Closed MichaelRoeder closed 7 years ago
Added the complete platform_controller_log.txt
Notes for easier reproducing:
Add the credentials of our admin to the compose file as described in https://github.com/hobbit-project/platform#using-benchmarks-from-the-hobbit-git
We know that this issue has been reported for rkaterinenko
. Create a user with this name and give it the role challenge-organiser
as described at https://github.com/hobbit-project/platform#details-of-the-user-management
For the experiment the DEBS GC challenge benchmark can be used.
the problem was the fact that challenge model doesn't include graph, following snippet fixes it:
private Model createExpModelForChallengeTask(Model model, String challengeTaskUri, String systemUri) {
Dataset dataset = DatasetFactory.create();
dataset.addNamedModel("http://temp.org/challenge", model);
String query = SparqlQueries.getCreateExperimentFromTaskQuery(Constants.NEW_EXPERIMENT_URI, challengeTaskUri,
systemUri, "http://temp.org/challenge");
if (query == null) {
LOGGER.error("Couldn't load SPARQL query to create an RDF model for a new experiment. Returning null.");
return null;
}
QueryExecution qe = QueryExecutionFactory.create(query, dataset);
return qe.execConstruct();
}
Problem
When a "normal" experiment is started the benchmark controller of an example benchmark is created in the following way
When creating a challenge task with the same (default) configuration as in the "normal" experiment the RDF model is not complete