Pregunta

I am trying to send the following query to DBpedia using dotNetRDF:

PREFIX rdf:  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX xsd:  <http://www.w3.org/2001/XMLSchema#>
PREFIX temp: <http://www.example.com/temp/>

CONSTRUCT {
  ?v1 a temp:r.
  ?v1 temp:p0 ?v2.
}
WHERE {
  OPTIONAL {
    {
      {
        SELECT ?v1
        WHERE {
          ?v1 rdfs:comment ?v2.
        }
        LIMIT 10
      }.
      ?v1 rdfs:comment ?v2.
    }.
  }.
}

However, I am getting an RdfParseException saying that an unexpected r character was found (114 being the decimal character code of r).

When I use the web-based DBpedia SPARQL form for the same query, I get back what looks like reasonable results. Therefore, I think there is no problem with my SPARQL query, or at least none that DBpedia could not cope with.

I have tried to follow the advice from this question about a similar problem by enabling Options.HttpDebugging and Options.HttpFullDebugging. As described in the accepted answer to that question, this does print the received response to stderr. It does not, however, evoke "a different error"; instead the query seems to be successful, however the returned graph is empty (it contains no triples).

Here is a minimal sample that reproduces the issue:

using System;
using System.Runtime.ConstrainedExecution;

using VDS.RDF;
using VDS.RDF.Query;
using VDS.RDF.Writing;

namespace DotNetRdfParsingError
{
    class Program
    {
        public static void Main(string[] args)
        {
            var endpoint = new SparqlRemoteEndpoint(new Uri("http://dbpedia.org/sparql"));

            string query =
                "PREFIX rdf:  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>\n" +
                "PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>\n" +
                "PREFIX xsd:  <http://www.w3.org/2001/XMLSchema#>\n" +
                "PREFIX temp: <http://www.example.com/temp/>\n" +
                "\n" +
                "CONSTRUCT {\n" +
                "  ?v1 a temp:r.\n" +
                "  ?v1 temp:p0 ?v2.\n" +
                "}\n" +
                "WHERE {\n" +
                "  OPTIONAL {\n" +
                "    {\n" +
                "      {\n" +
                "        SELECT ?v1\n" +
                "        WHERE {\n" +
                "          ?v1 rdfs:comment ?v2.\n" +
                "        }\n" +
                "        LIMIT 10\n" +
                "      }.\n" +
                "      ?v1 rdfs:comment ?v2.\n" +
                "    }.\n" +
                "  }.\n" +
                "}";

            Console.WriteLine("Without HTTP debugging:");
            try {
                var graph = endpoint.QueryWithResultGraph(query);
                PrintGraph(graph);
            }
            catch (Exception ex) {
                Console.WriteLine("EXCEPTION:\n" + ex.ToString());
            }

            Console.WriteLine("With HTTP debugging:");
            try {
                Options.HttpDebugging = true;
                Options.HttpFullDebugging = true;
                var graph = endpoint.QueryWithResultGraph(query);
                PrintGraph(graph);
            }
            catch (Exception ex) {
                Console.WriteLine("EXCEPTION:\n" + ex.ToString());
            }
        }

        private static void PrintGraph(IGraph graph)
        {
            if (graph != null) {
                using (var stringWriter = new System.IO.StringWriter()) {
                    var turtleWriter = new TurtleWriter();
                    turtleWriter.PrettyPrintMode = true;
                    turtleWriter.Save(graph, stringWriter);
                    Console.WriteLine("Graph:\n\n" + stringWriter.ToString() + "\n");
                }
            } else {
                Console.WriteLine("Graph is null.");
            }
        }
    }
}

This produces the following output:

