feature/trig_87

src/main/antlr/TriG.g4

@@ -0,0 +1,321 @@
+// $antlr-format alignTrailingComments true, columnLimit 150, minEmptyLines 1, maxEmptyLinesToKeep 1, reflowComments false, useTab false
+// $antlr-format allowShortRulesOnASingleLine false, allowShortBlocksOnASingleLine true, alignSemicolons hanging, alignColons hanging
+
+grammar TriG;
+
+trigDoc
+    : ( directive | block )* EOF
+    ;
+
+block
+    : triplesOrGraph
+    | wrappedGraph
+    | triples2
+    | Graph_w labelOrSubject wrappedGraph
+    ;
+
+triplesOrGraph
+    : labelOrSubject (wrappedGraph | predicateObjectList '.')
+    ;
+
+triples2
+    : blankNodePropertyList predicateObjectList? '.'
+    | collection predicateObjectList '.'
+    ;
+
+wrappedGraph
+    : '{' triplesBlock? '}'
+    ;
+
+triplesBlock
+    : triples ('.' triplesBlock?)?
+    ;
+
+labelOrSubject
+    : iri
+    | blankNode
+    ;
+
+directive
+    : prefixID
+    | base
+    | sparqlPrefix
+    | sparqlBase
+    ;
+
+prefixID
+    : '@prefix' PNAME_NS IRIREF '.'
+    ;
+
+base
+    : '@base' IRIREF '.'
+    ;
+
+sparqlPrefix
+    : Prefix_w PNAME_NS IRIREF
+    ;
+
+sparqlBase
+    : Base_w IRIREF
+    ;
+
+triples
+    : subject predicateObjectList
+    | blankNodePropertyList predicateObjectList?
+    ;
+
+predicateObjectList
+    : verb objectList (';' (verb objectList)?)*
+    ;
+
+objectList
+    : object (',' object)*
+    ;
+
+verb
+    : predicate
+    | 'a'
+    ;
+
+subject
+    : iri
+    | blank
+    ;
+
+predicate
+    : iri
+    ;
+
+object
+    : iri
+    | blank
+    | blankNodePropertyList
+    | literal
+    ;
+
+literal
+    : rDFLiteral
+    | numericLiteral
+    | BooleanLiteral
+    ;
+
+blank
+    : blankNode
+    | collection
+    ;
+
+blankNodePropertyList
+    : '[' predicateObjectList ']'
+    ;
+
+collection
+    : '(' object* ')'
+    ;
+
+numericLiteral
+    : INTEGER
+    | DECIMAL
+    | DOUBLE
+    ;
+
+rDFLiteral
+    : string LANGTAG
+    | string ('^^' iri)?
+    ;
+
+string
+    : STRING_LITERAL_QUOTE
+    | STRING_LITERAL_SINGLE_QUOTE
+    | STRING_LITERAL_LONG_SINGLE_QUOTE
+    | STRING_LITERAL_LONG_QUOTE
+    ;
+
+iri
+    : prefixedName
+    | IRIREF
+    ;
+
+prefixedName
+    : PNAME_LN
+    | PNAME_NS
+    ;
+
+blankNode
+    : BLANK_NODE_LABEL
+    | ANON
+    ;
+
+WS
+    : (('\u0020' | '\u0009' | '\u000A' | '\u000D' ) )+ -> skip
+    ;
+
+// Terminals
+
+Graph_w options { caseInsensitive=true; }
+    : 'GRAPH'
+    ;
+
+Base_w options { caseInsensitive=true; }
+    : 'BASE'
+    ;
+
+Prefix_w options { caseInsensitive=true; }
+    : 'PREFIX'
+    ;
+
+BooleanLiteral
+    : 'true'
+    | 'false'
+    ;
+
+IRIREF
