqaninfo(2,3,4,5);
if (!numbersagree(S("first"),S("last")))
fail();
@POST
L("tmp5") = N(5);
group(5,5,"_noun");
pncopyvars(L("tmp5"),N(5));
if (!N(4))
S("firstan") = N(5);
S("firstn") = S("last") = S("lastn")
= S("lastan") = N(5); # Reset stuff.
groupnp();
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_prep _conj _verb _vg)]
_det [plus]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xWILD [one lookahead match=(\, _qEOS _xEND _dbldash _prepj)]
@@
_xNIL <-
_xWILD [one match=(_prep _conj _verb _vg)]
_det [star]
_xWILD [plus match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xWILD [one lookahead match=(\, _qEOS _xEND _dbldash _prep)]
@@
_xNIL <-
_xWILD [one match=(_prep _conj _verb _vg)]
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [plus]
_xALPHA
_xWILD [one lookahead match=(\, _qEOS _xEND _dbldash _prep)]
@@
# alpha alpha
# TOO BROAD.
@CHECK
if (N("verb",2) || N("verb",3))
fail();
if (!N("adj",2) && !N("noun",2))
fail();
if (!N("adj",3) && !N("noun",3))
fail();
@POST
L("tmp3") = N(3);
L("tmp2") = N(2);
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(3,3,"_noun");
pncopyvars(L("tmp3"),N(3));
group(2,3,"_np");
pncopyvars(L("tmp3"),N(2));
clearpos(N(2),1,1); # Zero out token info.
@RULES
_xNIL <-
_xWILD [one match=(\, _dbldash)]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep _conj)]
@@
# low for
@CHECK
if (N("adj",2) && N("noun",2))
fail(); # Can't decide here.
if (!N("adj",2) && !N("noun",2))
fail();
if (N("verb",2))
fail();
@POST
L("tmp2") = N(2);
if (N("noun",2))
group(2,2,"_noun");
else
group(2,2,"_adj");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(2)) == "_adj")
fixadj(N(2));
nountonp(2,1);
@RULES
_xNIL <-
_xWILD [one match=(\,)]
_xALPHA
_xWILD [one lookahead match=(_prep _conj)]
@@
# verb alpha
@CHECK
if (N("pos num",2) != 2)
fail();
if (!N("verb",2))
fail();
if (!vconjq(N(1),"inf"))
fail();
# Check for helping verbs etc.
@POST
L("tmp2") = N(2);
if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
_verb
_xALPHA
@@
# well over
# well under
@POST
L("tmp2") = N(2);
group(2,2,"_adv");
pncopyvars(L("tmp2"),N(2));
if (pnname(N(1)) == "_vg")
if (!N("voice",1))
N("voice",1) = "active";
@RULES
_xNIL <-
_xWILD [one match=(_verb _vg)]
well
_xWILD [s one lookahead match=(over under)]
@@
# _adjc is constituent level concept, like _vg, _np.
# vg adj
@POST
L("tmp3") = N(3);
group(3,4,"_adjc");
pncopyvars(L("tmp3"),N(3));
N("ignorepos",3) = 1; # commented 04/19/07 AM. (why)
N("posarr len",3) = 0; # 04/21/07 AM.
@RULES
_xNIL <-
_xWILD [one match=(_noun _np _vg)]
_xWILD [star match=(_advl _adv)]
_adj
_xWILD [star match=(_advl _adv)]
_xWILD [one lookahead match=(_prep _fnword _qEOS _xEND \,)]
@@
@POST
singler(2,7);
@RULES
_np <-
_xWILD [one match=(_xSTART)]
_det [star]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_adj [plus]
_noun [plus]
_xWILD [one fail=(_xALPHA _noun _adj)]
@@
# verb alpha dqan
# vg alpha dqan
@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 if (N("noun",2))
{
group(2,2,"_noun");
pncopyvars(L("tmp2"),N(2));
fixnoun(N(2));
}
@RULES
_xNIL <-
_xWILD [one match=(_vg _verb)]
_xALPHA
_noun [plus]
@@
# Convert pronouns at some point.
@PRE
<1,1> varz("proposs");
@CHECK
if (pnname(N(1)) == "_np")
fail();
@POST
nountonp(1,1);
@RULES
_xNIL <-
_xWILD [s one match=(_proSubj _proObj) except=(_whword)]
@@
|
# Count nouns with variable "number" with value not equal to "plural"
@PRE
<1,1> varne("number","plural");
@POST
++G("count nonplural nouns");
@RULES
_xNIL <-
_noun
@@ |
@NODES _LINE
@RULES
_jobTitle <- _jobTitleRoot [s]
_xWHITE [s]
_jobTitleRoot [s] @@
_jobTitle <- _fieldName [s]
_xWHITE [s]
_jobTitleRoot [s] @@ |
# Fetch the string value from a KB value object.
L("return_str") = getstrval(L("val")); |
# Move concept before previous sibling. (Moves concept to the left in its list.)
movecleft(L("con")); |
@CODE
DisplayKB(findconcept(findroot(), "phrases"), 1);
@@CODE |
@NODES _LINE
@POST
X("word") = N("$text",2);
single();
@RULES
_word <-
_xSTART
_xALPHA ### (1)
@@
@POST
X("pos") = N("$text",3);
single();
@RULES
_pos <-
pos
\=
_xALPHA ### (1)
@@
|
# Rename con's named attribute to new.
renameattr(L("con"), L("name"), L("new")); |
@DECL
# Take KB string values and convert to array
# Args
# first_val: first value in kb value list
StrValsToArray(L("first_val")) {
L("arr");
L("arr_idx") = 0;
while (L("first_val")) {
L("arr")[L("arr_idx")] = getstrval(L("first_val"));
L("arr_idx")++;
L("first_val") = nextval(L("first_val"));
}
return L("arr");
}
# Get count of values in attribute list.
ValsLength(L("first_val")) {
L("arr_len") = 0;
while (L("first_val")) {
L("arr_len")++;
L("first_val") = nextval(L("first_val"));
}
return L("arr_len");
}
@@DECL |
@DECL
IsPOS(L("node"), L("pos")) {
L("vars") = pnvarnames(L("node"));
L("i") = 0;
while (L("i") < arraylength(L("vars"))) {
L("var") = L("vars")[L("i")];
if (strcontains(L("pos"), L("var"))) {
return 1;
}
L("i")++;
}
return 0;
}
@@DECL |
@NODES _LINE
@PRE
<1,1> cap();
<3,3> cap();
@RULES
# Ex: Professional\_Objective
_ObjectiveHeaderPhrase [layer=(_header )] <- Professional [s] _xWHITE [star s] Objective [s] @@
|
@CODE
DictionaryStart();
DictionaryWord("he","gender","male","str");
DictionaryWord("she","gender","female","str");
DictionaryWord("january","date","month","str");
DictionaryWord("january","number","1","num");
DictionaryEnd();
@@CODE |
# Same as addword; more principled function name.
L("return_con") = dictgetword(L("str")); |
@NODES _ROOT
@POST
singler(2,2);
@RULES
_entries <-
_split ### (1)
_xWILD [fails=(_split)] ### (2)
@@
|
@NODES _term
@POST
excise(1,1);
@RULES
_xNIL <-
_xWHITE
@@
|
@NODES _LINE
# Joe (SchoolNamePhrase University) -> ModSchoolPhrase(Joe University)
@RULES
_SchoolNamePhrase [base] <-
_xWILD
[one s fails=(
_xPUNCT _xWHITE _xNUM _degree _degreePhrase
_major _minor _gpa from at in by
_DateRange _SingleDate _CityState )] # 10/09/99 PS
_xWHITE [one s] # 10/07/99 AM.
