On 16/05/16 18:25, Praveen Balaji wrote:
Thanks Andy. Yes, I need to handle warnings because part of my use case is
to exit on malformed subjects, even if that doesn't mean a malformed TTL.
I'll try it out. I was hoping I don't have to use a low-level API :)
Rather than hook into the parser process itself, a different design is
to create processing pipeline.
Parsers send their output to a StreamRDF.
RDFDatMgr.parse(StreamRDF sink, String uri, Lang lang)
A StreamRDF is an interface for a processing pipeline. StreamRDFLib has
helpers for StreamRDF to put data into graphs and datasets.
You could write a StreamRDF that checks the triples going though and
send them onto to another StreamRDF (inherit from StreamRDFWrapper).
Then you can customize the checking (e.g. subjects only, make then
errors, whatever). And it works for N-Triples that usually runs with
minimal checking.
The checking of IRIs is done by org.apache.jena.riot.system.IRIResolver.
One thing - IRI parsing is a noticable cost - it is worth having a
cache. See CheckerIRI.
URIs in code repeat a lot across triples.
Andy
Cheers
Praveen
On Mon, May 16, 2016, 10:08 AM Andy Seaborne <[email protected]> wrote:
On 16/05/16 15:02, Praveen Balaji wrote:
Hi all
I'm trying to read from Turtle file using Jena Riot API. My input had a
malformed subject URI and the application exited without an exception. I
configured log4j and I was able to then see an error message.
Is there a way to configure an error handler on the parser so I am
notified
of errors and can handle them as needed?
Thanks
Is the logging output saying its warning?
A Turtle syntax error causes the parser to throw an exception. A warning
is note that a URI is not good as per the RFCs but it is not a syntax
error so the parser continues.
If you really want warnings passed out - set the error handler via the
low level APIs:
InputStream in = .... ;
StreamRDF dest = StreamRDFLib.graph(model.getGraph()) ;
ReaderRIOT r = RDFDataMgr.createReader(Lang.TTL) ;
r.setErrorHandler(null) ;
r.read(in, null, null, dest, null) ;
// data in the model.
Andy
