rdflib 2.3.0 → 2.3.1-0007ffed

package/src/n3parser.js

@@ -212,10 +212,20 @@ var signed_integer = new RegExp("^[-+]?[0-9]+", 'g');
 var number_syntax = new RegExp("^([-+]?[0-9]+)(\\.[0-9]+)?([eE][-+]?[0-9]+)?", 'g');
 var datetime_syntax = new RegExp('^[0-9][0-9][0-9][0-9]-[0-9][0-9]-[0-9][0-9](T[0-9][0-9]:[0-9][0-9](:[0-9][0-9](\\.[0-9]*)?)?)?Z?');
 
+// Reused in tight loops to detect whitespace or comment after a dot
+var wsOrHash = new RegExp("[\\s#]");
+
 var digitstring = new RegExp("^[0-9]+", 'g');
 var interesting = new RegExp("[\\\\\\r\\n\\\"]", 'g');
 var langcode = new RegExp("^[a-zA-Z0-9]+(-[a-zA-Z0-9]+)*", 'g');
 
+// Returns true when a dot at position i should terminate a name,
+// i.e., when the next character is whitespace, a comment start, or EOF
+function dotTerminatesName(str, i) {
+    var next = str.charAt(i + 1);
+    return next === '' || wsOrHash.test(next);
+}
+
 function createSinkParser(store, openFormula, thisDoc, baseURI, genPrefix, metaURI, flags, why) {
     return new SinkParser(store, openFormula, thisDoc, baseURI, genPrefix, metaURI, flags, why);
 }
@@ -554,7 +564,7 @@ export class SinkParser {
         <- prop -<
         _operator_*/
 
-        var j = this.skipSpace(str, i);
+    var j = this.skipSpace(str, i);
         if ((j < 0)) {
             return j;
         }
@@ -703,22 +713,17 @@ export class SinkParser {
                 if ((j >= 0)) {
                     var subj = objs[0];
                     if ((pyjslib_len(objs) > 1)) {
-
                         var __obj = new pyjslib_Iterator(objs);
                         try {
                             while (true) {
                                 var obj = __obj.next();
-
-
                                 this.makeStatement(new pyjslib_Tuple([this._context, this._store.sym(DAML_sameAs_URI), subj, obj]));
-
                             }
                         } catch (e) {
                             if (e != StopIteration) {
                                 throw e;
                             }
                         }
-
                     }
                     var j = this.skipSpace(str, j);
                     if ((j < 0)) {
@@ -743,6 +748,16 @@ export class SinkParser {
             if ((j < 0)) {
                 throw BadSyntax(this._thisDoc, this.lines, str, i, "EOF when ']' expected after [ <propertyList>");
             }
+            if ((str.slice( j,  ( j + 1 ) ) == ".")) {
+                // If a dot is found after a blank node, treat it as a statement terminator.
+                // Do NOT consume the '.' here: statement terminators are handled centrally by
+                // checkDot() (called by directiveOrStatement after statement()). Consuming the dot
+                // locally would bypass that unified logic and could cause inconsistencies.
+                // We do consume ']' below because it is a structural closer of the blank node,
+                // not a statement terminator.
+                res.push(subj);
+                return j; // leave '.' for checkDot()
+            }
             if ((str.slice( j,  ( j + 1 ) ) != "]")) {
                 throw BadSyntax(this._thisDoc, this.lines, str, j, "']' expected");
             }
@@ -1208,7 +1223,17 @@ export class SinkParser {
             return -1;
         }
         var i = j;
-        while ((i < pyjslib_len(str)) && (_notNameChars.indexOf(str.charAt(i)) < 0)) {
+        while ((i < pyjslib_len(str))) {
+            var c = str.charAt(i);
+            if (c === '.') {
+                if (dotTerminatesName(str, i)) {
+                    break; // treat as statement terminator, not part of name
+                }
+                // else: accept '.' as part of name
+            } else if (_notNameChars.indexOf(c) >= 0) {
+                // Other invalid characters terminate the name
+                break;
+            }
             var i =  ( i + 1 ) ;
         }
         res.push(str.slice( j, i));
@@ -1235,13 +1260,15 @@ export class SinkParser {
             var i =  ( i + 1 ) ;
             while ((i < pyjslib_len(str))) {
                 var c = str.charAt(i);
-                if ((_notNameChars.indexOf(c) < 0)) {
-                    var ln =  ( ln + c ) ;
-                    var i =  ( i + 1 ) ;
-                }
-                else {
+                if (c === '.') {
+                    if (dotTerminatesName(str, i)) {
+                        break; // dot ends the name here
+                    }
+                } else if (_notNameChars.indexOf(c) >= 0) {
                     break;
                 }
+                var ln =  ( ln + c ) ;
+                var i =  ( i + 1 ) ;
             }
         }
         else {
@@ -1253,13 +1280,15 @@ export class SinkParser {
             var ln = "";
             while ((i < pyjslib_len(str))) {
                 var c = str.charAt(i);
-                if ((_notNameChars.indexOf(c) < 0)) {
-                    var ln =  ( ln + c ) ;
-                    var i =  ( i + 1 ) ;
-                }
-                else {
+                if (c === '.') {
+                    if (dotTerminatesName(str, i)) {
+                        break; // dot ends the name here
+                    }
+                } else if (_notNameChars.indexOf(c) >= 0) {
                     break;
                 }
+                var ln =  ( ln + c ) ;
+                var i =  ( i + 1 ) ;
             }
             res.push(new pyjslib_Tuple([pfx, ln]));
             return i;

