Requirements:
- Java 7+
- Maven
License:
- The code is released under the Apache 2.0 license
The project has two main files:
- (eu.planetdata.srbench.oracle.repository.)SRBenchImporter : imports the RDF input stream in the oracle
- (eu.planetdata.srbench.oracle.)Oracle: executes the test queries over the RDF stream, produces the results. If a results to be compared is provided, the Oracle verifies that it is correct.
Configuration:
- The project can be configured editing the setup.properties file. See https://github.com/dellaglio/csrbench-oracle/wiki/Configuration for a description of the properties.
Import the data stream into the oracle:
- It is done through the SRBenchImporter
- Data should be described in Turtle and it should be stored in the folder data. Filenames should follow the pattern data_xx.ttl, where xx is the application timestamp associated to the triples in the file.
Execute the Oracle:
- Run the Oracle. Depending on the setup configuration, for each query to be executed, two possible behaviours are possible:
- If there is no answer associated to the query, the oracle produces all the possible answers according to the input configuration;
- If there is an answer associated to the query, the oracle starts to produce an answer and comparing it with the input one. If it finds a match, it stops, otherwise it repeats.
- When the Oracle runs, it creates an html file with a summary of the tests and the results.
Publications:
- Daniele Dell'Aglio, Jean-Paul Calbimonte, Marco Balduini, Óscar Corcho, Emanuele Della Valle: On Correctness in RDF Stream Processor Benchmarking. International Semantic Web Conference (2) 2013: 326-342