_xWILD [t one match=( _SchoolNamePhrase _SchoolRoot )]
@@
_CompleteSchoolName [] <- _CityState [s]
_xWHITE [s]
_CompleteSchoolName [s] @@ |
@PATH _ROOT _translations _headerZone _LINE
@RULES
_translation <-
_xWILD [plus match=(_xALPHA)] ### (1)
@@ |
@NODES _ROOT
@POST
S("pos") = N("$text",2);
S("con") = getconcept(G("funcs"),N("$text",2));
single();
@RULES
_pos <-
const ### (1)
_xALPHA ### (2)
\= ### (3)
_brackets ### (4)
@@
|
# Fetch string-value of attribute (must be first).
L("return_str") = strval(L("con"), L("name")); |
###############################################
# FILE: custom_funcs
# SUBJ: comment
# AUTH: Ashton
# MODIFIED:
###############################################
@DECL
# Concatenate
# Concatenates two non-empty arrays
#
# Args: Two non-empty arrayd
# Returns: concatenated array
Concatenate(L("arr_1"), L("arr_2")) {
L("new_arr") = L("arr_1");
L("idx") = 0;
while (L("idx") < arraylength(L("arr_2"))) {
L("new_arr")[arraylength(L("new_arr"))] = L("arr_2")[L("idx")];
L("idx")++;
}
return L("new_arr");
}
# Swap
# Swap elements at first and second indices in array
#
# Args:
# L("arr") = Non-empty array
# L("first_idx") = Index of first element to swap
# L("second_idx") = Index of second element to swap
#
# Returns: Array with swapped elements
#
# Note that indices must be < array len
Swap(L("arr"), L("first_idx"), L("second_idx")) {
L("temp") = L("arr")[L("first_idx")];
L("arr")[L("first_idx")] = L("arr")[L("second_idx")];
L("arr")[L("second_idx")] = L("temp");
return L("arr");
}
# Quicksort wrapper to handle indices, since this can be a challenge.
# Use QuickSortPartition for subarray
QuickSort(L("arr")) {
if (arraylength(L("arr")) <= 1) {
return L("arr");
}
L("start") = 0;
L("end") = arraylength(L("arr")) - 1;
return QuickSortPartition(L("arr"), L("start"), L("end"));
}
# QuickSortPartition
# Performs quicksort on array from <low> to <high>
#
# Args:
# L("arr"): Array to sort
# L("low"): Starting index of array
# L("high"): Upper index of array
#
# Returns:
# Sorted array
#
# N.B. low and high must be >=0 and < array len
QuickSortPartition(L("arr"), L("low"), L("high")) {
if (L("low") < L("high")) {
# Get pivot index
L("pivot") = L("arr")[L("high")];
L("i") = L("low") - 1;
L("j") = L("low");
while (L("j") < L("high")) {
if (L("arr")[L("j")] <= L("pivot")) {
L("i")++;
L("arr") = Swap(L("arr"), L("i"), L("j"));
}
L("j")++;
}
L("arr") = Swap(L("arr"), L("i")+1, L("high"));
L("pivot") = L("i") + 1;
# Sort each partition of array recursively
L("arr") = QuickSortPartition(L("arr"), L("low"), L("pivot")-1);
L("arr") = QuickSortPartition(L("arr"), L("pivot")+1, L("high"));
# Return concatenated array
# L("sorted") = L("left_arr");
# L("sorted")[arraylength(L("sorted"))] = L("pivot");
# return ConcatArrays(L("sorted"), L("right_arr"));
}
return L("arr");
}
# Naive tokenization function
GetTokens(L("str")) {
L("new_str");
L("str") = strsubst(L("str"), "n't", " not");
L("str") = strclean(L("str"));
L("i") = 0;
while (L("i") < strlength(L("str"))) {
L("char") = strchar(L("str"), L("i"));
L("char_to_add");
if (strisalpha(L("char")) || L("char") == " ") {
L("char_to_add") = L("char");
}
else {
L("char_to_add") = " ";
}
if (!L("new_str")) {
L("new_str") = L("char_to_add");
}
else{
L("new_str") = L("new_str") + L("char_to_add");
}
L("i")++;
}
L("tokens") = split(L("new_str"), " ");
L("return");
L("j") = 0;
"debug.txt" << arraylength(L("tokens")) << "\n";
"debug.txt" << L("new_str") << "\n";
"debug.txt" << L("str") << "\n";
while (L("j") < arraylength(L("tokens"))) {
L("this_str") = L("tokens")[L("j")];
if (L("this_str")) {
L("this_str") = strclean(L("this_str"));
L("this_str") = strtolower(L("this_str"));
if ((strlength(L("this_str")) > 2) && (spellword(L("this_str")))) {
L("return")[arraylength(L("return"))] = stem(L("this_str"));
}
}
L("j")++;
}
return L("return");
}
# GetUniqueWords
# Args:
# L("str"): Input string to get words from
#
#
# Should take parsetree node as well
# GetUniqueStrWords(L("str")) {
# # Assumes non-tokenized string
# L("str")
# if (L("vocab")) {
# for L("")
# }
# # }
# Binary Search
# Args
# arr: sorted array in which to search
BinarySearchPartition(L("arr"), L("low"), L("high"), L("val")) {
"binary.txt" << "low: " << L("low") << " high: " << L("high") << "\n";
if (L("low") > L("high")) {
return -1;
}
L("mid") = (L("high") + L("low")) / 2;
"binary.txt" << "mid: " << L("mid") << "\n";
if (L("arr")[L("mid")] == L("val")) {
return L("mid");
}
else if (L("val") < L("arr")[L("mid")]) {
L("high") = L("mid") - 1;
return BinarySearchPartition(L("arr"), L("low"), L("high"), L("val"));
}
else {
L("low") = L("mid") + 1;
return BinarySearchPartition(L("arr"), L("low"), L("high"), L("val"));
}
}
# Binary Search
# Args
# arr: sorted array to search
# val: value to search for
#
# Returns
# integer: -1 if not in array, otherwise idx in array
BinarySearch(L("arr"), L("val")) {
L("len") = arraylength(L("arr"));
L("low") = 0;
if (L("len") == 1) {
if (L("arr")[L("low")] == L("val")) {
return 0;
}
else {
return -1;
}
}
return BinarySearchPartition(L("arr"), L("low"), L("len")-1, L("val"));
