Add more datatypes

lszeremeta · Nov 30, 2021 · 4a52868 · 4a52868
1 parent 96b38ee
commit 4a52868
Show file tree

Hide file tree

Showing 3 changed files with 82 additions and 44 deletions.
diff --git a/other-notations/YARSpg.ebnf b/other-notations/YARSpg.ebnf
@@ -33,7 +33,7 @@
  IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 */
 
-yarspg   ::= (statement metadata?)*
+yarspg   ::= ( statement metadata? )*
 statement
          ::= node
            | edge
@@ -43,39 +43,35 @@ statement
            | edge_schema
            | variable_declaration_schema
            | graph
-           | graphs_schema
-metadata
-         ::= '+' prop_list
+           | graph_schema
+metadata ::= '+' prop_list
 variable ::= '$' variable_name
 variable_declaration
          ::= variable '=' prop ( ',' prop )*
 variable_declaration_schema
          ::= variable '=' prop_schema ( ',' prop_schema )*
 variable_name
          ::= ALNUMPLUS
-graph_id
-         ::= DEFAULT
+graph_id ::= DEFAULT
            | ALNUMPLUS
 prop_list
-         ::= ( '[' ( ( prop | variable ) ( ',' ( prop | variable ) )* )? ']' )?
+         ::= '[' ( ( prop | variable ) ( ',' ( prop | variable ) )* )? ']'
 meta_prop
          ::= '@' '<' ( key ':' value ( ',' key ':' value )* )? '>'
 graphs_list
-         ::= '/' graph_id (',' graph_id)* '/'
+         ::= '/' graph_id ( ',' graph_id )* '/'
 graph    ::= '/' graph_id '/' ( '{' ( graph_label ( ',' graph_label )* )? '}' )? prop_list?
-node
-         ::= '(' node_id ( '{' ( node_label ( ',' node_label )* )? '}' )? prop_list? ')' graphs_list?
-edge
-         ::= directed
+node     ::= '(' node_id ( '{' ( node_label ( ',' node_label )* )? '}' )? prop_list? ')' graphs_list?
+edge     ::= directed
            | undirected
 directed ::= '(' node_id ')' '-' '(' edge_id? ( '{' ( edge_label ( ',' edge_label )* )? '}' )? prop_list? ')' '->' '(' node_id ')' graphs_list?
 undirected
          ::= '(' node_id ')' '-' '(' edge_id? ( '{' ( edge_label ( ',' edge_label )* )? '}' )? prop_list? ')' '-' '(' node_id ')' graphs_list?
 node_id  ::= ALNUMPLUS
 node_label
          ::= STRING
-prop     ::= key ':' value meta_prop?       
-edge_id   ::= ALNUMPLUS
+prop     ::= key ':' value meta_prop?
+edge_id  ::= ALNUMPLUS
 edge_label
          ::= STRING
 graph_label
@@ -89,17 +85,20 @@ complex_value
          ::= set_value
            | list_value
            | struct_value
-set_value      ::= '{' ( primitive_value | set_value ) meta_prop? ( ',' ( primitive_value | set_value ) meta_prop? )* '}'
-list_value     ::= '[' ( primitive_value | list_value ) meta_prop? ( ',' ( primitive_value | list_value ) meta_prop? )* ']'
-struct_value   ::= '{' key ':' ( primitive_value | struct_value ) meta_prop? ( ',' key ':' ( primitive_value | struct_value ) meta_prop? )* '}'
+set_value
+         ::= '{' ( primitive_value | set_value ) meta_prop? ( ',' ( primitive_value | set_value ) meta_prop? )* '}'
+list_value
+         ::= '[' ( primitive_value | list_value ) meta_prop? ( ',' ( primitive_value | list_value ) meta_prop? )* ']'
+struct_value
+         ::= '{' key ':' ( primitive_value | struct_value ) meta_prop? ( ',' key ':' ( primitive_value | struct_value ) meta_prop? )* '}'
 node_schema
          ::= 'S' '(' node_id_schema ( '{' ( node_label ( ',' node_label )* )? '}' )? prop_list_schema? ')' graphs_list?
 node_id_schema
          ::= ALNUMPLUS
 prop_list_schema
          ::= '[' ( ( prop_schema | variable ) ( ',' ( prop_schema | variable ) )* )? ']'
 prop_schema
-         ::= key ':' value_schema id_key?
+         ::= key ':' value_schema
 meta_prop_schema
          ::= '@' '<' ( key ':' value_schema ( ',' key ':' value_schema )* )? '>'
 cardinality
@@ -111,30 +110,40 @@ max_cardinality
          ::= MAX card_num
 card_num ::= UNSIGNED_INT
 value_schema
-         ::= primitive_value_schema ( UNIQUE | NULL )? OPTIONAL? meta_prop_schema?
-           | complex_value_schema NULL? OPTIONAL? meta_prop_schema?
+         ::= ( primitive_value_schema ( UNIQUE | NULL )? | complex_value_schema NULL? ) OPTIONAL? meta_prop_schema?
 primitive_value_schema
          ::= DECIMAL
            | SMALLINT
            | INTEGER
+           | UINTEGER
            | BIGINT
            | FLOAT
            | REAL
            | DOUBLE
            | BOOL
-           | STRING_TYPE
+           | STRINGT
            | DATE
            | DATETIME
+           | LOCALDATETIME
            | TIME
+           | LOCALTIME
+           | DURATION
+           | BINARY
            | ALNUMPLUS
 complex_value_schema
-         ::= set_schema
+         ::= multiset_schema
+           | set_schema
            | list_schema
+           | listd_schema
            | struct_schema
+multiset_schema
+         ::= MULTISET '(' ( primitive_value_schema | set_schema ) NULL? cardinality? ')' meta_prop_schema?
 set_schema
          ::= SET '(' ( primitive_value_schema | set_schema ) NULL? cardinality? ')' meta_prop_schema?
 list_schema
          ::= LIST '(' ( primitive_value_schema | list_schema ) NULL? cardinality? ')' meta_prop_schema?
+listd_schema
+         ::= LISTD '(' ( primitive_value_schema | list_schema ) NULL? cardinality? ')' meta_prop_schema?
 struct_schema
          ::= STRUCT '(' ( primitive_value_schema | struct_schema ) NULL? cardinality? ')' meta_prop_schema?
 edge_schema
@@ -143,7 +152,7 @@ edge_schema
 directed_schema
          ::= 'S' '(' node_id_schema ')' '-' ( '(' ( '{' ( edge_label ( ',' edge_label )* )? '}' )? prop_list_schema? ')' )? '->' '(' node_id_schema ')' graphs_list?
 undirected_schema
-         ::= 'S' '(' node_id_schema ')' '-' ( '(' ( '{' ( edge_label ( ',' edge_label )* )? '}' )? prop_list_schema? ')' )? '-' '(' node_id_schema1 ')' graphs_list?
+         ::= 'S' '(' node_id_schema ')' '-' ( '(' ( '{' ( edge_label ( ',' edge_label )* )? '}' )? prop_list_schema? ')' )? '-' '(' node_id_schema ')' graphs_list?
 graph_schema
          ::= 'S' '/' graph_id '/' prop_list_schema?
 
