-
Notifications
You must be signed in to change notification settings - Fork 0
REBench cli in Action
Manzoor Ali edited this page May 9, 2022
·
4 revisions
To run the REBench from CLI, follow the following step-by-step guide:
-
Download the REBench-Cli from here
-
Extract the zip folder which contain REBench-cli.jar and a personalized query.
-
User can make changes in the personalized query according to the system requirements. The default personalized query as follows:
PREFIX reld:<https://reld.dice-research.org/schema/>
SELECT DISTINCT ?rId (AVG(?nToken) as ?avgToken) (count(?ne) as ?NE)
WHERE {
?rId reld:hasSentence ?sentence .
?sentence reld:hasSubject ?sub .
?sentence reld:hasObject ?obj .
?sentence reld:numOfTokens ?nToken .
?sentence reld:numOfPunctuations ?numPunc .
?sentence reld:hasNamedEntity ?ne .
}
- Run the following command in your terminal:
java -jar REBench.jar -m feasible-exmp -n 10 -e http://localhost:8890/sparql -q personalized_query.txt -o feasible-exmp-10qa-benchmark.ttl
[Note] User can change the size file names etc. in the above command. For example, in the above command, we are selecting 10 relations.
- After running the above command, you will see an output like the following:
- The system will produce two types of output, relation.txt file and feasible-exmp-10qa-benchmark.ttl file. 6.(a) The relation.txt file only contains the IRIs of the selected relations
relation.txt
<https://reld.dice-research.org/resource/R-3053>
<https://reld.dice-research.org/resource/R-3088>
<https://reld.dice-research.org/resource/R-3136>
<https://reld.dice-research.org/resource/R-3143>
<https://reld.dice-research.org/resource/R-3144>
<https://reld.dice-research.org/resource/R-3633>
<https://reld.dice-research.org/resource/R-3703>
<https://reld.dice-research.org/resource/R-4021.0>
<https://reld.dice-research.org/resource/R-5490.0>
<https://reld.dice-research.org/resource/R-5494.0>
We have provided a script which can extract the required data from the RELD dataset in clean Json format for the selected relations. 6.(b) The feasible-exmp-10qa-benchmark.ttl file contain all the output data including relations, sentences, and features in .ttl format, like see example.