Uzupełnienie zbioru selektorów

Wojciech Jaworski
1 parent 28b17074
Showing 11 changed files with 52 additions and 10 deletions
LCGlexicon/ENIAM_LCGlexicon.ml
LCGlexicon/ENIAM_LCGlexiconTypes.ml
LCGlexicon/ENIAMcategoriesPL.ml
semantics/ENIAMsemGraph.ml
semantics/ENIAMsemGraphOf.ml
semantics/ENIAMsemLatexOf.ml
semantics/ENIAMsemLexicon.ml
semantics/ENIAMsemStringOf.ml
semantics/ENIAMsemTypes.ml
semantics/ENIAMsemValence.ml
semantics/ENIAMsemXMLof.ml
@@ -251,10 +251,12 @@ let make_node id orth lemma pos syntax weight cat_list is_raised =
       | Ctype -> ("CTYPE", SubstVar "ctype") :: attrs
       | Mode -> ("MODE", SubstVar "mode") :: attrs
       | Psem -> ("PSEM", SubstVar "psem") :: attrs
+      | Icat -> attrs
       | Inumber -> attrs
       | Igender -> attrs
       | Iperson -> attrs
       | Nperson -> attrs
+      | Ncat -> attrs
       | Plemma -> attrs
       | Unumber -> attrs
       | Ucase -> attrs
@@ -27,7 +27,7 @@ type categories = {lemma: string; pos: string; pos2: string; cat: string; coerce
 type selector =
     Lemma | (*NewLemma |*) Pos | Pos2 | Cat | Coerced | Number | Case | Gender | Person | Grad | Praep |
     Acm | Aspect | Negation | Mood | Tense | Nsyn | Nsem | Ctype | Mode | Psem |
-    Inumber | Igender | Iperson | Nperson | Plemma |
+    Icat | Inumber | Igender | Iperson | Nperson | Ncat | Plemma |
     Unumber | Ucase | Ugender | Uperson | Amode
  
 module OrderedSelector = struct
@@ -431,7 +431,7 @@ let clarify_categories proper cat coerced = function
 let selector_names = StringSet.of_list [
     "lemma";"pos";"pos2";"cat";"coerced";"number";"case";"gender";"person";"grad";
     "praep";"acm";"aspect";"negation";"mood";"tense";"nsyn";"nsem";"ctype";"mode";"psem";
-    "inumber";"igender";"iperson";"nperson";"plemma";
+    "icat";"inumber";"igender";"iperson";"nperson";"ncat";"plemma";
     "unumber";"ucase";"ugender";"uperson";"amode"]
  
  
@@ -458,10 +458,12 @@ let string_of_selector = function
   | Ctype -> "ctype"
   | Mode -> "mode"
   | Psem -> "psem"
+  | Icat -> "icat"
   | Inumber -> "inumber"
   | Igender -> "igender"
   | Iperson -> "iperson"
   | Nperson -> "nperson"
+  | Ncat -> "ncat"
   | Plemma -> "plemma"
   | Unumber -> "unumber"
   | Ucase -> "ucase"
@@ -497,10 +499,12 @@ let selector_of_string = function
   | "ctype" -> Ctype
   | "mode" -> Mode
   | "psem" -> Psem
+  | "icat" -> Icat
   | "inumber" -> Inumber
   | "igender" -> Igender
   | "iperson" -> Iperson
   | "nperson" -> Nperson
+  | "ncat" -> Ncat
   | "plemma" -> Plemma
   | "unumber" -> Unumber
   | "ucase" -> Ucase
@@ -34,7 +34,7 @@ let rec make_args_list = function
 let symbols = StringSet.of_list [
   "symbol"; "date"; "date-interval"; "hour-minute"; "hour"; "hour-minute-interval"; "hour-interval";
   "year"; "year-interval"; "day"; "day-interval"; "day-month"; "day-month-interval"; "month-interval"; "roman"; "roman-interval";
-  "match-result"; "url"; "email"; "obj-id"; "building-number";
+  "match-result"; "url"; "email"; "phone-number"; "obj-id"; "building-number";
   "month-lex"; "day-lex"]
  