@@ -158,26 +167,34 @@ OPTIONAL ::= ( 'O' | 'o' ) ( 'P' | 'p' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'O' | 'o'
 DECIMAL  ::= ( 'D' | 'd' ) ( 'E' | 'e' ) ( 'C' | 'c' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'A' | 'a' ) ( 'L' | 'l' )
 SMALLINT ::= ( 'S' | 's' ) ( 'M' | 'm' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' )
 INTEGER  ::= ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'G' | 'g' ) ( 'E' | 'e' ) ( 'R' | 'r' )
+UINTEGER ::= ( 'U' | 'u' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'G' | 'g' ) ( 'E' | 'e' ) ( 'R' | 'r' )
 BIGINT   ::= ( 'B' | 'b' ) ( 'I' | 'i' ) ( 'G' | 'g' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' )
 FLOAT    ::= ( 'F' | 'f' ) ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'A' | 'a' ) ( 'T' | 't' )
 REAL     ::= ( 'R' | 'r' ) ( 'E' | 'e' ) ( 'A' | 'a' ) ( 'L' | 'l' )
 DOUBLE   ::= ( 'D' | 'd' ) ( 'O' | 'o' ) ( 'U' | 'u' ) ( 'B' | 'b' ) ( 'L' | 'l' ) ( 'E' | 'e' )
 BOOL     ::= ( 'B' | 'b' ) ( 'O' | 'o' ) ( 'O' | 'o' ) ( 'L' | 'l' )
-STRING_TYPE
-         ::= ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'G' | 'g' )
+STRINGT  ::= ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'G' | 'g' )
 DATE     ::= ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' )
 DATETIME ::= ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' )
+LOCALDATETIME
+         ::= ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'C' | 'c' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' )
 TIME     ::= ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' )
+LOCALTIME
+         ::= ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'C' | 'c' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' )
+DURATION ::= ( 'D' | 'd' ) ( 'U' | 'u' ) ( 'R' | 'r' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'O' | 'o' ) ( 'N' | 'n' )
+BINARY   ::= ( 'B' | 'b' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'A' | 'a' ) ( 'R' | 'r' ) ( 'Y' | 'y' )
 