package/src/serialize.ts

@@ -72,7 +72,8 @@ export default function serialize (
         return executeCallback(null, documentString)
       case TurtleContentType:
       case TurtleLegacyContentType:
-        sz.setFlags('si') // Suppress = for sameAs and => for implies
+        // Suppress = for sameAs and => for implies; preserve any user-specified flags (e.g., 'o')
+        sz.setFlags('si' + (opts.flags ? (' ' + opts.flags) : ''))
         documentString = sz.statementsToN3(newSts)
         return executeCallback(null, documentString)
       case NTriplesContentType:
@@ -80,7 +81,8 @@ export default function serialize (
         documentString = sz.statementsToNTriples(newSts)
         return executeCallback(null, documentString)
       case JSONLDContentType:
-        sz.setFlags('si dr') // turtle + dr (means no default, no relative prefix)
+        // turtle + dr (means no default, no relative prefix); preserve user flags
+        sz.setFlags('si dr' + (opts.flags ? (' ' + opts.flags) : ''))
         documentString = sz.statementsToJsonld(newSts) // convert via turtle
         return executeCallback(null, documentString)
       case NQuadsContentType:

package/src/serializer.js

@@ -5,14 +5,12 @@
 ** This is or was https://github.com/linkeddata/rdflib.js/blob/main/src/serializer.js
 ** Licence: MIT
 */
-import NamedNode from './named-node'
-import BlankNode from './blank-node'
+import * as ttl2jsonld from '@frogcat/ttl2jsonld'
+import solidNs from 'solid-namespace'
+import CanonicalDataFactory from './factories/canonical-data-factory'
 import * as Uri from './uri'
 import * as Util from './utils-js'
-import CanonicalDataFactory from './factories/canonical-data-factory'
 import { createXSD } from './xsd'
-import solidNs from 'solid-namespace'
-import * as ttl2jsonld from '@frogcat/ttl2jsonld'
 
 
 export default function createSerializer(store) {
@@ -52,6 +50,12 @@ export class Serializer {
     return this
   }
 
+  /**
+   * Set serializer behavior flags. Letters can be combined with spaces.
+   * Examples: 'si', 'deinprstux', 'si dr', 'o'.
+   * Notable flags:
+   *  - 'o': do not abbreviate to a prefixed name when the local part contains a dot
+   */
   setFlags(flags) {
     this.flags = flags || '';
     return this
@@ -251,10 +255,33 @@ export class Serializer {
     return this.statementsToN3(f.statements)
   }
 
-  _notQNameChars = '\t\r\n !"#$%&\'()*.,+/;<=>?@[\\]^`{|}~'
+  _notQNameChars = '\t\r\n !"#$%&\'()*,+/;<=>?@[\\]^`{|}~' // issue#228
   _notNameChars =
     (this._notQNameChars + ':')
 