}
# FilterDuplicates
# Returns list as a set of unique elements.
# Quicksort, then iterate. C
# Averge time complexity O(Nlog(N))
# Build a hash function later for O(N) performance
# Args
# arr: array to sort
# Returns
# sorted array of unique values in list
FilterDuplicates(L("arr"), L("sorted")) {
L("len") = arraylength(L("arr"));
if (L("len") <= 1) {
return L("arr");
}
L("i") = 0;
# Sort list
if (!L("sorted")) {
L("sorted_arr") = QuickSort(L("arr"));
}
else {
L("sorted_arr") = L("arr");
}
L("set_idx") = 0;
L("set");
while (L("i") < L("len")) {
L("curr_item") = L("sorted_arr")[L("i")];
L("i")++;
# Move i to next unique item
while (L("sorted_arr")[L("i")] == L("curr_item")) {
L("i")++;
}
if (!L("set_idx")){
L("set") = L("curr_item");
}
else {
L("set")[L("set_idx")] = L("curr_item");
}
L("set_idx")++;
}
return L("set");
}
# Take KB string values and convert to array
# Args
# first_val: first value in kb value list
StrValsToArray(L("first_val")) {
L("arr");
L("arr_idx") = 0;
while (L("first_val")) {
L("arr")[L("arr_idx")] = getstrval(L("first_val"));
L("arr_idx")++;
L("first_val") = nextval(L("first_val"));
}
return L("arr");
}
@@DECL |
@CODE
# Traverse rare split of eui_to_codes concept, saving sum total
# and adding eui/count to kb
G("rare_eui_sum") = 0;
G("rare_eui_exp_sum") = 0;
L("iter") = down(G("eui_to_codes_rare"));
while (L("iter")) {
# Get all codes for eui
L("codes") = findvals(L("iter"), "codes");
# Create concept for code in P_eui
L("eui_con") = makeconcept(G("P_eui_rare"), conceptname(L("iter")));
# Get number of codes in codes attribute
L("num_codes") = ValsLength(L("codes"));
# Add to sum total
G("rare_eui_sum") = G("rare_eui_sum") + L("num_codes");
# Add to exp total (softmax denominator)
G("rare_eui_exp_sum") = flt(G("rare_eui_exp_sum")) + exp(L("num_codes"));
# Add to numval
addnumval(L("eui_con"), "codes", L("num_codes"));
# Advance iterator to sibling concept
L("iter") = next(L("iter"));
}
"p_rare.txt" << "\nTotal sum: " << G("rare_eui_sum") << "\n";
"p_rare.txt" << "Total exp sum: " << G("rare_eui_exp_sum") << "\n\n";
# Calulate softmax P
L("iter") = down(G("P_eui_rare"));
while (L("iter")) {
# Get codes attribute
L("frequency") = numval(L("iter"), "codes");
# Calculate exp(eui) / (sum to j [exp(eui_j)])
L("exp_n") = flt(exp(L("frequency")))
L("p") = L("exp_n") / flt(G("rare_eui_exp_sum"));
"p_rare.txt" << conceptname(L("iter")) << ": " << L("exp_n") << " / total = " << L("p") << "\n";
# Add p as attribute under eui
addstrval(L("iter"), "p", str(L("p")));
# Advance iterator to next sibling
L("iter") = next(L("iter"));
}
@@CODE |
# @MULTI _section _sentence _subsection
# @POST
# L("pn_iter") = N(1);
# L("wrote") = 0;
# while (L("pn_iter")) {
# L("bases") = pnvar(L("pn_iter"), "bases")
# if (L("bases")) {
# if (strcontains("|",L("bases"))) {
# L("all_bases") = split(L("bases"), "|");
# AddUniqueCon(G("note_euis"), pnname(L("pn_iter")));
# "words.txt" << pnname(L("pn_iter")) << ":\n\t" << L("bases") << "\n";
# }
# AddUniqueCon(G("note_words"), pnname(L("pn_iter")));
# "words.txt" << pnname(L("pn_iter")) << ":\n\t" << L("bases") << "\n";
# L("wrote") = 1;
# }
# L("pn_iter") = pnnext(L("pn_iter"));
# }
# @RULES
# _xNIL <-
# _xANY
# _xWILD [fails=(_xEND)]
# _xEND
# @@
|
# Split a string into pieces using a separator character
@CODE
L("arr") = split("ab|cd|efg",
"|");
"output.txt"
<< L("arr") << "\n";
"output.txt"
<< L("arr")[2] << "\n";
prints out:
ab cd efg
efg |
@CODE
L("hello") = 0;
@@CODE
#@PATH _ROOT _TEXTZONE _sent
@NODES _sent
# ^ alpha ,
@PRE
<2,2> var("adv");
<2,2> varz("inc100 ^-alpha-comma");
@POST
alphatoadv(2);
N("inc100 ^-alpha-comma",2) = 1;
@RULES
_xNIL <-
_xSTART
_xALPHA
\, [lookahead]
@@
# TOO BROAD. Could be imperative etc. #
# ^ alpha
#@PRE
#<2,2> varz("inc
#@POST
# if (plural(N(2)))
# {
# if (N("verb",2))
# {
# # Clear out verb.
# N("verb",2) = 0;
# --N("pos num",2);
# alphaunambigred(2);
# }
# }
#@RULES
#_xNIL <-
# _xSTART
# _xALPHA
# @@
# ^ alpha that
@CHECK
if (!N("verb",2))
fail();
@POST
L("tmp2") = N(2);
group(2,2,"_verb");
pncopyvars(L("tmp2"),N(2));
# Todo: reduce to vg.
@RULES
_xNIL <-
_xSTART
_xALPHA
that [s lookahead]
@@
# ^ np that alpha
# ^ noun that alpha
@PRE
<4,4> var("verb");
@CHECK
if (!vconjq(N(4),"inf") && !vconjq(N(4),"-s"))
fail();
@POST
alphatoverb(4,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_np _noun)]
that [s]
_xALPHA
@@
# ^ verb that verb
@CHECK
if (N("mypos",3))
fail();
if (!vconjq(N(2),"-ing"))
fail();
@POST
chpos(N(3),"DT"); # that/DT.
nountonp(3,1);
N("number",3) = "singular";
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_verb _vg)]
that [s except=(_np _det)]
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# ^ verb that alpha prep
@CHECK
if (!N("noun",4))
fail();
@POST
L("tmp4") = N(4);
group(4,4,"_noun");
pncopyvars(L("tmp4"),N(4));
if (!N("mypos",3))
chpos(N(3),"DT"); # that/DT
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_verb _vg)]
that [s]
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# ^ that noun
@CHECK
if (N("mypos",2))
fail();
@POST
if (singular(N(4)))
{
chpos(N(2),"DT"); # that/DT.
pnrename(N(2),"_det");
}
else
chpos(N(2),"IN"); # that/IN.
@RULES
_xNIL <-
_xSTART
that [s except=(_det)]
_xWILD [star match=(_adv _advl)]
_xWILD [one match=(_noun)]
@@
# ^ that alpha noun
@CHECK
if (!N("verb",3))
fail();
S("ed") = vconjq(N(3),"-ed");
if (!S("ed")
&& !vconjq(N(3),"-s"))
fail();
@POST
if (!N("mypos",2))
chpos(N(2),"DT"); # that/DT
pnrename(N(2),"_det");
L("tmp3") = N(3);
group(3,3,"_verb");
pncopyvars(L("tmp3"),N(3));
if (S("ed"))
chpos(N(3),"VBD");
else
chpos(N(3),"VBZ");
@RULES
_xNIL <-
_xSTART
that [s]
_xALPHA
_noun [lookahead]
@@
# ^ that alpha
@PRE
<2,2> varz("mypos"); # 05/04/07 AM.
@CHECK
if (!N("noun",3) && !N("adj",3))
fail();
if (N("verb",3))
{
if (!vconjq(N(3),"inf"))
fail();
}
@POST
if (!N("mypos",2))
chpos(N(2),"DT"); # that/DT
pnrename(N(2),"_det");
@RULES
_xNIL <-
_xSTART
that [s]
_xALPHA
@@
# ^ fnword det alpha alpha
# ^ fnword dqa alpha alpha
# NIBBLE.
@POST
# Need agreement, plurality checks.
if (N("noun",6) && N("adj",6))
{
if (plural(N(6)))
alphatonoun(6);
else if (N("noun",7) || N("adj",7))
alphatoadj(6);
else
fixnphead(6);
}
else if (N("noun",6))
alphatonoun(6);
else if (N("adj",6))
alphatoadj(6);
else if (N("adv",6))
alphatoadv(6);
else
alphatoadj(6); # Probly some verby form.
