1
0
forked from GitHub/gf-core

query language generalized and extended ; added README

This commit is contained in:
aarne
2010-06-19 16:24:48 +00:00
parent a0f2ff0772
commit 041e5e2a33
3 changed files with 197 additions and 84 deletions

View File

@@ -10,8 +10,10 @@ cat
Query ; Query ;
Answer ; Answer ;
Set ; -- the set requested, e.g. "all persons" Set ; -- the set requested, e.g. "all persons"
Relation ; -- something of something, e.g. "subregion of Bulgaria" Interrogative ; -- interrog. pron. e.g. "who"
Function ; -- something of something, e.g. "subregion of Bulgaria"
Kind ; -- type of things, e.g. "person" Kind ; -- type of things, e.g. "person"
Relation ; -- relation between things,e.g. "employed at"
Property ; -- property of things, e.g. "employed at Google" Property ; -- property of things, e.g. "employed at Google"
Individual ; -- one entity, e.g. "Google" Individual ; -- one entity, e.g. "Google"
Name ; -- person, company... e.g. "Eric Schmidt" Name ; -- person, company... e.g. "Eric Schmidt"
@@ -22,13 +24,16 @@ fun
MAnswer : Answer -> Move ; MAnswer : Answer -> Move ;
QSet : Set -> Query ; -- (give me | what are | which are | ) (S | the names of S | S's names) QSet : Set -> Query ; -- (give me | what are | which are | ) (S | the names of S | S's names)
QWhere : Set -> Query ; -- where are S QWhere : Set -> Query ; -- where are S
QWhat : Kind -> Property -> Query ; -- what K P QWhat : Interrogative -> Property -> Query ; -- who P
QWho : Property -> Query ; -- who P QWhatWhat : Interrogative -> Interrogative -> Relation -> Query ; -- who R what
QRel : Relation -> Set -> Query ; -- what R does S have QWhatWhere : Interrogative -> Relation -> Query ; -- who R where --- overgenerating
QInfo : Set -> Query ; -- (give me | ) (information about | all about) S QRelWhere : Set -> Relation -> Query ; -- where does S R --- overgenerating
QCalled : Individual -> Query ; -- how is X (also | otherwise) (called | named | known) ; QFun : Function -> Set -> Query ; -- what R does S have
QWhether : Answer -> Query ; -- is S P --- not in the corpus, but should be ?? QFunPair : Set -> Function -> Query ; -- S and their R's
QInfo : Set -> Query ; -- (give me | ) (information about | all about) S
QCalled : Individual -> Query ; -- how is X (also | otherwise) (called | named | known) ;
QWhether : Answer -> Query ; -- is S P --- not in the corpus, but should be ??
AKind : Set -> Kind -> Answer ; -- S is a K AKind : Set -> Kind -> Answer ; -- S is a K
AInd : Set -> Individual -> Answer ; -- S is I AInd : Set -> Individual -> Answer ; -- S is I
@@ -36,7 +41,7 @@ fun
AProp : Set -> Property -> Answer ; -- S is P AProp : Set -> Property -> Answer ; -- S is P
SAll : Kind -> Set ; -- all Ks | the Ks SAll : Kind -> Set ; -- all Ks | the Ks
SRel : Set -> Relation -> Set ; -- S's Rs SFun : Set -> Function -> Set ; -- S's Rs
SOne : Kind -> Set ; -- one K SOne : Kind -> Set ; -- one K
SIndef : Kind -> Set ; -- a K SIndef : Kind -> Set ; -- a K
SDef : Kind -> Set ; -- the K SDef : Kind -> Set ; -- the K
@@ -45,12 +50,16 @@ fun
SInd : Individual -> Set ; -- X SInd : Individual -> Set ; -- X
SInds : [Individual] -> Set ; -- X and Y SInds : [Individual] -> Set ; -- X and Y
KRelSet : Relation -> Set -> Kind ; -- R of S | S's R IWho : Interrogative ; -- who
KRelsSet : Relation -> Relation -> Set -> Kind ; -- R and Q of S IWhat : Interrogative ; -- what
KRelKind : Kind -> Relation -> Set -> Kind ; -- K that is R of S IWhich : Kind -> Interrogative ; -- which K | what K | which Ks | what Ks
KRelPair : Kind -> Relation -> Kind ; -- S's with their R's
KFunSet : Function -> Set -> Kind ; -- R of S | S's R
KFunsSet : Function -> Function -> Set -> Kind ; -- R and Q of S
KFunKind : Kind -> Function -> Set -> Kind ; -- K that is R of S
KFunPair : Kind -> Function -> Kind ; -- S's with their R's
KProp : Property -> Kind -> Kind ; -- P K | K that is P KProp : Property -> Kind -> Kind ; -- P K | K that is P
KRel : Relation -> Kind ; -- R ---?? KFun : Function -> Kind ; -- R ---??