+MULTISET ::= ( 'M' | 'm' ) ( 'U' | 'u' ) ( 'L' | 'l' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'E' | 'e' ) ( 'T' | 't' )
 SET      ::= ( 'S' | 's' ) ( 'E' | 'e' ) ( 'T' | 't' )
 LIST     ::= ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'T' | 't' )
+LISTD    ::= ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'T' | 't' ) ( 'D' | 'd' )
 STRUCT   ::= ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'U' | 'u' ) ( 'C' | 'c' ) ( 'T' | 't' )
 /* END CASE-INSENSITIVE NAMES */
 
 COMMENT  ::= '#' [^#xd#xa#xc]*
-STRING   ::= '"' ([^"\#xd#xa] | "'" | '\"')* '"'
+STRING   ::= '"' ( [^"\#xd#xa] | "'" | '\"' )* '"'
 UNSIGNED_INT
-         ::= [0-9]*
+         ::= [0-9]+
 ALNUMPLUS
-         ::= [a-zA-Z_][a-zA-Z0-9_]*
-WS       ::= ([#x20#x9#xa])+
+         ::= [a-zA-Z_] [a-zA-Z0-9_]*
+WS       ::= [ #x9#xa]+
diff --git a/yarspg/YARSpg.g4 b/yarspg/YARSpg.g4
@@ -198,31 +198,38 @@ value_schema
     ;
 
 primitive_value_schema
-    // NUMBER
     : DECIMAL
     | SMALLINT
     | INTEGER
+    | UINTEGER
     | BIGINT
     | FLOAT
     | REAL
     | DOUBLE
-    // BOOL
     | BOOL
-    // STRING
-    | STRING_TYPE
-    // DATETYPE
+    | STRINGT
     | DATE
-    | DATETIME    
+    | DATETIME
+    | LOCALDATETIME 
     | TIME
+    | LOCALTIME
+    | DURATION
+    | BINARY
     | ALNUMPLUS
     ;
 
 complex_value_schema
-    : set_schema
+    : multiset_schema
+    | set_schema
     | list_schema
+    | listd_schema
     | struct_schema
     ;
 
+multiset_schema
+    : MULTISET '(' (primitive_value_schema | set_schema) NULL? cardinality? ')' meta_prop_schema?
+    ;
+
 set_schema
     : SET '(' (primitive_value_schema | set_schema) NULL? cardinality? ')' meta_prop_schema?
     ;
@@ -231,6 +238,10 @@ list_schema
     : LIST '(' (primitive_value_schema | list_schema) NULL? cardinality? ')' meta_prop_schema?
     ;
 