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_fnword)]
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_xALPHA
@@
# ^ noun modal alpha
@CHECK
if (!N("verb",6))
fail();
if (!vconjq(N(6),"inf"))
fail();
@POST
L("tmp6") = N(6);
group(6,6,"_verb");
pncopyvars(L("tmp6"),N(6));
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_pro _noun _np)]
_xWILD [star match=(_adv _advl)]
_modal
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_noun _np _pro)]
@@
# ^ noun alpha that
# ^ pro alpha that
# that
@PRE
<4,4> var("verb");
@POST
alphatovg(4,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_pro _noun _np)]
_xWILD [star match=(_adv _advl)]
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [s one match=(that)]
@@
# ^ dqan alpha that
@PRE
<7,7> var("verb");
@POST
alphatovg(7,"active","VBP");
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_xWILD [star match=(_adv _advl)] ### (6)
_xALPHA ### (7)
_xWILD [star lookahead match=(_adv _advl)] ### (8)
that [s one] ### (9)
@@
# ^ pro alpha
@CHECK
if (!N("verb",3))
fail();
@POST
alphatoverb(3,"active","VBP");
@RULES
_xNIL <-
_xSTART
_proSubj [s]
_xALPHA
_xWILD [one lookahead fail=(_modal _verb _vg)]
@@
# ^ noun alpha
@PRE
<4,4> var("verb");
@CHECK
if (!plural(N(2)))
fail();
if (!vconjq(N(4),"inf"))
fail();
@POST
alphatovg(4,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_noun _np)]
_adv [star]
_xALPHA
@@
# ^ np alpha alpha
@PRE
<4,4> var("verb");
@CHECK
if (pnname(N(2)) == "_pro" && !N("prosubj",2))
fail();
if (vconjq(N(4),"-ing") || vconjq(N(4),"-en"))
fail();
if (singular(N(2)) && vconjq(N(4),"inf"))
fail();
if (plural(N(2)) && vconjq(N(4),"-s"))
fail();
@POST
alphatovg(4,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_noun _np _pro)]
_adv [plus]
_xALPHA
@@
# ^ dqan alpha dqan
@PRE
<6,6> var("verb");
@POST
# todo: Check agreement.
alphatovg(6,"active","VBP");
@RULES
_xNIL <-
_xSTART
_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
\, )]
@@
# ^ 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
\,
@@
# ^ dqa alpha noun
@PRE
<5,5> var("noun");
<6,6> vareq("sem","person");
@CHECK
if (!N(2) && !N(3) && !N(4))
fail();
@POST
fixnphead(5);
dqaninfo(2,3,4,5);
groupnp();
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_xALPHA
_noun
@@
# ^ dqan alpha as verb
# Ex: People regarded as being weird....
@PRE
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"-en"))
fail();
@POST
alphatovg(6,"passive","VBN");
N("ellipted-that-is",6) = 1;
# L("tmp6") = N(6);
# group(6,6,"_clause");
# setunsealed(6,"true"); #
# N("voice",6) = "passive";
# N("vg node",6) = L("tmp6");
chpos(N(7),"IN"); # as/IN
pnrename(N(7),"_fnword"); # prep -> fnword
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
_xALPHA
as [s]
_xWILD [one match=(_verb _vg)]
@@
# ^ dqan alpha prep
@CHECK
if (!N("verb",6))
fail();
if (phrprepverbq(N(6),N(7)))
succeed();
if (N("adj",6) || N("noun",6))
fail();
# Agreement...
if (vconjq(N(6),"-ing"))
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
_xWILD [one lookahead match=(_prep)]
@@
# ^ dqan alpha alpha dqan [idiom: let stand]
# eg: court let stand a ruling
# let
# stand
@POST
alphatoverb(7,"active","VB"); # Cleanup. # 05/28/07 AM.
alphatoverb(6,"active","VBD"); # Cleanup. # 05/28/07 AM.
group(6,7,"_vg");
N("voice",6) = "active";
@RULES
_xNIL <-
_xSTART
_xWILD [star match=(_det _pro)]
_xWILD [star match=(_quan _num _xNUM)]
_adj [star]
_noun [plus]
let
stand
_xWILD [one lookahead match=(_det _pro _quan _num _xNUM
_adj _noun _np _adv _advl)]
@@
# ^ dqan alpha adv alpha
@PRE
<4,4> var("noun");
<6,6> var("verb");
@CHECK
if (!plural(N(4)))
fail();
if (!vconjq(N(6),"inf"))
fail();
@POST
alphatoverb(6,"active","VBP");
fixnphead(4);
@RULES
_xNIL <-
_xSTART
_adj
_noun
_xALPHA
_adv
_xALPHA
_xWILD [one lookahead match=(_fnword)]
@@
# ^ dqan alpha
@CHECK
if (N("verb",4))
fail();
if (!N("noun",4) && !N("adj",4))
fail();
if (singular(N(3)) && singular(N(4)))
succeed();
fail();
@POST
alphatonoun(4);
@RULES
_xNIL <-
_xSTART
_det [opt]
_noun
_xALPHA
@@
# ^ dqan alpha alpha noun
@PRE
<6,6> var("verb");
@CHECK
if (!vconjq(N(6),"-ed"))
fail();
if (!N("noun",6) && !N("adj",6))
fail();
@POST
alphatovg(6,"active","VBD");
@RULES
_xNIL <-
_xSTART ### (1)
_xWILD [star match=(_det _pro)] ### (2)
_xWILD [star match=(_quan _num _xNUM)] ### (3)
_adj [star] ### (4)
_noun [plus] ### (5)
_xALPHA ### (6)
_xALPHA ### (7)
@@
# Special cases. Can't resolve yet.
@POST
noop();
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(following)]
@@
# ^ alpha det
@PRE
<2,2> varz("inc100 a-det");
@POST
L("done") = 0;
if (!L("done") && N("verb",2))
{
# Presumably gerund tagging will dominate in Treebank.
if (vconjq(N(2),"-ing")) # Gerund case.
{
alphatovg(2,"active","VBG"); # Conform Treebank.
L("done") = 1;
}
else if (vconjq(N(2),"inf"))
{
alphatovg(2,"active","VB");
L("done") = 1;
}
}
if (!L("done") && N("adv",2))
{
alphatoadv(2);
L("done") = 1;
}
if (!L("done") && N("adj",2))
{
alphatoadj(2);
L("done") = 1;
}
N("inc100 a-det",2) = 1;
@RULES
_xNIL <-
_xSTART
_xALPHA
_xWILD [one lookahead match=(_det)]
@@
# ^ alpha verb
@PRE
<2,2> varz("inc100 av");
@CHECK
if (!N("noun",2) && !N("unknown",2))
fail();
@POST
if (singular(N(2)) && !N("mass",2) && !N("mypos",2))
cappos(N(2),0);
fixnphead(2);
N("inc100 av",2) = 1; # Loop guard.
@RULES
_xNIL <-
_xSTART
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_verb _vg _modal)]
@@
# ^ alpha , pro
@CHECK
if (!N("adv",2))
fail();
@POST
alphatoadv(2);
@RULES
_xNIL <-
_xSTART
_xALPHA
\, [lookahead]
_pro
@@
# ^ alpha alpha verb
@PRE
<2,3> var("noun"); # Require possible nouns.
