Merge pull request #48 from herminiogg/develop

Develop

build.sbt

@@ -2,7 +2,7 @@ organization := "es.weso"
 
 name := "ShExML"
 
-version := "0.2.0"
+version := "0.2.1"
 
 scalaVersion := "2.12.4"
 

films.shexml

@@ -42,7 +42,7 @@ EXPRESSION films <films_xml_file.film_xml UNION films_json_file.film_json>
 :Films :[films.id] {
     schema:name [films.name] ;
     :year dbr:[films.year] ;
-    schema:countryOfOrigin dbr:[films.country] ;
+    schema:2232-22 dbr:[films.country] ;
     schema:director dbr:[films.directors] ;
     :screenwritter dbr:[films.screenwritters] ;
     schema:musicBy dbr:[films.music] ;

src/main/java/es/weso/antlr/ShExMLLexer.g4

@@ -9,6 +9,7 @@ SOURCE: S O U R C E -> mode(DECLARATION_CONTENT) ;
 QUERY: Q U E R Y ;
 ITERATOR: I T E R A T O R -> mode(DECLARATION_CONTENT) ;
 FIELD: F I E L D -> mode(DECLARATION_CONTENT) ;
+AUTOINCREMENT: A U T O I N C R E M E N T -> mode(AUTOINCREMENT_CONTENT);
 EXPRESSION: E X P R E S S I O N ;
 UNION: U N I O N ;
 JOIN: J O I N ;
@@ -30,8 +31,8 @@ COMMA: ',' ;
 QUOTE: '"' ;
 AND: '&' | 'AND' ;
 A: ('a' | 'A');
-STRING_OR_VAR: LETTER (LETTER | DIGIT | '_')* ;
-URI_VAR: (LETTER | DIGIT | '_')* ':' ;
+STRING_OR_VAR: (LETTER | DIGIT | '_' | '-' | '\\.' | '%2E')+ ;
+URI_VAR: (LETTER | DIGIT | '_' | '\\.' | '%2E')* ':' ;
 STRINGOPERATOR: '"' (LETTER | DIGIT | ALLOWED_CHARACTERS | ' ')+ '"' ;
 XMLSCHEMADATATYPE: XMLSCHEMADATATYPEPREFIX ':' ALLOWEDTYPES ;
 SHAPELINK: '@' (URI_VAR STRING_OR_VAR | '<' STRING_OR_VAR '>') ;
@@ -54,6 +55,17 @@ STRING_OR_VAR_QUERY: LETTER (LETTER | DIGIT | '_')* ;
 URI_VAR_QUERY: (LETTER | DIGIT | '_')* ':' ;
 WS_DECLARATION: [ \t\n\r] -> skip ;
 
+mode AUTOINCREMENT_CONTENT ;
+LESS_SYMBOL_AUTOINCREMENT: '<' ;
+TO: T O ;
+BY: B Y ;
+ADD_AUTOINCREMENT: '+' ;
+DIGITS: DIGIT+ ;
+STRING_OR_VAR_AUTOINCREMENT: LETTER (LETTER | DIGIT | '_')* ;
+STRINGOPERATOR_AUTOINCREMENT: '"' (LETTER | DIGIT | ALLOWED_CHARACTERS | ' ')+ '"' ;
+GREATER_SYMBOL_AUTOINCREMENT: '>' -> mode(DEFAULT_MODE) ;
+WS_AUTOINCREMENT: [ \t\n\r] -> skip ;
+
 
 fragment LETTER: [a-zA-Z] | '\u00C0'..'\u00D6' | '\u00D8'..'\u00F6' | '\u00F8'..'\u00FF' ;
 fragment DIGIT: [0-9] ;

src/main/java/es/weso/antlr/ShExMLLexer.tokens

@@ -3,54 +3,63 @@ SOURCE=2
 QUERY=3
 ITERATOR=4
 FIELD=5
-EXPRESSION=6
-UNION=7
-JOIN=8
-MATCHER=9
-AS=10
-MATCHING=11
-LESS_SYMBOL=12
-GREATER_SYMBOL=13
-BRACKET_LEFT=14
-BRACKET_RIGHT=15
-PAREN_LEFT=16
-PAREN_RIGHT=17
-BRACE_LEFT=18
-BRACE_RIGHT=19
-SEMICOLON=20
-DOT=21
-ADD=22
-COMMA=23
-QUOTE=24
-AND=25
-A=26
-STRING_OR_VAR=27
-URI_VAR=28
-STRINGOPERATOR=29
-XMLSCHEMADATATYPE=30
-SHAPELINK=31
-LANGTAG=32
-COMMENT=33
-WS=34
-URL=35
-JSONPATH=36
-XMLPATH=37
-CSVPERROW=38
-QUERY_PART=39
-GREATER_SYMBOL_QUERY=40
-WS_QUERY=41
-LESS_SYMBOL_QUERY=42
-STRING_OR_VAR_QUERY=43
-URI_VAR_QUERY=44
-WS_DECLARATION=45
-'['=14
-']'=15
-'('=16
-')'=17
-'{'=18
-'}'=19
-';'=20
-'.'=21
-'+'=22
-','=23
-'"'=24
+AUTOINCREMENT=6
+EXPRESSION=7
+UNION=8
+JOIN=9
+MATCHER=10
+AS=11
+MATCHING=12
+LESS_SYMBOL=13
+GREATER_SYMBOL=14
+BRACKET_LEFT=15
+BRACKET_RIGHT=16
+PAREN_LEFT=17
+PAREN_RIGHT=18
+BRACE_LEFT=19
+BRACE_RIGHT=20
+SEMICOLON=21
+DOT=22
+ADD=23
+COMMA=24
+QUOTE=25
+AND=26
+A=27
+STRING_OR_VAR=28
+URI_VAR=29
+STRINGOPERATOR=30
+XMLSCHEMADATATYPE=31
+SHAPELINK=32
+LANGTAG=33
+COMMENT=34
+WS=35
+URL=36
+JSONPATH=37
+XMLPATH=38
+CSVPERROW=39
+QUERY_PART=40
+GREATER_SYMBOL_QUERY=41
+WS_QUERY=42
+LESS_SYMBOL_QUERY=43
+STRING_OR_VAR_QUERY=44
+URI_VAR_QUERY=45
+WS_DECLARATION=46
+LESS_SYMBOL_AUTOINCREMENT=47
+TO=48
+BY=49
+ADD_AUTOINCREMENT=50
+DIGITS=51
+STRING_OR_VAR_AUTOINCREMENT=52
+STRINGOPERATOR_AUTOINCREMENT=53
+GREATER_SYMBOL_AUTOINCREMENT=54
+WS_AUTOINCREMENT=55
+'['=15
+']'=16
+'('=17
+')'=18
+'{'=19
+'}'=20
+';'=21
+'.'=22
+','=24
+'"'=25