+    : '<' (PN_CHARS | '.' | ':' | '#' | '@' | '%' | '&' | '$' | '!' | '\'' | '*' | '+' | '/' | '(' | ')' | '-' | ',' | '?' | '~' | UCHAR)* '>'
+    ;
+
+PNAME_NS
+    : PN_PREFIX? ':'
+    ;
+
+PNAME_LN
+    : PNAME_NS PN_LOCAL
+    ;
+
+BLANK_NODE_LABEL
+    : '_:' (PN_CHARS_U | '0' .. '9') ((PN_CHARS | '.')* PN_CHARS)?
+    ;
+
+LANGTAG
+    : '@' ('a'.. 'z' | 'A' .. 'Z')+ ('-' ('a'.. 'z' | 'A' .. 'Z' | '0' .. '9')* )*
+    ;
+
+INTEGER
+    : ('+' | '-' )? ('0' .. '9')+
+    ;
+
+DECIMAL
+    : ('+' | '-' )? ('0' .. '9')* '.' ('0' .. '9')+
+    ;
+
+DOUBLE
+    : ('+' | '-' )? (('0' .. '9')+ '.' ('0' .. '9')* EXPONENT
+    | '.' ('0' .. '9')+ EXPONENT
+    | ('0' .. '9')+ EXPONENT)
+    ;
+
+EXPONENT
+    : ('e' | 'E') ('+' | '-' )? ('0' .. '9')+
+    ;
+
+STRING_LITERAL_QUOTE
+    : '"' ((~[\u0022\u005C\u0010\u0013]) | ECHAR | UCHAR)* '"'
+    ;
+
+STRING_LITERAL_SINGLE_QUOTE
+    : '\'' ((~[\u0027\u005C\u0010\u0013]) | ECHAR | UCHAR)* '\''
+    ;
+
+STRING_LITERAL_LONG_SINGLE_QUOTE
+    : '\'\'\'' (('\'' | '\'\'')? ( (~['\\] ) | ECHAR | UCHAR))* '\'\'\''
+    ;
+
+STRING_LITERAL_LONG_QUOTE
+    : '"""' (('"' | '""')? ( (~["'] ) | ECHAR | UCHAR))* '"""'
+    ;
+
+UCHAR
+    : '\\u' HEX HEX HEX HEX
+    | '\\U' HEX HEX HEX HEX HEX HEX HEX HEX
+    ;
+
+ECHAR options { caseInsensitive=true; }
+    : '\\' [tbnrf"'\\]
+    ;
+
+WHITESPACE
+    : [\u0020\u0009\u000A\u000D]
+    ;
+
+ANON
+    : '[' WHITESPACE* ']'
+    ;
+
+PN_CHARS_BASE
+    : 'A' .. 'Z'
+    | 'a' .. 'z'
+    | '\u00C0' .. '\u00D6'
+    | '\u00D8' .. '\u00F6'
+    | '\u00F8' .. '\u02FF'
+    | '\u0370' .. '\u037D'
+    | '\u037F' .. '\u1FFF'
+    | '\u200C' .. '\u200D'
+    | '\u2070' .. '\u218F'
+    | '\u2C00' .. '\u2FEF'
+    | '\u3001' .. '\uD7FF'
+    | '\uF900' .. '\uFDCF'
+    | '\uFDF0' .. '\uFFFD'
+//    | '\u10000' .. '\uEFFFF'
+    ;
+
+PN_CHARS_U
+    : PN_CHARS_BASE
+    | '_'
+    ;
+
+PN_CHARS
+    : PN_CHARS_U
+    | '-'
+    | [0-9]
+    | [\u00B7]
+    | [\u0300-\u036F]
+    | [\u203F-\u2040]
+    ;
+
+PN_PREFIX
+    : PN_CHARS_BASE ((PN_CHARS | '.')* PN_CHARS)?
+    ;
+
+PN_LOCAL
+    :  	(PN_CHARS_U | ':' | [0-9] | PLX) ((PN_CHARS | '.' | ':' | PLX)* (PN_CHARS | ':' | PLX))?
+    ;
+
+PLX
+    : PERCENT
+    | PN_LOCAL_ESC
+    ;
+
+PERCENT
+    : '%' HEX HEX
+    ;
+
+HEX
+    : [0-9a-fA-F]
+    ;
+
+PN_LOCAL_ESC
+    : '\\' (
+        '_'
+        | '~'
+        | '.'
+        | '-'
+        | '!'
+        | '$'
+        | '&'
+        | '\''
+        | '('
+        | ')'
+        | '*'
+        | '+'
+        | ','
+        | ';'
+        | '='
+        | '/'
+        | '?'
+        | '#'
+        | '@'
+        | '%'
+        )
+    ;
+
+LC
+    : '#' ~[\r\n]+ -> channel(HIDDEN)
+    ;