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_modal _verb _vg)]
@@
# ^ alpha alpha ,
@PRE
<3,3> var("noun"); # Require possible noun.
@CHECK
if (!N("adj",2) && !N("noun",2))
fail();
if (N("verb",3))
{
if (!vconjq(N(3),"inf")
&& !vconjq(N(3),"-s"))
fail();
}
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xWILD [one lookahead match=( \, )]
@@
# ^ alpha alpha pro
@PRE
<2,3> var("noun"); # Require possible nouns.
@CHECK
if (!singular(N(2)) || N("unknown",2))
fail();
# Todo: not proper name.
@POST
fixnphead(3);
fixnpnonhead(2);
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xWILD [s one lookahead match=(
himself herself itself themselves ourselves yourselves
_noun _verb _vg _adv
)]
@@
# ^ alpha alpha prep
# Note: ellipsis. "people happy with ..."
@CHECK
if (N("inc100 a-a-prep",2))
fail();
@POST
N("inc100 a-a-prep",2) = 1;
if (N("noun",2))
{
if (plural(N(2)))
{
if (N("verb",3))
{
alphatoverb(3,"active","VBP");
fixnphead(2);
}
else if (N("adj",3))
{
alphatoadj(3);
fixnphead(2);
}
}
}
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_prep)]
@@
# ^ alpha alpha adj
@CHECK
if (!N("verb",3) || !N("noun",2))
fail();
if (!vconjq(N(3),"-ed"))
fail();
@POST
alphatoverb(3,"active","VBP");
if (singular(N(2)) && !N("mass",2))
cappos(N(2),0);
fixnphead(2);
@RULES
_xNIL <-
_xSTART
_xALPHA
_xALPHA
_xWILD [star lookahead match=(_adv _advl)]
_xWILD [one match=(_det _quan _num _xNUM _adj _np)]
@@
# question patterns.
# why are alpha
@POST
if (N("noun",4))
{
if (plural(N(4)))
fixnphead(4);
else
fixnpnonhead(4);
}
else if (N("adj",4))
fixnpnonhead(4);
else if (N("adv",4))
{
L("tmp4") = N(4);
group(4,4,"_adv");
pncopyvars(L("tmp4"),N(4));
fixadv(N(4));
}
else # verb or otherwise.
fixnpnonhead(4);
@RULES
_xNIL <-
_xSTART
_xWILD [s one match=(_whword)]
are [s]
_xALPHA
@@
# ^ yet
@CHECK
if (N("mypos",2))
fail();
@POST
chpos(N(2),"CC"); # yet/CC
pnrename(N(2),"_adv");
@RULES
_xNIL <-
_xSTART
yet [s]
@@
# ^ adj alpha alpha [NIBBLE]
@CHECK
if (!N("adj",3) && !N("noun",3))
fail();
@POST
if (singular(N(3)))
fixnpnonhead(3);
else
fixnphead(3);
@RULES
_xNIL <-
_xSTART
_adj
_xALPHA
_xALPHA
@@
# ^ noun 's alpha
@CHECK
if (!N("adj",4))
fail();
@POST
alphatoadj(4);
pnrename(N(3),"_verb");
chpos(N(3),"VBZ"); # 's/VBZ
N("copula",3) = 1;
@RULES
_xNIL <-
_xSTART
_noun
_aposS
_xALPHA
_xWILD [one lookahead match=(_dbldash \, \; _qEOS)]
@@
# ^ noun verb noun , alpha noun
@PRE
<6,6> var("verb");
@CHECK
L("v") = N("verb node",3);
S("c") = samevconj(L("v"),N(6));
if (!S("c"))
fail();
@POST
alphatovg(6,"active","VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_noun _np)]
_xWILD [one match=(_verb _vg)]
_xWILD [one match=(_noun _np)]
\,
_xALPHA
_xWILD [one lookahead match=(_noun _np _det _quan
_num _xNUM _adj _xALPHA)]
@@
# to alpha alpha dqan
@PRE
<2,2> var("adv");
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"inf"))
fail();
@POST
alphatovg(3,"active","VB");
alphatoadv(2);
@RULES
_xNIL <-
to [s]
_xALPHA
_xALPHA
_xWILD [one lookahead match=(_det _pro _quan _num _xNUM _np)]
@@
# ^ np , dqan alpha ,
@PRE
<6,6> var("verb");
@CHECK
if (plural(N(5)) && vconjq(N(6),"-s"))
fail();
if (singular(N(5)) && vconjq(N(6),"inf"))
fail();
if (!verbfeat(N(6),"T5"))
fail();
@POST
alphatovg(6,0,"VBP");
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(_np _noun)]
\,
_det
_noun
_xALPHA
\, [lookahead]
@@
# ^ alpha num
@PRE
<2,2> var("verb");
@CHECK
if (!vconjq(N(2),"inf"))
fail();
@POST
alphatovg(2,"active","VB");
@RULES
_xNIL <-
_xSTART
_xALPHA
_xWILD [one lookahead match=(_quan _num _xNUM)]
@@
# ^ and then
@PRE
<3,3> vareq("stem","then");
@POST
group(2,3,"_adv");
@RULES
_xNIL <-
_xSTART
_conj
_fnword [opt]
@@
# ^ before alpha
# ^ before ving
@PRE
<3,3> var("verb");
@CHECK
if (!vconjq(N(3),"-ing"))
fail();
@POST
alphatovg(3,"active","VBG");
@RULES
_xNIL <-
_xSTART
before [s]
_xALPHA
@@
# Sentinel.
@POST
# When done matching rules against starts,
# finish with current segment.
# NLP++ opt: would be nice to have a way to exit the
# current dominating context (ie, the current _sent).
# something like an xcontinue() or xnext() function.
noop();
@RULES
_xNIL <-
_xANY [plus]
@@
|
@NODES _LINE
@POST
S("header") = N("$text",2);
X("header") = 1;
S("count") = strlength(N("$text",1));
"headers.txt" << SpacesStr(S("count")) << S("count") << " " << S("header") << "\n";
if (S("count") > G("max header"))
G("max header") = S("count");
single();
@RULES
_header <-
_xWILD [min=2 match=(\=)] ### (1)
_xWILD [plus fail=(\=)] ### (2)
_xWILD [min=2 match=(\=)] ### (3)
@@
|
@NODES _ROOT
@POST
AddUniqueStr(G("alphaNumeric"), "terms", N("$text", 1) + N("$text", 2));
single();
@RULES
_alphaNum <-
_xALPHA ### (1)
_xNUM ### (2)
@@ |
@PATH _ROOT
@POST
S("ref") = N("ref",1);
"bib.txt" << "Ref: " << S("ref") << "\n";
single();
@RULES
_bibItem <-
_bibRef ### (1)
_xWILD [fail=(_bibRef _endBib)] ### (2)
@@
|
# Create two children concepts then delete them both. Then create them again and deletes one by name:
if(findconcept(findroot(),"a concept"))
rmconcept(findconcept(findroot(),"a concept"));
G("aParent")= makeconcept(findroot(),"a concept")
G("aChild") = makeconcept(G("aParent"),"a child");
G("aChild") = makeconcept(G("aParent"),"a sibling");
rmchildren(G("aParent"));
G("aChild") = makeconcept(G("aParent"),"a child");
G("aChild") = makeconcept(G("aParent"),"a sibling");
rmchild(G("aParent"),"a sibling"); |
@CODE
DictionaryClear();
DictionaryWord("intact","nominal", "true", "str");
DictionaryEnd();
@@CODE
|
@NODES _ROOT
@RULES
_paragraph [unsealed] <-
_sentence [s plus] ### (1)
_xWILD [s matches=(_paragraphSeparator _xEND)] ### (2)
@@ |
@NODES _ROOT
@PRE
<2,2> uppercase();
@POST
S("header") = N("$text", 2);
singler(2,3);
@RULES
_zoneHeader <-
_xWILD [one matches=(\n \r _header)] ### (1)
_xALPHA [plus] ### (2)
_xWILD [one matches=(\n \r)] ### (3)
