Validatrr

Validation using rule-based reasoning.

installation

first have EYE and Python installed
npm install
node gather.js

run

./bin/validate_owl -i [inputfile] -o outputfile -s [comma-delimited schemas, e.g. 'skos,foaf'] --ram [amount of RAM, e.g. 2g, 512m] -r [reasoning profile, rdfs and/or owl]

--ram and -r are optional.

the prefices of the schemas are saved locally at resources/ontologies/prefix.json. If a prefix isn't found, you can also enter the full URI If you add no schemas, N3Unit will look for all included schemas (best effort)

test

npm test

The testing goes as follows:

1. gather all rule files (see gather.js)
   1. take extra files rdfsResource and rdfsSubClassOf
   2. take all rule files under resources/rules/*/rule.n3
   3. combine them in a pvm under dist/n3unit.pvm
2. for every profile (in this case only owl)
   1. take the testing ontology file (in this case test/owl/ontology.ttl)
   2. for every test file (in this case test/owl/*Correct-Wrong*.ttl)
      1. run the validation with the pvm, the ontology, the test file, and the profile (in this case profiles/owl.n3)
      2. put the ttl output in tmp/tmp-CONSTRAINT.n3
      3. put the stderr output in tmp/tmp-CONSTRAINT_err.n3
      4. check: if it should be correct, you should receive 0 violations, and otherwise you should receive more than 0

Docker - Validate

Docker does the installation for you, so you only need to do

docker build -t n3unit .
docker run n3unit -it -i test/test.ttl -o test.ttl -s foaf

There's a test file at test/test.ttl

Docker - RML Validate

Docker does the installation for you, so you only need to do

docker build -f cmd.Dockerfile -t n3unit-rml .
docker run -it -v [absolute path to local resources folder]:/usr/local/n3unit/resources n3unit-rml [path to rml file] > output.ttl

Make sure the path to rml file is within the root folder of this project

TODO

• Documentation about ./bin/cmd
• Documentation about ./bin/fetchTurtle
• Separate libs for fetchTurtle and eyeHandler
• Constraint generator (from generic to anything and vice versa)