+listd_schema
+    : LISTD '(' (primitive_value_schema | list_schema) NULL? cardinality? ')' meta_prop_schema?
+    ;
+
 struct_schema
     : STRUCT '(' (primitive_value_schema | struct_schema) NULL? cardinality? ')' meta_prop_schema?
     ;
@@ -264,19 +275,26 @@ OPTIONAL: ( 'O' | 'o' ) ( 'P' | 'p' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'O' | 'o' )
 DECIMAL: ( 'D' | 'd' ) ( 'E' | 'e' ) ( 'C' | 'c' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'A' | 'a' ) ( 'L' | 'l' );
 SMALLINT: ( 'S' | 's' ) ( 'M' | 'm' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' );
 INTEGER: ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'G' | 'g' ) ( 'E' | 'e' ) ( 'R' | 'r' );
+UINTEGER: ( 'U' | 'u' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'G' | 'g' ) ( 'E' | 'e' ) ( 'R' | 'r' );
 BIGINT: ( 'B' | 'b' ) ( 'I' | 'i' ) ( 'G' | 'g' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'T' | 't' );
 FLOAT: ( 'F' | 'f' ) ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'A' | 'a' ) ( 'T' | 't' );
 REAL: ( 'R' | 'r' ) ( 'E' | 'e' ) ( 'A' | 'a' ) ( 'L' | 'l' );
 DOUBLE: ( 'D' | 'd' ) ( 'O' | 'o' ) ( 'U' | 'u' ) ( 'B' | 'b' ) ( 'L' | 'l' ) ( 'E' | 'e' );
 BOOL: ( 'B' | 'b' ) ( 'O' | 'o' ) ( 'O' | 'o' ) ( 'L' | 'l' );
-STRING_TYPE: ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'G' | 'g' );
+STRINGT: ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'G' | 'g' );
 DATE: ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' );
 DATETIME: ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' );
+LOCALDATETIME: ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'C' | 'c' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'D' | 'd' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'E' | 'e' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' );
 TIME: ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' );
+LOCALTIME: ( 'L' | 'l' ) ( 'O' | 'o' ) ( 'C' | 'c' ) ( 'A' | 'a' ) ( 'L' | 'l' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'M' | 'm' ) ( 'E' | 'e' );
+DURATION: ( 'D' | 'd' ) ( 'U' | 'u' ) ( 'R' | 'r' ) ( 'A' | 'a' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'O' | 'o' ) ( 'N' | 'n' );
+BINARY: ( 'B' | 'b' ) ( 'I' | 'i' ) ( 'N' | 'n' ) ( 'A' | 'a' ) ( 'R' | 'r' ) ( 'Y' | 'y' );
 
 // complex datetypes
+MULTISET: ( 'M' | 'm' ) ( 'U' | 'u' ) ( 'L' | 'l' ) ( 'T' | 't' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'E' | 'e' ) ( 'T' | 't' );
 SET: ( 'S' | 's' ) ( 'E' | 'e' ) ( 'T' | 't' );
 LIST: ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'T' | 't' );
+LISTD: ( 'L' | 'l' ) ( 'I' | 'i' ) ( 'S' | 's' ) ( 'T' | 't' ) ( 'D' | 'd' );
 STRUCT: ( 'S' | 's' ) ( 'T' | 't' ) ( 'R' | 'r' ) ( 'U' | 'u' ) ( 'C' | 'c' ) ( 'T' | 't' );
 // END CASE-INSENSITIVE NAMES
 