@@ |
# Find entire path of dictionary concept for the given string. (If not present, don't add the word.)
@CODE
"output.txt" << "1 " << conceptname(addword("hello")) << "\n";
"output.txt" << "2 " << conceptname(wordindex("hello")) << "\n";
"output.txt" << "3 " << findwordpath("hello") << "\n";
"output.txt" << "4 " << findwordpath("olleh") << "\n";
"output.txt" << "5 " << wordpath("foobaz") << "\n";
"output.txt" << "6 " << conceptname(dictfindword("hello")) << \n";
rmword("foobaz");
Prints out:
1 hello
2 he
3 "concept" "sys" "dict" "a" "h" "he" "hello"
4
5 "concept" "sys" "dict" "a" "f" "fo" "foobaz"
6 hello |
@NODES _paragraph
@POST
singler(2,2);
@RULES
_endSent <-
_xWILD [s one match=(_xALPHA _xNUM)] ### (1)
_xWILD [one trig match=(\. \? \!)] ### (2)
_xWILD [one match=(_xWHITE _xALPHA _xNUM _xEND _xCTRL)] ### (3)
@@
|
@NODES _LINE
@POST
S("con") = N("$text",2);
X("con") = N("$text",2);
single();
@RULES
_con <-
\" ### (1)
_xWILD [fail=(\")] ### (2)
\" ### (3)
@@
|
@CODE
DisplayKB(G("resumes"),1);
@@CODE |
@NODES _LINE
@POST
L("node") = N(1);
L("last") = 0;
L("final") = 0;
"number.txt" << phrasetext() << "\n";
while (L("node")) {
L("name") = pnname(L("node"));
if (L("name") != "," && L("name") != "-" && L("name") != "and" && L("name") != "&") {
L("num") = num(pnvar(L("node"),"numeric"));
L("power") = num(pnvar(L("node"),"power"));
"number.txt" << pnname(L("node")) << " " << L("num") << " " << L("last") << " => ";
if (L("last power") <= 1 && L("power") == 1) {
L("last") = L("last") * 100 + L("num");
} else if (L("power") <= 2) {
if (L("num") > L("last")) {
if (L("last"))
L("last") = L("last") * L("num");
else
L("last") = L("num");
} else {
L("last") = L("last") + L("num");
}
"number.txt" << L("last");
} else {
L("final") = L("final") + L("last") * L("num");
"number.txt" << "final = " << L("final") << " ";
L("last") = 0;
}
"number.txt" << "\n";
L("last power") = L("power");
}
L("node") = pnnext(L("node"));
}
L("final") = L("final") + L("last");
"number.txt" << "=====> " << L("final") << "\n\n";
S("numeric") = L("final");
single();
@RULES
_number <-
_xWILD [plus match=(_number \- \, and \&)] ### (1)
@@
|
@NODES _term
@POST
G("root") = makeconcept(G("radlex"), N("$text", 1));
exitpass();
@RULES
_xNIL <-
_xWILD [fail=(\t \n \r)] ### (1) ### (2)
_xWILD [one matches=(\n \r)] ### (3)
@@
|
# Rename an attribute named oldAttrNameStr to newAttrNameStr belonging to the concept aConcept
if(findconcept(findroot(),"apples"))
rmconcept(findconcept(findroot(),"apples"));
G("apples") = makeconcept(findroot(),"apples");
addstrval(G("apples"),"have","seeds");
renameconcept(G("apples"),"fruit");
"output.txt" << conceptname(G("apples")) << "\n";
renameattr(G("apples"),"have","are");
"output.txt" << attrname(findattr(G("apples"),"are")) << "\n";
"output.txt" << strval(G("apples"),"are") << "\n";
# rmconcept(G("apples"));
This code prints out:
fruit
are
seeds |
@NODES _ROOT
@PRE
<3,3> var("lang");
<5,5> var("pos");
@RULES
_lang <-
\{ ### (1)
\{ ### (2)
_xALPHA ### (3)
\- ### (4)
_xALPHA ### (5)
_xWILD [fail=(\})] ### (6)
\} ### (7)
\} ### (8)
@@
@RULES
_curly <-
\{ ### (1)
\{ ### (2)
_xWILD [fail=(\})] ### (3)
\} ### (4)
\} ### (5)
@@
@POST excise(1,2); @RULES _xNIL <- \[ \[ @@
@POST excise(1,2); @RULES _xNIL <- \] \] @@
@POST excise(1,2); @RULES _xNIL <- \( \( @@
@POST excise(1,2); @RULES _xNIL <- \) \) @@
@POST excise(1,2); @RULES _xNIL <- \< \< @@
@POST excise(1,2); @RULES _xNIL <- \> \> @@
@POST excise(1,2); @RULES _xNIL <- \> \> @@
@POST excise(1,3); @RULES _xNIL <- \' \' \' [opt] @@
|
# Check if first char of string1 is uppercase
@CHECK
if (strisupper(N("$text",1)))
fail();
@RULES
_modal <- will @@ |
@NODES _LINE
@POST
X("code",2) = strtolower(N("$text",2));
single();
@RULES
_code <-
\[ ### (1)
_xALPHA ### (2)
\] ### (3)
@@
|
@MULTI _ROOT
@PRE
<1,1> var("key_words");
@POST
"pass_25.txt" << N("key_words", 1) << "\n";
extractCodeRanks(N("key_words", 1), "diagnoses");
extractCodeRanks(N("key_words", 1), "procedures");
@RULES
_xNIL <-
_xWILD [one matches=( _section _subsection _sentence _looseText _item)]
@@ |
@CODE
# Get our eui to codes mapping
G("eui_to_codes") = getconcept(findroot(),"eui_to_codes");
G("eui_to_codes_top") = getconcept(G("eui_to_codes"),"top");
G("eui_to_codes_rare") = getconcept(G("eui_to_codes"),"rare");
DisplayKB(findroot(), 2);
# Check whether P_eui concept exists, if not, create it.
G("P_eui") = getconcept(findroot(),"P_eui");
if (!G("P_eui")) {
G("P_eui") = makeconcept(findroot(), "P_eui");
}
rmchildren(G("P_eui"));
# Make concept for top split of P_eui
G("P_eui_top") = getconcept(G("P_eui"),"top");
if (!G("P_eui_top")) {
G("P_eui_top") = makeconcept(G("P_eui"), "top");
}
rmchildren(G("P_eui_top"));
# Make concept for rare split of P_eui
G("P_eui_rare") = getconcept(G("P_eui"),"rare");
if (!G("P_eui_rare")) {
G("P_eui_rare") = makeconcept(G("P_eui"), "rare");
}
rmchildren(G("P_eui_rare"));
@@CODE |
@NODES _ROOT
@RULES
_patientID <-
\[ ### (1)
\* ### (2)
\* ### (3)
_xWILD [fails=(\*)] ### (4)
\* ### (5)
\* ### (6)
\] ### (7)
@@
@RULES
_time <-
_xNUM
\:
_xNUM
\: [opt]
_xNUM [opt]
\_ [opt]
_xWILD [opt match=(am pm AM PM)]
@@
@POST
excise(6,6);
excise(4,4);
excise(2,2);
single();
# Initialism
@RULES
_init <-
_xWILD [one matches=(_xNUM _xALPHA)] ### (1)
\. ### (2)
_xWILD [one matches=(_xNUM _xALPHA)] ### (3)
\. ### (4)
_xWILD [opt matches=(_xNUM _xALPHA)] ### (5)
\. [opt] ### (6)
@@
@POST
excise(2,2);
singler(1,1);
@RULES
_init <-
# Add abbreviations here, in the form dr., jr., etc
_xWILD [one match=(Dr DR dr q etc)]
\.