IName : Name -> Individual ; IName : Name -> Individual ;
@@ -59,6 +68,8 @@ fun
PIs : Individual -> Property ; PIs : Individual -> Property ;
PRelation : Relation -> Set -> Property ;
-- the test lexicon -- the test lexicon
cat cat
@@ -68,32 +79,37 @@ fun
NCountry : Country -> Name ; NCountry : Country -> Name ;
PCountry : Country -> Property ; PCountry : Country -> Property ;
Located : Set -> Property ; Located : Relation ;
In : Set -> Property ; In : Relation ;
HaveTitleAt : JobTitle -> Set -> Property ; HaveTitleAt : JobTitle -> Relation ;
HaveTitle : JobTitle -> Property ; EmployedAt : Set -> Relation ;
Employed : Set -> Property ; HaveTitle : Relation ;
Employed : Relation ;
Named : Name -> Property ; Named : Name -> Property ;
Start : Name -> Property ; Start : Name -> Property ;
Organization : Kind ; Organization : Kind ;
Company : Kind ;
Place : Kind ; Place : Kind ;
Person : Kind ; Person : Kind ;
Location : Relation ; Location : Function ;
Region : Relation ; Region : Function ;
Subregion : Relation ; Subregion : Function ;
USA : Country ; USA : Country ;
California : Country ; California : Country ;
Bulgaria : Country ; Bulgaria : Country ;
OblastSofiya : Name ; OblastSofiya : Name ;
RName : Relation ; FName : Function ;
RNickname : Relation ; FNickname : Function ;
RJobTitle : Relation ; FJobTitle : Function ;
SJobTitle : JobTitle -> Set ; -- a programmer
CEO : JobTitle ; CEO : JobTitle ;
ChiefInformationOfficer : JobTitle ; ChiefInformationOfficer : JobTitle ;

View File

@@ -15,9 +15,11 @@ lincat
Query = Utt ; Query = Utt ;
Answer = Cl ; -- Utt ; Answer = Cl ; -- Utt ;
Set = NP ; Set = NP ;
Relation = {cn : CN ; prep : Prep} ; Interrogative = IP ;
Function = {cn : CN ; prep : Prep} ;
Kind = CN ; Kind = CN ;
Property = {ap : AP ; vp : VP} ; Property = {ap : AP ; vp : VP} ;
Relation = {ap : AP ; vp : VP ; prep : Prep} ;
Individual = NP ; Individual = NP ;
Name = NP ; Name = NP ;
[Individual] = [NP] ; [Individual] = [NP] ;
@@ -39,19 +41,34 @@ lin
| mkUtt (mkQS (mkQCl (L.CompIP (L.IdetIP (mkIDet which_IQuant))) ss)) | mkUtt (mkQS (mkQCl (L.CompIP (L.IdetIP (mkIDet which_IQuant))) ss))
| mkUtt ss ; | mkUtt ss ;
QWhat k p =
mkUtt (mkQS (mkQCl (mkIP (what_IQuant | which_IQuant) (sgNum | plNum) k) p.vp)) ;
QWho p = mkUtt (mkQS (mkQCl whoSg_IP p.vp)) ;
QWhere s = QWhere s =
mkUtt (mkQS (mkQCl where_IAdv s)) mkUtt (mkQS (mkQCl where_IAdv s))
| mkUtt (mkQS (mkQCl where_IAdv (mkCl s (mkA "located" | mkA "situated")))) ; | mkUtt (mkQS (mkQCl where_IAdv (mkCl s (mkA "located" | mkA "situated")))) ;
QRel r s = QWhat i p = mkUtt (mkQS (mkQCl i p.vp)) ;
QWhatWhat i j p = mkUtt (mkQS (QuestQVP i (AdvQVP p.vp (mkIAdv p.prep j)))) ;
QWhatWhere i p = mkUtt (mkQS (QuestQVP i (AdvQVP p.vp where_IAdv))) ;
QRelWhere s p = mkUtt (mkQS (mkQCl where_IAdv (mkCl s p.vp))) ;
QFun r s =
mkUtt mkUtt
(mkImp (mkVP give_V3 (mkNP and_Conj s (mkNP (mkQuant they_Pron) plNum r.cn)) (mkNP i_Pron))) (mkImp (mkVP give_V3
(mkNP and_Conj s (mkNP (mkQuant they_Pron) plNum r.cn)) (mkNP i_Pron)))
| mkUtt (mkQS (mkQCl (mkIP what_IQuant plNum r.cn) s have_V2)) | mkUtt (mkQS (mkQCl (mkIP what_IQuant plNum r.cn) s have_V2))