Without HTTP debugging:
EXCEPTION:
VDS.RDF.Parsing.RdfParseException: [Line 1 Column 9] Unexpected Character (Code 114) r was encountered
   bei VDS.RDF.Parsing.Tokens.NTriplesTokeniser.GetNextToken()
   bei VDS.RDF.Parsing.Tokens.BufferedTokenQueue.BufferInternal()
   bei VDS.RDF.Parsing.Tokens.BufferedTokenQueue.InitialiseBuffer(Int32 amount)
   bei VDS.RDF.Parsing.NTriplesParser.Parse(TokenisingParserContext context)
   bei VDS.RDF.Parsing.NTriplesParser.Load(IRdfHandler handler, TextReader input)
   bei VDS.RDF.Parsing.NTriplesParser.Load(IRdfHandler handler, StreamReader input)
   bei VDS.RDF.Query.SparqlRemoteEndpoint.QueryWithResultGraph(IRdfHandler handler, String sparqlQuery)
   bei VDS.RDF.Query.SparqlRemoteEndpoint.QueryWithResultGraph(String sparqlQuery)
   bei DotNetRdfParsingError.Program.Main(String[] args) in c:\Users\uname\Documents\Test\DOTNET\DotNetRdfParsingError\DotNetRdfParsingError\Program.cs:Zeile 43.
With HTTP debugging:
# HTTP DEBUGGING #
HTTP Request to http://dbpedia.org/sparql?query=PREFIX%20rdf%3A%20%20%3Chttp%3A%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%23%3E%0APREFIX%20rdfs%3A%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E%0APREFIX%20xsd%3A%20%20%3Chttp%3A%2F%2Fwww.w3.org%2F2001%2FXMLSchema%23%3E%0APREFIX%20temp%3A%20%3Chttp%3A%2F%2Fwww.example.com%2Ftemp%2F%3E%0A%0ACONSTRUCT%20%7B%0A%20%20%3Fv1%20a%20temp%3Ar.%0A%20%20%3Fv1%20temp%3Ap0%20%3Fv2.%0A%7D%0AWHERE%20%7B%0A%20%20OPTIONAL%20%7B%0A%20%20%20%20%7B%0A%20%20%20%20%20%20%7B%0A%20%20%20%20%20%20%20%20SELECT%20%3Fv1%0A%20%20%20%20%20%20%20%20WHERE%20%7B%0A%20%20%20%20%20%20%20%20%20%20%3Fv1%20rdfs%3Acomment%20%3Fv2.%0A%20%20%20%20%20%20%20%20%7D%0A%20%20%20%20%20%20%20%20LIMIT%2010%0A%20%20%20%20%20%20%7D.%0A%20%20%20%20%20%20%3Fv1%20rdfs%3Acomment%20%3Fv2.%0A%20%20%20%20%7D.%0A%20%20%7D.%0A%7D

GET
Accept:text/plain,text/ntriples,text/ntriples+turtle,application/rdf-triples,application/x-ntriples,text/plain,text/ntriples,text/ntriples+turtle,application/rdf-triples,application/x-ntriples,text/turtle,application/x-turtle,application/turtle,text/turtle,application/x-turtle,application/turtle,text/n3,text/rdf+n3,text/n3,text/rdf+n3,application/rdf+xml,text/xml,application/xml,application/rdf+xml,text/xml,application/xml,application/json,text/json,application/rdf+json,application/json,text/json,application/rdf+json,text/html,application/xhtml+xml,text/html,application/xhtml+xml,*/*;q=0.5


HTTP Response from http://dbpedia.org/sparql?query=PREFIX%20rdf%3A%20%20%3Chttp%3A%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%23%3E%0APREFIX%20rdfs%3A%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E%0APREFIX%20xsd%3A%20%20%3Chttp%3A%2F%2Fwww.w3.org%2F2001%2FXMLSchema%23%3E%0APREFIX%20temp%3A%20%3Chttp%3A%2F%2Fwww.example.com%2Ftemp%2F%3E%0A%0ACONSTRUCT%20%7B%0A%20%20%3Fv1%20a%20temp%3Ar.%0A%20%20%3Fv1%20temp%3Ap0%20%3Fv2.%0A%7D%0AWHERE%20%7B%0A%20%20OPTIONAL%20%7B%0A%20%20%20%20%7B%0A%20%20%20%20%20%20%7B%0A%20%20%20%20%20%20%20%20SELECT%20%3Fv1%0A%20%20%20%20%20%20%20%20WHERE%20%7B%0A%20%20%20%20%20%20%20%20%20%20%3Fv1%20rdfs%3Acomment%20%3Fv2.%0A%20%20%20%20%20%20%20%20%7D%0A%20%20%20%20%20%20%20%20LIMIT%2010%0A%20%20%20%20%20%20%7D.%0A%20%20%20%20%20%20%3Fv1%20rdfs%3Acomment%20%3Fv2.%0A%20%20%20%20%7D.%0A%20%20%7D.%0A%7D
HTTP/1.1 200 OK

Connection:keep-alive
Vary:Accept-Encoding
Accept-Ranges:bytes
Content-Length:5451
Content-Type:text/plain
Date:Fri, 18 Oct 2013 08:06:33 GMT
Server:Virtuoso/07.00.3204 (Linux) i686-generic-linux-glibc212-64  VDB

