rdf-interchange

Examples

There are 14 examples. They are deliberately flat: every example is the same five-file pipeline.

<name>-input.ttl|trig   RDF input
<name>-input.pl         rdf-to-prolog result
<name>-rules.pl         Prolog rules
<name>-output.pl        Prolog test result as ground rdf/4 facts
<name>-output.ttl|trig  prolog-to-rdf result

Every *-rules.pl exposes the same two predicates:

result_rdf(S, P, O, G).  % enumerate the result
write_results.           % print every result as an rdf/4 fact

Example: trust flow

The five files are:

odrl-dpv-fpv-trust-flow-input.ttl
odrl-dpv-fpv-trust-flow-input.pl
odrl-dpv-fpv-trust-flow-rules.pl
odrl-dpv-fpv-trust-flow-output.pl
odrl-dpv-fpv-trust-flow-output.ttl

Regenerate the input Prolog facts:

rdf-to-prolog examples/odrl-dpv-fpv-trust-flow-input.ttl \
  -o examples/odrl-dpv-fpv-trust-flow-input.pl

Load the input and rules in a Prolog engine and query:

?- result_rdf(S, P, O, G).

Or run write_results/0 to print the complete result as rdf/4 facts. With EyeProlog 1.5.26 or newer, --quiet keeps the resolved command goal out of the materialized fact stream:

eyeprolog --quiet --goal write_results \
  examples/odrl-dpv-fpv-trust-flow-input.pl \
  examples/odrl-dpv-fpv-trust-flow-rules.pl \
  > examples/odrl-dpv-fpv-trust-flow-output.pl

The same rules can, for example, be materialized with SWI-Prolog:

swipl -q \
  -s examples/odrl-dpv-fpv-trust-flow-input.pl \
  -s examples/odrl-dpv-fpv-trust-flow-rules.pl \
  -g write_results -t halt \
  > examples/odrl-dpv-fpv-trust-flow-output.pl

Only the command-line flags are host-specific.

eyepl evaluates the same rules and lists the answers of the query as result_rdf/4 facts:

printf '?- result_rdf(S, P, O, G).\n' > query.pl
eyepl examples/odrl-dpv-fpv-trust-flow-input.pl \
  examples/odrl-dpv-fpv-trust-flow-rules.pl query.pl > result.pl
prolog-to-rdf result.pl -o result.ttl

prolog-to-rdf reads those result_rdf/4 facts like rdf/4 facts.

The rules are ISO Prolog with two widely supported additions. Four examples declare :- table for helpers that are called with some arguments already known, and three rank results with aggregate_all(count, Goal, Count). EyeProlog, SWI-Prolog, and eyepl run all 14 examples. A table declaration evaluates a helper on demand; engines that otherwise run rules forward, such as eyepl, need it for helpers that cannot start without those arguments.

Convert the Prolog result back to RDF:

prolog-to-rdf examples/odrl-dpv-fpv-trust-flow-output.pl \
  -o examples/odrl-dpv-fpv-trust-flow-output.ttl

For TriG inputs, the final filename is *-output.trig and named graphs are preserved.

This site is open source. Improve this page.