| mkUtt (mkQS (mkQCl whatSg_IP | mkUtt (mkQS (mkQCl whatSg_IP
(mkClSlash (mkClSlash s have_V2) (mkAdv as_Prep (mkNP aPl_Det r.cn))))) ; (mkClSlash (mkClSlash s have_V2) (mkAdv as_Prep (mkNP aPl_Det r.cn))))) ;
QFunPair s f =
let
ss0 : NP = s
| mkNP (mkNP thePl_Det L.name_N) (mkAdv possess_Prep s) ;
ss : NP = mkNP and_Conj ss0 (mkNP (mkQuant they_Pron) plNum f.cn)
| mkNP ss0 (mkAdv with_Prep (mkNP (mkQuant they_Pron) plNum f.cn))
in
mkUtt (mkImp (mkVP give_V3 ss (mkNP i_Pron)))
| mkUtt (mkQS (mkQCl (L.CompIP whatPl_IP) ss))
| mkUtt (mkQS (mkQCl (L.CompIP (L.IdetIP (mkIDet which_IQuant))) ss))
| mkUtt ss ;
QInfo s = QInfo s =
let let
info : NP = mkNP (all_NP | (mkNP information_N)) (mkAdv about_Prep s) ; info : NP = mkNP (all_NP | (mkNP information_N)) (mkAdv about_Prep s) ;
@@ -72,7 +89,7 @@ lin
AProp s p = (mkCl s p.vp) ; AProp s p = (mkCl s p.vp) ;
SAll k = mkNP all_Predet (mkNP aPl_Det k) | mkNP thePl_Det k ; SAll k = mkNP all_Predet (mkNP aPl_Det k) | mkNP thePl_Det k ;
SRel s r = mkNP (GenNP s) plNum r.cn | mkNP (GenNP s) sgNum r.cn ; SFun s r = mkNP (GenNP s) plNum r.cn | mkNP (GenNP s) sgNum r.cn ;
SOne k = mkNP n1_Numeral k ; SOne k = mkNP n1_Numeral k ;
SIndef k = mkNP a_Det k ; SIndef k = mkNP a_Det k ;
SDef k = mkNP the_Det k ; SDef k = mkNP the_Det k ;
@@ -81,20 +98,27 @@ lin
SInd i = i ; SInd i = i ;
SInds is = mkNP and_Conj is ; SInds is = mkNP and_Conj is ;
KRelSet r s = IWhich k =
mkIP what_IQuant (sgNum | plNum) k
| mkIP which_IQuant (sgNum | plNum) k ;
IWho = whoSg_IP | whoPl_IP ;
IWhat = whatSg_IP | whatPl_IP ;
KFunSet r s =
mkCN r.cn (mkAdv r.prep s) ; mkCN r.cn (mkAdv r.prep s) ;
KRelsSet r q s = KFunsSet r q s =
mkCN (ConjCN and_Conj (BaseCN r.cn q.cn)) (mkAdv r.prep s) ; mkCN (ConjCN and_Conj (BaseCN r.cn q.cn)) (mkAdv r.prep s) ;
KRelKind k r s = KFunKind k r s =
mkCN k (mkRS (mkRCl that_RP (mkVP (mkNP aPl_Det (mkCN r.cn (mkAdv r.prep s)))))) ; mkCN k (mkRS (mkRCl that_RP (mkVP (mkNP aPl_Det (mkCN r.cn (mkAdv r.prep s)))))) ;
KRelPair k r = mkCN k (mkAdv with_Prep (mkNP (mkQuant they_Pron) plNum r.cn)) ; KFunPair k r = mkCN k (mkAdv with_Prep (mkNP (mkQuant they_Pron) plNum r.cn)) ;
KProp p k = KProp p k =
mkCN p.ap k mkCN p.ap k
| mkCN k (mkRS (mkRCl that_RP p.vp)) ; | mkCN k (mkRS (mkRCl that_RP p.vp)) ;
KRel r = r.cn ; KFun r = r.cn ;
IName n = n ; IName n = n ;
@@ -103,6 +127,11 @@ lin
PIs i = propVP (mkVP i) ; PIs i = propVP (mkVP i) ;
PRelation r s = {
ap = AdvAP r.ap (mkAdv r.prep s) ;
vp = mkVP r.vp (mkAdv r.prep s)
} ;
BaseIndividual = mkListNP ; BaseIndividual = mkListNP ;
ConsIndividual = mkListNP ; ConsIndividual = mkListNP ;
@@ -125,7 +154,7 @@ oper
-- lexical constructors -- lexical constructors
mkName : Str -> NP = mkName : Str -> NP =
\s -> mkNP (mkPN s) ; \s -> mkNP (mkPN s) ;
mkRelation : Str -> {cn : CN ; prep : Prep} = mkFunction : Str -> {cn : CN ; prep : Prep} =
\s -> {cn = mkCN (mkN s) ; prep = possess_Prep} ; \s -> {cn = mkCN (mkN s) ; prep = possess_Prep} ;
propAP : AP -> {ap : AP ; vp : VP} = \ap -> { propAP : AP -> {ap : AP ; vp : VP} = \ap -> {