 let rec get_person = function
@@ -172,6 +172,7 @@ let create_normal_concept tokens lex_sems t cat coerced =
       | "GRAD",Val "com" -> {c with c_relations=Tuple[c.c_relations;SingleRelation (Val "com")]}
       | "GRAD",Val "sup" -> {c with c_relations=Tuple[c.c_relations;SingleRelation (Val "sup")]}
       | "ASPECT",_ -> c
+      | "CTYPE",_ -> c (* FIXME1: trzeba zaznaczyć pytajność w grafie, CTYPE pojawia się w dwu węzłach *)
 (*       | "TYPE",Val "int" -> {c with c_quant=Tuple[c.c_quant;Val "interrogative"]} *)
       | "TYPE",_ -> c (* FIXME *)
       | "PERS",_ -> c
@@ -273,13 +274,19 @@ let create_normal_concept tokens lex_sems t cat coerced =
     make_relation t (Context cx) else
   if t.lemma = "<root>" then t.args else
   if t.lemma = "<merge>" then RemoveRelation("null","",t.args) else
-  if t.pos = "interp" && t.lemma = "?" && t.args = Dot then SingleRelation(Val "int") else
+  (* if t.pos = "interp" && t.lemma = "?" && t.args = Dot then SingleRelation(Val "int") else *)
+  if t.pos = "interp" && t.lemma = "?" then
+    make_relation t (AddSingleRelation(Val "int",RemoveRelation("CORE","",t.args))) else (* FIXME1: to powinno tworzyć kontekst i zaznaczać ze jest interrogative *)
   if t.pos = "interp" && t.lemma = ":" then
     make_relation t (RemoveRelation("CORE","",t.args)) else
   if t.pos = "interp" && t.lemma = "</sentence>" then
     let l = (*List.rev*) (make_args_list t.args) in
     Xlist.fold (List.tl l) (RemoveRelation("null","",List.hd l)) (fun t s -> AddRelation(t,"Next","Clause",RemoveRelation("null","",s))) else
   if t.pos = "interp" && t.lemma = "<sentence>" then t.args else
+  if t.pos = "interp" && t.lemma = "</query>" then
+    let l = (*List.rev*) (make_args_list t.args) in
+    Xlist.fold (List.tl l) (List.hd l) (fun t s -> AddRelation(t,"Next","Sentence",s)) else
+  if t.pos = "interp" && t.lemma = "<query>" then t.args else
 (*  if t.pos = "interp" && t.lemma = "”s" then
     let l = List.rev (make_args_list t.args) in
     let x = Xlist.fold (List.tl l) (List.hd l) (fun t s -> AddRelation(RemoveRelation t,"Next","Sentence",RemoveRelation s)) in
@@ -340,6 +347,7 @@ let rec translate_node tokens lex_sems t =
     | "NEGATION",s -> t,("NEGATION",s) :: attrs,cat,coerced
     | "MOOD",s -> t,("MOOD",s) :: attrs,cat,coerced
     | "TENSE",s -> t,("TENSE",s) :: attrs,cat,coerced
+    | "CTYPE",s -> t,("CTYPE",s) :: attrs,cat,coerced
     | "controller",s -> t,("controller",s) :: attrs,cat,coerced
     | "controllee",s -> t,("controllee",s) :: attrs,cat,coerced
     | "coref",s -> t,attrs,cat,coerced
@@ -388,6 +396,7 @@ let rec make_tree_rec references = function
   | SingleRelation r  -> SingleRelation r
   (* | TripleRelation(r,a,s,t) -> TripleRelation(r,a,make_tree_rec references s,make_tree_rec references t) *)
   | AddRelation(t,r,a,s) -> AddRelation(make_tree_rec references t,r,a,make_tree_rec references s)
+  | AddSingleRelation(r,s) -> AddSingleRelation(r,make_tree_rec references s)
   | RemoveRelation(r,a,t) -> RemoveRelation(r,a,make_tree_rec references t)
   | SetContextName(s,t) -> SetContextName(s,make_tree_rec references t)
   | CreateContext(s,t) -> CreateContext(s,make_tree_rec references t)
@@ -414,6 +423,7 @@ let rec validate_translation r = function
   | SingleRelation _  -> ()
   (* | TripleRelation(_,_,s,t) -> validate_translation r s; validate_translation r t *)
   | AddRelation(t,_,_,s) -> validate_translation r t; validate_translation r s
+  | AddSingleRelation(_,s) -> validate_translation r s
   | RemoveRelation(_,_,t) -> validate_translation r t
   | SetContextName(s,t) -> validate_translation r t
   | CreateContext(s,t) -> validate_translation r t
@@ -508,6 +518,13 @@ let rec reduce_tree = function
   (* | TripleRelation(r,a,s,t) -> TripleRelation(r,a,reduce_tree s,reduce_tree t) *)
 (*  | AddRelation(Concept c,r,a,s) -> reduce_tree (Concept{c with c_relations=Tuple[Relation(Val r,Val a,s);c.c_relations]})
   | AddRelation(Context c,r,a,s) -> reduce_tree (Context{c with cx_relations=Tuple[Relation(Val r,Val a,s);c.cx_relations]})*)
+  | AddSingleRelation(r,t) ->
+      (match reduce_tree t with
+        Concept t -> Concept{t with c_relations=Tuple[t.c_relations;SingleRelation r]}
+      | Context({cx_sense=Val "czy"} as t) -> Context t
+      | Context({cx_sense=Val "jaki"} as t) -> Context t
+      | Context({cx_sense=Dot} as t) -> Context{t with cx_sense=Val "czy"}
+      | t -> AddSingleRelation(r,t))
   | AddRelation(t,r,a,s) -> simplify_tree_add_relation r a (reduce_tree s) (reduce_tree t)
 (*      let t = reduce_tree t in
       let s = reduce_tree s in
@@ -270,6 +270,13 @@ let rec print_graph2_rec file edge_rev edge_label edge_style edge_head upper = f
           let _ = print_graph2_rec file false "" "" "" id t in
           let _ = print_graph2_rec file false "" "" "" id s in
           id
+  | AddSingleRelation(role,t) ->
+          let id = !id_counter in
+          incr id_counter;
+          fprintf file "  %d [shape=circle,label=\"AddSingleRelation\\n%s\"]\n" id (ENIAMsemStringOf.linear_term 0 role);
+          print_edge2 file edge_rev edge_label edge_style edge_head "" upper id;
+          let _ = print_graph2_rec file false "" "" "" id t in
+          id
   | RemoveRelation(role,role_attr,t) ->
           let id = !id_counter in
           incr id_counter;
@@ -75,6 +75,7 @@ let rec linear_term c = function
   | SingleRelation r -> "{\\bf singlerelation}(" ^ linear_term 0 r ^ ")"
   (* | TripleRelation(r,a,c,t) -> "{\\bf triplerelation}(" ^ (*linear_term 0*) r ^ "," ^ (*linear_term 0*) a ^ "," ^ linear_term 0 c ^ "," ^ linear_term 0 t ^ ")" *)
   | AddRelation(t,r,a,s) -> "{\\bf addrelation}(" ^ linear_term 0 t ^ "," ^ r ^ "," ^ a ^ "," ^ linear_term 0 s ^ ")"
+  | AddSingleRelation(r,s) -> "{\\bf addrelation}(" ^ linear_term 0 r ^ "," ^ linear_term 0 s ^ ")"
   | RemoveRelation(r,a,t) -> "{\\bf removerelation}(" ^ r ^ "," ^ a ^ "," ^ linear_term 0 t ^ ")"
   | SetContextName(s,t) -> "{\\bf setcontextname}(" ^ linear_term 0 s ^ "," ^ linear_term 0 t ^ ")"
   | CreateContext(s,t) -> "{\\bf createcontext}(" ^ linear_term 0 (Context s) ^ "," ^ linear_term 0 t ^ ")"
@@ -45,7 +45,7 @@ let parse_multi p = function
   | tokens -> tokens,p
  