+  // Validate if a string is a valid PN_LOCAL per Turtle 1.1 spec
+  // Allows dots inside the local name but not as trailing character
+  // Also allows empty local names (for URIs ending in / or #)
+  isValidPNLocal(local) {
+    // Empty local name is valid (e.g., ex: for http://example.com/)
+    if (local.length === 0) return true
+    
+    // Cannot end with a dot
+    if (local[local.length - 1] === '.') return false
+    
+    // Check each character (allow dots mid-string)
+    for (var i = 0; i < local.length; i++) {
+      var ch = local[i]
+      // Dot is allowed unless it's the last character (checked above)
+      if (ch === '.') continue
+      // Other characters must not be in the blacklist
+      if (this._notNameChars.indexOf(ch) >= 0) {
+        return false
+      }
+    }
+    return true
+  }
+
   explicitURI(uri) {
     if (this.flags.indexOf('r') < 0 && this.base) {
       uri = Uri.refTo(this.base, uri)
@@ -628,13 +655,17 @@ export class Serializer {
     if (j >= 0 && this.flags.indexOf('p') < 0 &&
       // Can split at namespace but only if http[s]: URI or file: or ws[s] (why not others?)
       (uri.indexOf('http') === 0 || uri.indexOf('ws') === 0 || uri.indexOf('file') === 0)) {
-      var canSplit = true
-      for (var k = j + 1; k < uri.length; k++) {
-        if (this._notNameChars.indexOf(uri[k]) >= 0) {
-          canSplit = false
-          break
-        }
-      }
+      var localid = uri.slice(j + 1)
+      var namesp = uri.slice(0, j + 1)
+      // Don't split if namespace is just the protocol (e.g., https://)
+      // A valid namespace should have content after the protocol
+  var minNamespaceLength = uri.indexOf('://') + 4 // e.g., "http://x" minimum
+  // Also don't split if namespace is the base directory (would serialize as relative URI)
+  var baseDir = this.base ? this.base.slice(0, Math.max(this.base.lastIndexOf('/'), this.base.lastIndexOf('#')) + 1) : null
+  var namespaceIsBaseDir = baseDir && namesp === baseDir
+  // If flag 'o' is present, forbid dots in local part when abbreviating
+  var forbidDotLocal = this.flags.indexOf('o') >= 0 && localid.indexOf('.') >= 0
+  var canSplit = !namespaceIsBaseDir && !forbidDotLocal && namesp.length > minNamespaceLength && this.isValidPNLocal(localid)
 /*
       if (uri.slice(0, j + 1) === this.base + '#') { // base-relative
         if (canSplit) {
@@ -645,8 +676,6 @@ export class Serializer {
       }
 */
       if (canSplit) {
-        var localid = uri.slice(j + 1)
-        var namesp = uri.slice(0, j + 1)
         if (this.defaultNamespace && this.defaultNamespace === namesp &&
             this.flags.indexOf('d') < 0) { // d -> suppress default
           if (this.flags.indexOf('k') >= 0 &&
@@ -883,10 +912,10 @@ export class Serializer {
             break
           case 'Literal':
             results = results.concat(['<' + t +
-            (st.object.datatype.equals(this.xsd.string)
-              ? ''
-              : ' rdf:datatype="' + escapeForXML(st.object.datatype.uri) + '"') +
-            (st.object.language ? ' xml:lang="' + st.object.language + '"' : '') +
+            (st.object.language ? ' xml:lang="' + st.object.language + '"' :
+              (st.object.datatype.equals(this.xsd.string)
+                ? ''
+                : ' rdf:datatype="' + escapeForXML(st.object.datatype.uri) + '"')) +
             '>' + escapeForXML(st.object.value) +
             '</' + t + '>'])
             break
@@ -949,9 +978,10 @@ export class Serializer {
             break
           case 'Literal':
             results = results.concat(['<' + qname(st.predicate) +
-              (st.object.datatype.equals(this.xsd.string) ? '' : ' rdf:datatype="' + escapeForXML(st.object.datatype.value) + '"') +
-              (st.object.language ? ' xml:lang="' + st.object.language + '"' : '') +
-              '>' + escapeForXML(st.object.value) +
+              (st.object.language ? 
+                ' xml:lang="' + st.object.language + '"' :
+                (st.object.datatype.equals(this.xsd.string) ? '' : ' rdf:datatype="' + escapeForXML(st.object.datatype.value) + '"')) 
+              + '>' + escapeForXML(st.object.value) +
               '</' + qname(st.predicate) + '>'])
             break
           case 'Collection':