@@ -138,9 +167,22 @@ oper
vp = vp vp = vp
} ; } ;
relAP : AP -> Prep -> {ap : AP ; vp : VP ; prep : Prep} = \ap,p -> {
ap = ap ;
vp = mkVP ap ;
prep = p
} ;
relVP : VP -> Prep -> {ap : AP ; vp : VP ; prep : Prep} = \vp,p -> {
ap = PartVP vp ;
vp = vp ;
prep = p
} ;
propCalled : NP -> {ap : AP ; vp : VP} = \i -> propCalled : NP -> {ap : AP ; vp : VP} = \i ->
propAP (mkAP (also_AdA | otherwise_AdA) (mkAP (mkA2 called_A []) i)) ; propAP (mkAP (also_AdA | otherwise_AdA) (mkAP (mkA2 called_A []) i)) ;
noPrep : Prep = mkPrep [] ;
-- lexicon -- lexicon
@@ -151,65 +193,63 @@ lin
NCountry c = c.np ; NCountry c = c.np ;
PCountry c = propAP (mkAP c.a) ; PCountry c = propAP (mkAP c.a) ;
Located i = propAP ( Located =
mkAP (mkA2 (mkA "located") in_Prep) i relAP (mkAP (mkA "located")) in_Prep
| mkAP (mkA2 (mkA "situated") in_Prep) i | relAP (mkAP (mkA "situated")) in_Prep
) ; ;
In i = propVP (mkVP (mkAdv in_Prep i)) ; In = relVP UseCopula in_Prep ;
Employed i = propAP ( Employed =
mkAP (mkA2 (mkA "employed") by8agent_Prep) i relAP (mkAP (mkA "employed")) by8agent_Prep
| mkAP (mkA2 (mkA "paid") by8agent_Prep) i | relAP (mkAP (mkA "paid")) by8agent_Prep
| mkAP (mkA2 (mkA "active") at_Prep) i | relAP (mkAP (mkA "active")) at_Prep
| mkAP (mkA2 (mkA "professionally active") at_Prep) i | relAP (mkAP (mkA "professionally active")) at_Prep
) | relVP (mkVP (mkV "work")) at_Prep
| | relVP (mkVP (mkV "collaborate")) in_Prep
propVP ( ;
mkVP (mkV2 (mkV "work") at_Prep) i
| mkVP (mkV2 (mkV "collaborate") in_Prep) i
) ;
HaveTitle t = HaveTitle =
propAP ( relAP (mkAP (mkA "employed")) as_Prep
mkAP (mkA2 (mkA "employed") as_Prep) (mkNP t) --- | relVP UseCopula noPrep
) | relVP (mkVP (mkV "work")) as_Prep
| | relVP (mkVP have_V2 (mkNP the_Det (mkCN (mkN2 (mkN "title"))))) possess_Prep
propVP ( ;
mkVP (mkNP a_Det t)
| mkVP (mkV2 (mkV "work") as_Prep) (mkNP t)
-- | mkVP have_V2 (mkNP the_Det (mkCN (mkN2 (mkN "title")) (mkNP t)))
) ;
HaveTitleAt t i = EmployedAt s =
propAP ( relAP (mkAP (mkA2 (mkA "employed") at_Prep) s) as_Prep
mkAP (mkA2 (mkA "employed") as_Prep) (mkNP (mkNP t) (mkAdv at_Prep i)) | relAP (mkAP (mkA2 (mkA "employed") by8agent_Prep) s) as_Prep
) | relVP (mkVP (mkV2 (mkV "work") at_Prep) s) as_Prep
| ;
propVP (
mkVP (mkVP (mkNP a_Det t)) (mkAdv at_Prep i) HaveTitleAt t =
| mkVP (mkVP (mkV2 (mkV "work") as_Prep) (mkNP t)) (mkAdv at_Prep i) relAP (mkAP (mkA2 (mkA "employed") as_Prep) (mkNP t)) at_Prep
-- | mkVP (mkVP have_V2 (mkNP the_Det (mkCN (mkN2 (mkN "title")) (mkNP t)))) | relAP (mkAP (mkA2 (mkA "employed") as_Prep) (mkNP t)) by8agent_Prep
-- (mkAdv at_Prep i) | relVP (mkVP (mkNP a_Det t)) at_Prep
) ; | relVP (mkVP (mkV2 (mkV "work") as_Prep) (mkNP t)) at_Prep
| relVP (mkVP have_V2 (mkNP the_Det (mkCN (mkN2 (mkN "title")) (mkNP t)))) at_Prep
;
Named n = propAP (mkAP (mkA2 (mkA "named") []) n) ; Named n = propAP (mkAP (mkA2 (mkA "named") []) n) ;
Start n = propVP (mkVP (mkV2 "start" with_Prep) n) ; Start n = propVP (mkVP (mkV2 "start" with_Prep) n) ;
Organization = mkCN (mkN "organization") ; Organization = mkCN (mkN "organization") ;
Company = mkCN (mkN "company") ;