 let parse_morf p = function
-    [T "1"] -> {p with is_necessary=Opt}
+    [T "1"] -> if p.is_necessary=Multi then p else {p with is_necessary=Opt}
   | tokens ->
       let l = Xlist.map (try Lexer.split_symbol (T "*") [] tokens with _ -> failwith "parse_morf: split_symbol *") (function
           [T s] -> Atom s
@@ -52,6 +52,7 @@ let rec linear_term c = function
   | SingleRelation r -> "singlerelation(" ^ linear_term 0 r ^ ")"
   (* | TripleRelation(r,a,c,t) -> "triplerelation(" ^ r ^ "," ^ a ^ "," ^ linear_term 0 c ^ "," ^ linear_term 0 t ^ ")" *)
   | AddRelation(t,r,a,s) -> "addrelation(" ^ linear_term 0 t ^ "," ^ r ^ "," ^ a ^ "," ^ linear_term 0 s ^ ")"
+  | AddSingleRelation(r,s) -> "addsinglerelation(" ^ linear_term 0 r ^ "," ^ linear_term 0 s ^ ")"
   | RemoveRelation(r,a,t) -> "removerelation(" ^ r ^ "," ^ a ^ "," ^ linear_term 0 t ^ ")"
   | SetContextName(s,t) -> "setcontextname(" ^ linear_term 0 s ^ "," ^ linear_term 0 t ^ ")"
   | CreateContext(s,t) -> "createcontext(" ^ linear_term 0 (Context s) ^ "," ^ linear_term 0 t ^ ")"
@@ -72,7 +72,8 @@ and linear_term =
   | RevRelation of string * string * linear_term (* role * role_attr * concept *)
   | SingleRelation of linear_term
   (* | TripleRelation of string * string * linear_term * linear_term (* role * role_attr * concept *) *)
-  | AddRelation of linear_term * string * string * linear_term (* nadrządnik * role * role_attr * podrzędnik *)
+  | AddRelation of linear_term * string * string * linear_term (* nadrzędnik * role * role_attr * podrzędnik *)
+  | AddSingleRelation of linear_term * linear_term (* role * podrzędnik *)
   | RemoveRelation of string * string * linear_term
   | SetContextName of linear_term * linear_term (* sense * concept *)
   | CreateContext of context * linear_term (* context * args *)
@@ -220,7 +220,9 @@ let string_of_arg arg =
   String.concat ", " (Xlist.map arg (fun ((arg_symbol,dir),t) -> (string_of_argdir dir) ^ arg_symbol ^ ":" ^ ENIAM_LCGstringOf.linear_term 0 t))
  