@prefix rdf:    <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix ns1:    <http://www.example.com/temp/> .
<http://dbpedia.org/resource/American_Legion_Field_(Florence)>  rdf:type    ns1:r ;
    ns1:p0  "American Legion Field is a baseball venue in Florence, South Carolina, USA. It is home to the Florence Red Wolves of the Coastal Plain League, a collegiate summer baseball league. The Red Wolves have played at the field since 1998. The venue was built sometime before 1981 and has a capacity of 3,500. The field's dimensions are 305 ft. down the foul lines, 335 ft. to the gaps, and 385 ft. to dead center field."@en .
@prefix dbpedia:    <http://dbpedia.org/resource/> .
dbpedia:Francisco_Bautista  rdf:type    ns1:r ;
    ns1:p0  "Francisco Bautista Cuamatzi is a male long-distance runner from Mexico. He represented his native country at the 2008 Summer Olympics in Beijing, PR China, where he finished in 66th place in the men's marathon event, clocking 2:29.28. Rojas set his personal best in the marathon on March 7, 2004 in Torre\u00F3n."@en ,
        "Francisco Bautista ist ein mexikanischer Marathonl\u00E4ufer. Bei den Halbmarathon-Weltmeisterschaften 1999 in Palermo kam er auf den 87. Platz. Im Jahr darauf wurde er F\u00FCnfter beim Marat\u00F3n de la Comarca Lagunera und Achter beim Mailand-Marathon. 2001 wurde er Sechster in der Comarca Lagunera, Zentralamerika/Karibik-Vizemeister im Halbmarathon, belegte bei den Leichtathletik-Weltmeisterschaften in Edmonton Rang 66 und gewann den Monterrey-Marathon."@de .
<http://dbpedia.org/resource/Tillandsia_\u0027Comet\u0027>  rdf:type    ns1:r ;
    ns1:p0  "'Comet' is a hybrid cultivar of the genus Tillandsia in the Bromeliad family."@en .
dbpedia:Glass_Sky   rdf:type    ns1:r ;
    ns1:p0  "Glass Sky is a Japanese manga anthology written and illustrated by Yugi Yamada. It is licensed in North America by Digital Manga Publishing, which released the manga through its June imprint, on December 12, 2007."@en .
dbpedia:Healthcare_in_Guntur    rdf:type    ns1:r ;
    ns1:p0  "The region of Guntur has many health care facilities supported by both the government and private institutions. It is one of the major cities on the east coast region of India in providing excellent medical and health care facilities."@en .
dbpedia:Lake_Nuijamaa   rdf:type    ns1:r ;
    ns1:p0  "Lake Nuijamaa is a lake on the border between Finland and Russia next to the town of Nuijamaa. It is part of the Saimaa Canal linking Vyborg Bay in the Baltic Sea to Lake Saimaa in the Finnish Lakeland."@en ,
        "\u041D\u0443\u0439\u044F\u043C\u0430\u044F\u0440\u0432\u0438\u00A0\u2014 \u043E\u0437\u0435\u0440\u043E \u043D\u0430 \u041A\u0430\u0440\u0435\u043B\u044C\u0441\u043A\u043E\u043C \u043F\u0435\u0440\u0435\u0448\u0435\u0439\u043A\u0435 \u043D\u0430 \u0433\u0440\u0430\u043D\u0438\u0446\u0435 \u0424\u0438\u043D\u043B\u044F\u043D\u0434\u0438\u0438 \u0438 \u0420\u043E\u0441\u0441\u0438\u0438. \u041E\u0437\u0435\u0440\u043E \u0440\u0430\u0441\u043F\u043E\u043B\u043E\u0436\u0435\u043D\u043E \u0432 20 \u043A\u043C \u043D\u0430 \u044E\u0433\u043E-\u0432\u043E\u0441\u0442\u043E\u043A \u043E\u0442 \u0444\u0438\u043D\u0441\u043A\u043E\u0433\u043E \u0433\u043E\u0440\u043E\u0434\u0430 \u041B\u0430\u043F\u043F\u0435\u044D\u043D\u0440\u0430\u043D\u0442\u0430 \u0438 \u0432 26\u00A0\u043A\u043C \u043D\u0430 \u0441\u0435\u0432\u0435\u0440\u043E-\u0437\u0430\u043F\u0430\u0434 \u043E\u0442 \u0440\u043E\u0441\u0441\u0438\u0439\u0441\u043A\u043E\u0433\u043E \u0412\u044B\u0431\u043E\u0440\u0433\u0430. \u0427\u0435\u0440\u0435\u0437 \u043D\u0435\u0433\u043E \u043F\u0440\u043E\u0445\u043E\u0434\u0438\u0442 \u0421\u0430\u0439\u043C\u0435\u043D\u0441\u043A\u0438\u0439 \u043A\u0430\u043D\u0430\u043B."@ru .
dbpedia:Michelangelo_Unterberger    rdf:type    ns1:r ;
    ns1:p0  "Michelangelo Unterberger (August 11, 1695 \u2013 June 27, 1758), was an Austrian painter. Unterberger was born at Cavalese, in what was then Austrian Tyrol, and was part of an Austrian family of artists. He was originally taught by Giuseppe Alberti, and painted primarily religious-themed works."@en ,
        "Michelangelo Unterberger, auch Michael Angelo Unterberger und Michelangelo Unterperger war ein S\u00FCdtiroler Maler des Barock."@de ,
        "Michelangelo Unterberger, aussi Michael Angelo Unterberger ou Michelangelo Unterperger est un peintre autrichien, originaire de la province autonome de Bolzano, et de l'\u00E9poque baroque."@fr .
<http://dbpedia.org/resource/Miroslav_Ivani%C5%A1evi%C4%87> rdf:type    ns1:r ;
    ns1:p0  "Miroslav Ivani\u0161evi\u0107 is a Montenegrin politician. He was appointed as the Montenegrin Minister of Finance in 1998. In 2007 he was accused of being involved in cigarette smuggling into Italy, but was found not guilty in 2010."@en .
dbpedia:Skin_for_Skin   rdf:type    ns1:r ;
    ns1:p0  "Skin for Skin \u00E8 per ora l'ultimo album pubblicato dalla heavy/christian metal band Bride, nel 2006."@it ,
        "Skin for Skin \u00E9 um \u00E1lbum da banda de heavy metal Bride, lan\u00E7ado em 2006 pela gravadora Retrospective Records. A banda sai de seu estilo alternativo e lan\u00E7a um [disco extremamente pesado e totalmente heavy metal.]"@pt ,
        "Skin for Skinis the twelfth studio album by the rock band Bride."@en .
<http://dbpedia.org/resource/Union_Church_(Buckfield,_Maine)>   rdf:type    ns1:r ;
    ns1:p0  "Union Church is a historic church off ME 140 in Buckfield, Maine. It was built in 1832 and added to the National Register of Historic Places in 1980."@en .