diff --git a/yarspg/examples/test-example.yarspg b/yarspg/examples/test-example.yarspg
@@ -149,22 +149,22 @@ S(ABC434)-({"bbb23b"}["bbbd32d": Bool, "value": String OPTIONAL])-(JSk342)
 S(ad234)-({"bbb23b"}[$var_234e_schema, "value": String NULL])-(gr525)
 S(asf32)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": String NULL])-(wef32)
 S(asf323)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String) NULL])-(wef323)
-S(asf3233)-({"bbb23b"}["bbbd32d": Bool UNIQUE @<"dfwe": Integer NULL>, "value": List(String) NULL @<"efe": List(List(String))>])-(wef3234)
+S(asf3233)-({"bbb23b"}["bbbd32d": Bool UNIQUE @<"dfwe": SInteger NULL>, "value": List(String) NULL @<"efe": List(List(String))>])-(wef3234)
 S(asf3245)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String) OPTIONAL])-(wef3245)
 S(asf3245)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String MIN 4) OPTIONAL])-(wef3245)
-S(asf3244)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String) OPTIONAL @<"psod": Struct(Integer) OPTIONAL>])-(wef3245)
+S(asf3244)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String) OPTIONAL @<"psod": Struct(SInteger) OPTIONAL>])-(wef3245)
 S(asf3244)-({"bbb23b"}["bbbd32d": Bool UNIQUE, "value": List(String) OPTIONAL @<"psod": Struct(Integer MAX 3) OPTIONAL>])-(wef3245)
 (b43b)-({"bbbb53b"}["bbbd32d": "false", "value": "null"])-(e)+["foaf:maker": "Łukasz Szeremeta and Dominik Tomaszuk", "<http://xmlns.com/dc/0.5/zyz>": "Test annotation"]
 (b43b)-({"bbbb53b"}["bbbd32d": "false", "value": "null"@<"psod": {"dff": "1", "ewf": "55"}>])-(e)+["foaf:maker": "Łukasz Szeremeta and Dominik Tomaszuk", "<http://xmlns.com/dc/0.5/zyz>": "Test annotation"]
 
 S(f23)-({"label"}["date": Date])->(f321)
-S(g34g2)-({"label2"}["time": Time])->(fwef334)
+S(g34g2)-({"label2"}["time": LocalTime])->(fwef334)
 S(weg12)-({"label3"}["timezp": Time])->(gerg534)
 S(Adg3)-({"label4"}["timezm": Time])->(saf24)
 S(Adg35)-({"label4"}["timezm": Time OPTIONAL])->(saf244)
-S(wf23)-({"label5"}["datetime": DateTime])->(erg32)
-S(wf2343)-({"label5"}["datetime": DateTime @<"info": String>])->(erg3264)
-S(sdaf56)-({"label5"}["datetime": DateTime UNIQUE])->(dsf53)
+S(wf23)-({"label5"}["datetime": LocalDateTime])->(erg32)
+S(wf2343)-({"label5"}["datetime": LocalDateTime @<"info": String>])->(erg3264)
+S(sdaf56)-({"label5"}["datetime": LocalDateTime UNIQUE])->(dsf53)
 S(sas34)-({"label6"}["datetimetzp": DateTime])->(sdag654)
 S(asdg41)-({"label7"}["datetimetzm": DateTime])->(asv94)
 (ABC)-({"label"}["date": "1999-01-01"])->(XYZ)
@@ -180,4 +180,7 @@ S(asdg41)-({"label7"}["datetimetzm": DateTime])->(asv94)
 (ABC8)-({"label7"}["datetimetzmminus": "-0003-01-05T11:15:32+05:00"])->(XYZ8)
 (ABC9)-({"label7"}["datetimetzmminusms": "-0003-01-05T11:15:32.432+05:00"])->(XYZ9)
 (ABC10)-({"label7"}["datetimetzmminusms2": "-0003-01-05T11:15:32.5+05:00"])->(XYZ10)
-(ABC10)-({"label7"}["datetimetzmms3": "2020-01-07T14:15:32.24+08:00"])->(XYZ10)
+(ABC10)-({"label7"}["datetimetzmms3": "2020-01-07T14:15:32.24+08:00"])->(XYZ10)
+S(asdg41)-({"label7"}["test": MyType])->(asv94)
+
+# Add DURATION, BINARY, SMALLINT, BIGINT examples