_xWILD [one matches=(\_ \, \:)]
@@ |
@CODE
L("hello") = 0;
@@CODE
#@PATH _ROOT _TEXTZONE _tok
@NODES _tok
# Treebank note: JJ is the default, since alpha-alpha
# nonheads of nps are about 77% adjective.
@POST
X("hyphenated") = 1;
X("pos_np") = "JJ"; # Default. # 06/07/06 AM.
if (strisupper(N("$text",2)) || strisupper(N("$text",4)))
X("cap") = 1;
xrename("_adj");
chpos(X(),"JJ");
@RULES
_xNIL <-
_xSTART
short
\-
term
_xEND
@@
_xNIL <-
_xSTART
worst
\-
case
_xEND
@@
_xNIL <-
_xSTART
pro
\-
forma
_xEND
@@
_xNIL <-
_xSTART
pre
\-
_xWILD [one match=(_xALPHA _xNUM)]
_xEND
@@
_xNIL <-
_xSTART
_xALPHA # run, drag
\-
down
_xEND
@@
_xNIL <-
_xSTART
_xALPHA
\-
free
_xEND
@@
# Treebank note: JJ is the default, since alpha-alpha
# nonheads of nps are about 77% adjective.
@POST
X("hyphenated") = 1;
X("pos_np") = "NN";
if (strisupper(N("$text",2)) || strisupper(N("$text",4)))
X("cap") = 1;
xrename("_adj");
chpos(X(),"NN");
@RULES
_xNIL <- # NN/JJ 15/5.
_xSTART
high
\-
technology
_xEND
@@
# buy-out
# buy - out
_xNIL <- # NN/JJ 499/25
_xSTART
buy
\-
out
_xEND
@@
_xNIL <- # NN/JJ 499/25
_xSTART
stock
\-
index
_xEND
@@
_xNIL <-
_xSTART
twin
\-
jet
_xEND
@@
_xNIL <-
_xSTART
close
\-
up
_xEND
@@
_xNIL <-
_xSTART
cop
\-
killer
_xEND
@@
@POST
X("hyphenated") = 1;
X("pos_np") = "NN";
if (strisupper(N("$text",2)) || strisupper(N("$text",4)))
X("cap") = 1;
xrename("_noun");
chpos(X(),"NNS");
@RULES
_xNIL <-
_xSTART
mid
\-
_xNUM
s
_xEND
@@
@POST
X("hyphenated") = 1;
X("pos_np") = "NN";
if (strisupper(N("$text",4)))
X("cap") = 1;
xrename("_noun");
chpos(X(),"NN");
@RULES
_xNIL <-
_xSTART
t
\-
shirt
_xEND
@@
@POST
X("hyphenated") = 1;
X("pos_np") = "NNS";
if (strisupper(N("$text",4)))
X("cap") = 1;
xrename("_noun");
chpos(X(),"NNS");
@RULES
_xNIL <-
_xSTART
t
\-
shirts
_xEND
@@
@PRE
<2,2> length(4);
<4,4> lengthr(1,4);
@POST
X("hyphenated") = 1;
X("pos_np") = "CD";
xrename("_noun");
chpos(X(),"CD");
@RULES
_xNIL <-
_xSTART
_xNUM
\-
_xNUM
_xEND
@@
@POST
X("hyphenated") = 1;
if (strisupper(N("$text",2)) || strisupper(N("$text",4)))
X("cap") = 1;
xrename("_noun");
if (singular(N(2)))
{
X("pos_np") = "NN";
chpos(X(),"NN");
}
else
{
X("pos_np") = "NNS";
chpos(X(),"NNS");
}
@RULES
_xNIL <-
_xSTART
_xWILD [one match=(passer passers)]
\-
by
_xEND
@@
@POST
X("idiom") = 1;
xrename("_verb");
chpos(X(),"VB");
X("verb") = 1;
X("pos num") = 1;
X("inf") = 1;
X("imperative") = 1;
@RULES
_xNIL <-
_xSTART
c
\'
mon
_xEND
@@
# Note: exclusively in "single-handedly".
@POST
N("adv") = 1;
N("sem") = N("stem") = "handed";
N("pos num") = 1;
N("pos") = "_adv";
if (N("unknown"))
N("unknown") = 0;
@RULES
_xNIL <- handedly @@
|
@PATH _ROOT _experienceZone _experienceInstance _LINE
# name regardless of position, etc. Should raise confidence here.
# May want to collect confidence in the instance.
@CHECK
if (
!X("company name",3)
&& (N("tmp") = (X("lineno") - X("anchor lineno",3)))
&& (N("tmp") == -1 || N("tmp") == 1) # Within 2 lines of anchor.
)
succeed();
fail();
@POST
X("company name",3) = N("$text");
@RULES
_xNIL <- _company [s] @@
# Don't get cocky, Luke!
@CHECK
if (
!X("job title",3)
&& (N("tmp") = (X("lineno") - X("anchor lineno",3)))
&& (N("tmp") == -1 || N("tmp") == 1) # Within 2 lines of anchor.
)
succeed();
fail();
@POST
X("job title") = N("$text");
@RULES
_xNIL <- _jobTitle [s] @@
_xNIL <- _jobPhrase [s] @@
@CHECK
if (
!X("country",3)
&& (N("tmp") = (X("lineno") - X("anchor lineno",3)))
&& (N("tmp") == -1 || N("tmp") == 1) # Within 2 lines of anchor.