src/main/java/es/weso/antlr/ShExMLParser.g4

@@ -7,20 +7,23 @@ package es.weso.antlr;
 options{tokenVocab=ShExMLLexer;}
 
 shExML: decl* shape* ;
-decl: (source | prefix | query | expression | matcher | iterator) ;
+decl: (source | prefix | query | expression | matcher | iterator | autoincrement) ;
 prefix: PREFIX variable LESS_SYMBOL_QUERY URL GREATER_SYMBOL_QUERY ;
 source: SOURCE variable LESS_SYMBOL_QUERY URL GREATER_SYMBOL_QUERY ;
 iterator: ITERATOR variable LESS_SYMBOL_QUERY queryClause GREATER_SYMBOL_QUERY '{' field+ nestedIterator* '}' ;
 nestedIterator: ITERATOR variable LESS_SYMBOL_QUERY QUERY_PART GREATER_SYMBOL_QUERY '{' field+ nestedIterator* '}' ;
 field: FIELD variable LESS_SYMBOL_QUERY QUERY_PART GREATER_SYMBOL_QUERY ;
 query: QUERY variable LESS_SYMBOL_QUERY queryClause GREATER_SYMBOL_QUERY ;
+autoincrement: AUTOINCREMENT variable LESS_SYMBOL_AUTOINCREMENT
+    (STRINGOPERATOR_AUTOINCREMENT ADD_AUTOINCREMENT)? DIGITS (TO DIGITS)? (BY DIGITS)?
+    (ADD_AUTOINCREMENT STRINGOPERATOR_AUTOINCREMENT)? GREATER_SYMBOL_AUTOINCREMENT ;
 expression: EXPRESSION variable LESS_SYMBOL exp GREATER_SYMBOL ;
 matcher: MATCHER variable LESS_SYMBOL matchers GREATER_SYMBOL ;
 matchers: replacedStrings AS (STRING_OR_VAR | STRINGOPERATOR)
     | matchers AND replacedStrings AS (STRING_OR_VAR | STRINGOPERATOR) ;
 replacedStrings: (STRING_OR_VAR | STRINGOPERATOR) ',' replacedStrings | (STRING_OR_VAR | STRINGOPERATOR) ;
 exp: union | join | stringOperation | iteratorQuery ;
-stringOperation: iteratorQuery '+' STRINGOPERATOR '+' iteratorQuery ;
+stringOperation: iteratorQuery ADD STRINGOPERATOR ADD iteratorQuery ;
 iteratorQuery: variable '.' composedVariable ;
 composedVariable: variable | variable '.' composedVariable ;
 queryClause: JSONPATH QUERY_PART | XMLPATH QUERY_PART | CSVPERROW ;
@@ -30,10 +33,10 @@ leftUnionOption: iteratorQuery | stringOperation ;
 rightUnionOption: iteratorQuery | union | stringOperation ;
 shape: tripleElement prefixVar '[' (exp | variable) ']' '{' (predicateObject ';')* predicateObject? '}' ;
 predicateObject: predicate (objectElement | shapeLink | literalValue) ;
-objectElement: prefixVar? '[' (exp | variable) (MATCHING variable)? ']' (XMLSCHEMADATATYPE | LANGTAG)? ;
+objectElement: prefixVar? ('[' (exp | variable) (MATCHING variable)? ']' | STRINGOPERATOR) (XMLSCHEMADATATYPE | LANGTAG)? ;
 shapeLink: SHAPELINK ;
 predicate: (literalValue | A) ;
 literalValue: prefixVar variable ;
 tripleElement: predicate | LESS_SYMBOL variable GREATER_SYMBOL ;
 prefixVar: variable | URL ;
-variable: STRING_OR_VAR | URI_VAR | STRING_OR_VAR_QUERY | URI_VAR_QUERY ;
+variable: STRING_OR_VAR | URI_VAR | STRING_OR_VAR_QUERY | URI_VAR_QUERY | STRING_OR_VAR_AUTOINCREMENT ;