src/main/java/fr/inria/corese/core/next/impl/common/util/IRIUtils.java

@@ -10,9 +10,10 @@
  */
 public class IRIUtils {
 
-    private static final Pattern IRI_PATTERN = Pattern.compile("^(([\\w\\-]+:\\/\\/([\\w\\-_:]+\\.)*[\\w\\-_:]*)(\\/([\\w\\-\\._\\:]+\\/)*))([\\w\\-\\._\\:]+)?(\\?[\\w\\-_\\:\\?\\=]+)?((\\#)?([\\w\\-_]+))?$");
+    private static final Pattern IRI_PATTERN = Pattern.compile("^(?<namespace>(?<protocol>[\\w\\-]+):(?<dblSlashes>\\/\\/)?(?<domain>([\\w\\-_:@]+\\.)*[\\w\\-_:]*))((?<path>\\/([\\w\\-\\._\\:]+\\/)*)(?<finalPath>[\\w\\-\\._\\:]+)?(?<query>\\?[\\w\\-_\\:\\?\\=]+)?(\\#)?(?<fragment>([\\w\\-_]+))?)?$");
     private static final Pattern STANDARD_IRI_PATTERN = Pattern.compile("^(([^:/?#\\s]+):)(\\/\\/([^/?#\\s]*))?([^?#\\s]*)(\\?([^#\\s]*))?(#(.*))?");
 
+
     /**
      * Prevent instantiation of the utility class.
      */
@@ -29,15 +30,24 @@ public static String guessNamespace(String iri) {
             Matcher matcher = IRI_PATTERN.matcher(iri);
 
             if(matcher.matches()) {
-                if((matcher.group(8) == null) || (matcher.group(6) == null && matcher.group(9) == null) ) { // If the IRI has no fragment or ends with a slash
-
-                    return matcher.group(1);
-                } else {
-                    // 1: Domain and path ending with a slash, 6: final path element without slash, 9: final # if there is a fragment
-                    return matcher.group(1) + matcher.group(6) + matcher.group(9);
+                if(matcher.group("protocol") != null && matcher.group("protocol").equals("_")) {
+                    return "";
+                }
+                StringBuilder namespace = new StringBuilder();
+                namespace.append(matcher.group("protocol")).append(":");
+                if(matcher.group("dblSlashes") != null) {
+                    namespace.append(matcher.group("dblSlashes"));
                 }
+                namespace.append(matcher.group("domain"));
+                if(matcher.group("path") != null) {
+                    namespace.append(matcher.group("path"));
+                }
+                if(matcher.group("fragment") != null && matcher.group("finalPath") != null) {
+                    namespace.append(matcher.group("finalPath")).append("#");
+                }
+                return namespace.toString();
             } else {
-                return "";
+                throw new IllegalStateException("No namespace found for the given IRI: " + iri + ".");
             }
         } catch (IllegalStateException e) {
             return "";
@@ -54,10 +64,10 @@ public static String guessLocalName(String iri) {
             Matcher matcher = IRI_PATTERN.matcher(iri);
 
             if(matcher.matches()) {
-                if(matcher.group(10) != null){ // If the IRI has a fragment
-                    return matcher.group(10);
-                } else if(matcher.group(6) != null ) { // If the IRI has no fragment but do not ends with a slash
-                    return matcher.group(6);
+                if(matcher.group("fragment") != null){ // If the IRI has a fragment
+                    return matcher.group("fragment");
+                } else if(matcher.group("finalPath") != null ) { // If the IRI has no fragment but do not ends with a slash
+                    return matcher.group("finalPath");
                 } else { // If the URI ends with a slash
                     return "";
                 }