)
succeed();
fail();
@POST
X("country",3) = N("$text");
@RULES
_xNIL <- _country [s] @@
|
@NODES _LINE
@POST
S("str") = N("str");
S("str") = 1;
excise(3,3);
excise(1,1);
single();
@RULES
_item <-
\" ### (1)
_xWILD [plus fail=(\")] ### (2)
\" ### (3)
@@
|
@CODE
L("hello") = 0;
@@CODE
@NODES _TEXTZONE
@POST
++X("line count",2);
splice(1,1);
@RULES
_xNIL <-
_LINE
@@
|
# Format the current date and time as a string
@CODE
"output.txt" << today() << "\n";
@@CODE |
@NODES _ROOT
@POST
G("file") << N("$text") << "\n";
@RULES
_xNIL <-
_LINE
@@
|
@CODE
G("word dict") = G("$kbpath") + "vocab.dict";
L("debugpath") = G("$kbpath") + "debug.txt";
@@CODE |
@NODES _ROOT
@PRE
<1,1> var("negation");
@POST
if (N("negation")=="CONJ") {
pnrename(N(1), "_conj");
}
else {
pnrename(N(1), "_negation");
}
@RULES
_xNIL <-
_phrase ### (1)
@@
@PRE
<1,1> var("negation");
@POST
S("negation") = pnvar(N(1), "negation");
single();
@RULES
_negation <-
_xALPHA [one] ### (1)
@@ |
@NODES _ROOT
@POST
"verb.txt" << N("$text",4) << "\n";
single();
@RULES
_header <-
\< ### (1)
h2 ### (2)
\> ### (3)
_xALPHA ### (4)
Conjugation ### (5)
\: ### (6)
Present ### (7)
Tense ### (8)
\< ### (9)
\/ ### (10)
h2 ### (11)
\> ### (12)
@@
@RULES
_tbodyStart <-
\< ### (1)
tbody ### (2)
\> ### (3)
@@
@RULES
_tbodyEnd <-
\< ### (1)
\/ ### (2)
tbody ### (3)
\> ### (4)
@@
|
@CODE
fileout("skills.txt");
#varstrs("skills")
@@CODE
@NODES _LINE
@POST
addstrs("skills",1);
prrange("skills.txt",1,1);
prlit("skills.txt","\n");
@RULES
_xNIL <- _skill [s] @@
|
@NODES _ROOT
@POST
S("header") = N("header",1);
S("count") = 5;
single();
@RULES
_headerZone <-
_headerFive ### (1)
_xWILD [plus fails=(_headerFive _headerFour _headerThree _headerTwo _xEND)] ### (2)
@@
|
@NODES _LINE
@POST
singler(3,3);
@RULES
_rouge <-
_space ### (1)
_xWILD [plus fails=(_space)] ### (2)
_space ### (3)
@@
|
@NODES _LINE
@RULES
_email <-
_xWILD [plus match=(_xALPHA _xNUM \_ \.)] ### (1)
\@ ### (2)
_xWILD [min=2 match=(_xALPHA _xNUM \.)] ### (3)
@@
|
@CODE
G("filepath") = G("$kbpath") + "\\" + "en-firstnames.dict";
L("type") = "app";
if (!G("$isdirrun") || G("$isfirstfile"))
G("file") = openfile(G("filepath"));
else
G("file") = openfile(G("filepath"),"app");
@@CODE |
@NODES _LINE
@POST
group(2,2,"_item");
@RULES
_xNIL <-
_xSTART ### (1)
_comma ### (2)
@@ |
@NODES _LINE
@RULES
# Ex: whose
_posPRO [layer=(_funWORD )] <- _xWILD [min=1 max=1 s match=("whose" "me" "myself" "my" "mine" "we" "us" "ourselves" "our"
"ours" "you" "yourself" "yourselves" "your" "yours" "he" "him" "himself" "his"
"she" "her" "herself" "hers" "it" "itself" "its" "they" "them" "themselves"
"their" "theirs" "who" "whom" "which" "that" "I" "everyone" "everybody" "every"
"everything" "each" "all" "many" "much" "more" "most" "few" "fewer" "fewest"
"little" "least" "several" "enough" "both" "one" "ones" "somebody" "someone" "either"
"something" "anybody" "anyone" "anything" "any" "some" "no" "nobody" "neither" "none"
"nothing")] @@
|
@CODE
sortchilds(G("adjMatrixData"));
L("i") = 0;
L("con") = down(G("adjMatrixData"));
while( L("con") ) {
addnumval(L("con"), "index", L("i"));
++L("i");
L("con") = next(L("con"));
}
@@CODE |
# Fetch numeric-value of attribute (must be first).
L("return_num") = numval(L("con"), L("name")); |
@CODE
prlit( "edu.txt", "KEY= 0-none,1-major,2-minor,3-school,\n" );
prlit( "edu.txt", " 4-degree,5-grade,6-date\n\n");
prlit( "edu.txt", "LINE DATA IN EDUCATION ZONE\n" );
prlit( "edu.txt", "---------------------------\n" );
prlit( "edu.txt", " NEW TOT FIRST LAST MAJ MIN SCH DEG GPA DAT\n" );
prlit( "edu.txt", " num key key ord ord ord ord ord ord\n" );
prlit( "edu.txt", "---------------------------------------------\n" );
@@CODE
@PATH _ROOT _educationZone
# Print out relevant vars for each line that has parts.
@CHECK
#Ngt(1, "eduparts", 0) # More than zero education parts.
if (N("eduparts",1) <= 0) fail();
@POST
"edu.txt" << " "
<< rightjustifynum(N("instance",1),3) << " ";
"edu.txt" << rightjustifynum(N("eduparts",1),3) << " ";
"edu.txt" << rightjustifynum(N("first edupart",1),5) << " ";
"edu.txt" << rightjustifynum(N("last edupart",1),4) << " ";
"edu.txt" << rightjustifynum(N("Omajor",1), 3) << " ";
"edu.txt" << rightjustifynum(N("Ominor",1), 3) << " ";
"edu.txt" << rightjustifynum(N("Oschool",1),3) << " ";
"edu.txt" << rightjustifynum(N("Odegree",1),3) << " ";
"edu.txt" << rightjustifynum(N("Ograde",1), 3) << " ";
"edu.txt" << rightjustifynum(N("Odate",1), 3) << "\n";
#noop();
@RULES
_xNIL <- _LINE @@
# Make note of lines that have no eduparts.
@CHECK
# Neq(1, "eduparts", 0)
if (!N("eduparts",1)) fail();
@POST
prlit( "edu.txt", " - - -\n");
@RULES
_xNIL <- _LINE @@
@POST
prlit( "edu.txt", " blank line\n");
@RULES
_xNIL <- _BLANKLINE @@
|
@NODES _ROOT
@RULES
_test <-
_code ### (1)
@@
|
@NODES _LINE
@RULES
# Ex: Extension\_
_phoneExtension <- _xWILD [min=1 max=1 s layer=("_extendWord") match=("Extension" "Ext" "Ex" "x")] _xWHITE [star s] _xNUM [s layer=("_extension")] @@
@PRE
<1,1> cap();
@RULES
# Ex: Ext.\_
_phoneExtension <- _xWILD [min=1 max=1 s layer=("_extendWord") match=("Ext" "Ex")] \. [s] _xWHITE [star s] _xNUM [s layer=("_extension")] @@
|
@CODE
#DisplayKB(G("labels"), 1);
DisplayKB(G("phrases"), 1);
# Replace "labels.kb" with file path to store in different dir
#kbdumptree(G("labels"), "labels.kb");
DumpKB(G("phrases"), "phrases");
DictionaryEnd();
@@CODE
|
@CODE
DisplayKB(G("kb"),0);
@@CODE |
@CODE
G("dictfile") = G("$kbpath") + "states.dict";
G("abbrevdict") = G("$kbpath") + "abbrev.dict";
G("kbbfile") = G("$kbpath") + "states.kbb";
G("citiesdict") = G("$kbpath") + "cities.dict";
G("debugfile") = G("$kbpath") + "debug.txt";
G("debug") = openfile(G("debugfile"),"app");
G("state name") = strsubst(G("$inputhead"),"-"," ");
G("states") = getconcept(findroot(),"states");
if (G("$isfirstfile") || !G("$isdirrun")) {
rmchildren(G("states"));
L("type") = "w";
} else {
L("type") = "app";
}
G("state") = getconcept(G("states"),G("state name"));
G("dict") = openfile(G("dictfile"),L("type"));
G("kbb") = openfile(G("kbbfile"),L("type"));
G("abbrev") = openfile(G("abbrevdict"),L("type"));
G("cities") = openfile(G("citiesdict"),L("type"));
@@CODE |
Subsets and Splits