 let string_of_position p =
-  (string_of_argdir p.dir) ^ String.concat "+" (StringSet.to_list p.morfs)
+  (string_of_argdir p.dir) ^
+  (if p.is_multi then "?" else "") ^
+  String.concat "+" (StringSet.to_list p.morfs)
  
 let manage_arg p t =
   let t = SetAttr("gf",Val (ENIAMwalStringOf.gf p.gf),t) in
@@ -383,8 +385,10 @@ let rec assign_frames_rec tokens lex_sems tree arg_symbols visited = function
           let frame = ENIAMsemLexicon.extend_frame phsymbol frame in
           (* print_endline "passed"; *)
           (attrs,symbol,frame,Xlist.rev_map frame.positions (translate_position (string_of_int t.id))) :: frames
-        with Not_found -> (*print_endline "rejected";*) frames) in
-      if frames = [] then failwith "assign_frames_rec: no frame" else
+        with Not_found ->
+          (* print_endline "rejected"; *)
+          frames) in
+      if frames = [] then failwith ("assign_frames_rec: no frame phsymbol='" ^ phsymbol ^ "' node='" ^ t.lemma ^ "'") else
       let prong_attrs = get_prong_attrs t.attrs in
       let e = ENIAM_LCGreductions.get_variant_label () in
       let l,_ = Xlist.fold frames ([],1) (fun (l,n) (attrs,symbol,frame,positions) ->
@@ -404,7 +408,9 @@ let rec assign_frames_rec tokens lex_sems tree arg_symbols visited = function
               ("fopinion",Val (ENIAMwalStringOf.opinion frame.fopinion)) ::
               ("sopinion",Val (ENIAMwalStringOf.opinion frame.sopinion)) :: attrs; args=args; symbol=symbol}) ::
               l,n+1))) in
-      if l = [] then ((*print_endline ("assign_frames_rec 4: no frame assingment found for " ^ t.lemma ^ " " ^ ENIAM_LCGstringOf.linear_term 0 t.symbol);*)raise (NoFrame(t.lemma,ENIAM_LCGstringOf.linear_term 0 t.symbol,visited))) else
+      if l = [] then (
+        (* print_endline ("assign_frames_rec 4: no frame assingment found for " ^ t.lemma ^ " " ^ ENIAM_LCGstringOf.linear_term 0 t.symbol); *)
+        raise (NoFrame(t.lemma,ENIAM_LCGstringOf.linear_term 0 t.symbol,visited))) else
       Variant(e,l),visited
   | Variant(e,l) ->
       let a = ref "" in
@@ -61,6 +61,9 @@ let rec linear_term = function
   | AddRelation(t,r,a,s) ->
       Xml.Element("AddRelation",["role",r;"role_attribute",a],
         [Xml.Element("",[],[linear_term t]);Xml.Element("",[],[linear_term s])])
+  | AddSingleRelation(r,s) ->
+      Xml.Element("AddSingleRelation",[],
+        [Xml.Element("",[],[linear_term r]);Xml.Element("",[],[linear_term s])])
   | RemoveRelation(r,a,t) -> Xml.Element("RemoveRelation",["role",r;"role_attribute",a],[linear_term t])
   | SetContextName(s,t) ->
       Xml.Element("SetContextName",[],[linear_term s;linear_term t])