Issue 172: Support for sh:closed sh:ByTypes

[HolgerKnublauch]

Closes #172

• See this document rendered online here

[TallTed]

Suggested change

	<code>
	function collectProperties(S)<br />
	&nbsp;&nbsp;add all IRI properties that can be reached from S via the SPARQL path<br />
	&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;sh:property/sh:path<br />
	&nbsp;&nbsp;if S is a SHACL instance of rdfs:Class in the shapes graph {<br />
	&nbsp;&nbsp;&nbsp;&nbsp;for each triple in the shapes graph matching (S rdfs:subClassOf ?o)<br />
	&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;collectProperties(?o)<br />
	&nbsp;&nbsp;&nbsp;&nbsp;for each triple in the shapes graph matching (?s sh:targetClass S)<br />
	&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;collectProperties(?s)<br />
	&nbsp;&nbsp;}<br/>
	&nbsp;&nbsp;if S is a SHACL instance of sh:NodeShape in the shapes graph<br />
	&nbsp;&nbsp;&nbsp;&nbsp;for each triple in the shapes graph matching (S sh:node ?o)<br />
	&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;collectProperties(?o)<br />
	<br />
	for each rdf:type T of the value node in the data graph<br />
	&nbsp;&nbsp;collectProperties(T)<br />
	<code>
	<pre>function collectProperties(S)
	add all IRI properties that can be reached from S via the SPARQL path
	sh:property/sh:path
	if S is a SHACL instance of rdfs:Class in the shapes graph, then {
	for each triple in the shapes graph matching (S rdfs:subClassOf ?o)
	collectProperties(?o)
	for each triple in the shapes graph matching (?s sh:targetClass S)
	collectProperties(?s)
	}
	if S is a SHACL instance of sh:NodeShape in the shapes graph, then
	for each triple in the shapes graph matching (S sh:node ?o)
	collectProperties(?o)
	for each rdf:type T of the value node in the data graph
	collectProperties(T)</pre>

[HolgerKnublauch]

The pre was adding syntax highlighting in unsuitable places in my testing. Have you tried to render your change?

[YoucTagh]

I tried out the suggested changes, and indeed some unwanted highlighting was added.

[TallTed]

I do my editing in GitHub's web interface, so I can't see my changes rendered until they're merged into a PR and PR-Preview is invoked.

I guess there's some CSS used on other <pre> blocks to prevent the unwanted syntax highlighting. Changing from the manual formatting to the <pre> isn't vital; if I find the CSS incantation I'll try to apply it later.

[ajnelson-nist]

Is this line going to be universally necessary?

[ajnelson-nist]

I see in test 4's ex:RootClass shape that rdf:type isn't ignored.