Place = mkCN (mkN "place") ; Place = mkCN (mkN "place") ;
Person = Person =
mkCN (mkN "person" "people") mkCN (mkN "person" "people")
| mkCN (mkN "person") ; | mkCN (mkN "person") ;
Location = mkRelation "location" ; Location = mkFunction "location" ;
Region = mkRelation "region" ; Region = mkFunction "region" ;
Subregion = mkRelation "subregion" | mkRelation "sub-region" ; Subregion = mkFunction "subregion" | mkFunction "sub-region" ;
RName = mkRelation "name" ; FName = mkFunction "name" ;
RNickname = mkRelation "nickname" ; FNickname = mkFunction "nickname" ;
RJobTitle = mkRelation "job title" | mkRelation "job" | mkRelation "position" | FJobTitle = mkFunction "job title" | mkFunction "job" | mkFunction "position" |
mkRelation "appointment" | mkRelation "job position" | mkRelation "mandate" | mkFunction "appointment" | mkFunction "job position" | mkFunction "mandate" |
mkRelation "title" ; mkFunction "title" | mkFunction "capacity" ;
SJobTitle t = mkNP a_Det t ;
USA = mkCountry "USA" "American" ; USA = mkCountry "USA" "American" ;
Bulgaria = mkCountry "Bulgaria" "Bulgarian" ; Bulgaria = mkCountry "Bulgaria" "Bulgarian" ;

57
examples/query/README Normal file
View File

@@ -0,0 +1,57 @@
Copyright (c) 2010 Aarne Ranta, under LGPL(3).
Part of MOLTO Project, WP 4.
Query language, based on the corpus from Ontotext.
Purpose: natural language queries to an ontology database.
Work in progress:
- 19 June parsing 28% 160/562
- 17 June 2010 first version, parsing under 10%
The corpus contains misspellings and ungrammatical sentences; these will (mostly) not
be covered by the grammar.
Test:
-- start GF with the grammar; notice that lib/present/ must have latest Eng libraries,
-- which can be provided by 'runghc Make present lang api langs=Eng' in lib/src/
% gf QueryEng.gf
-- parse a sentence and see all variants
> "p "Bulgarian people working at Google" | l -all
Regression test:
-- run the parser on the corpus
% gf QueryEng.gf <test.gfs > test.results8
-- compute the number of sentences not covered
% grep "no tree" test.results8 | wc
Semantics: generic logical semantics, that could be mapped to many query languages.
The denotations of the main categories are, assuming a domain of individuals:
Set ; P(P(D)) (generalized quantifier) -- the set requested, e.g. "all persons"
Function ; D -> P(D) -- something of something, e.g. "subregion of Bulgaria"
Kind ; P(D) -- type of things, e.g. "person"
Relation ; D -> D -> T -- relation between things,e.g. "employed at"
Property ; D -> T -- property of things, e.g. "employed at Google"
Individual ; D -- one entity, e.g. "Google"
Name ; D -- person, company... e.g. "Eric Schmidt"
Characteristics:
- simple AST's, lots of variants (easily hundreds per query)
- semantic overgeneration, e.g. "Google works at Larry Page"
- some ambiguities, e.g.
> give me the organizations and their locations
MQuery (QFun Location (SAll Organization))
MQuery (QFunPair (SAll Organization) Location)
Maybe harmless?
Resource grammar was not quite enough; added for instance multiple interrogatives
("who is employed as what where") in Extra and ExtraEng