# END HTTP DEBUGGING #
Graph:

@base <http://dbpedia.org/sparql>.

@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>.
@prefix xsd: <http://www.w3.org/2001/XMLSchema#>.

As can be seen there, the resulting graph only consists of some prefixes, but does not contain any nodes or triples.

The exception points to line 1, column 9, of the resulting graph, which - when looking at the received response, would be the beginning of the rdf prefix of the prefix declaration

@prefix rdf:    <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

One thing I find peculiar is that the exception is thrown by the NTriplesTokeniser somewhere inside the NTriplesParser, yet the returned graph looks to me like Turtle rather than N-Triples.

I am using the latest dotNetRDF version from NuGet, currently 1.0.1.2809.

Am I doing anything wrong there, or are there any additional settings for dotNetRDF required in this case?

EDIT: For the sake of completeness, I should add that on a local Virtuoso installation with some DBpedia data inside, there is no crash.

¿Fue útil?

Solución

Yes, it looks like DBPedia is doing the wrong thing and for some reason returning Turtle with the Content-Type of text/plain rather than the appropriate value for Turtle which would be application/turtle. Hence why dotNetRDF selects the NTriples parser and then runs into issues.

It would be worth reporting this to the DBPedia folks to see if they can fix this issue

dotNetRDF could improve how we form our Accept header to put text/plain lower down the list but this still wouldn't solve the fact that DBPedia appears to be doing content negotiation incorrectly.

You can use the RdfAcceptHeader property to work around this which as the documentation says:

Can be used to workaround buggy endpoints which don't like the broad Accept Header that dotNetRDF sends by default.

For example:

endpoint.RdfAcceptHeader = "application/turtle";

With this workaround in place the code given does not throw an exception.

Aside on HTTP Debugging

A final aside on the point about HTTP debugging evoking a different error, Options.HttpFullDebugging consumes the response stream so the parser receives a stream that is already at the end.

Depending on the parser this could cause a different parsing error, but in the case of NTriples which is a textual format an empty stream is a valid empty graph serialisation. The graph happens to have some prefixes because Graph instances in dotNetRDF always have the prefixes rdf, rdfs and xsd defined by default.

Licenciado bajo: CC-BY-SA con atribución
No afiliado a StackOverflow
scroll top