outE (out edges) and inE (in edges) methods

[cristianvasquez]

Hi,

I was trying to 'follow my nose' and explore properties and objects using clownface, but it was difficult for me to explore 'properties' (for example, to get the label of a property)

I added the outE (out edges) and inE (in edges) methods to experiment. I find them useful.

This is the inspiration: https://github.com/mpollmeier/gremlin-scala/blob/master/gremlin-scala/src/test/scala/gremlin/scala/ArrowSyntaxSpec.scala

I made this pull request to discuss the addition of these two functions

[cristianvasquez]

Motivation:

import rdf from 'rdf-ext'
import clownface from 'clownface'

const ns = {
  ex: rdf.namespace('http://example.org/'),
}
const dataset = rdf.dataset()
dataset.addAll([
  rdf.quad(ns.ex.bob, ns.ex.exoticProperty, rdf.literal('Bob')),
  rdf.quad(ns.ex.exoticProperty, ns.ex.label, rdf.literal('First name')),
])

const bob = clownface({ dataset: dataset })
   .namedNode(ns.ex.bob)

// What's the label of exoticProperty?
console.log(bob.outE(ns.ex.exoticProperty).out(ns.ex.label).value)

// outputs 'First name'

[tpluscode]

I support this effort, albeit I have not looked closely yet. That said, the motivation does not give it justice. What you describe can simply be done as

-bob.outE(ns.ex.exoticProperty).out(ns.ex.label).value
+bob.node(ns.ex.exoticProperty).out(ns.ex.label).value

What is missing, and came up a bunch of times, is to discover properties of bob

// will return pointer to ex:exoticProperty
const properties = bob.outE()

[cristianvasquez]

As it is, we can also use it to discover the properties of bob

I agree that that is the best use case.

import rdf from 'rdf-ext'
import clownface from 'clownface'

const ns = {
  ex: rdf.namespace('http://example.org/'),
}
const dataset = rdf.dataset()
dataset.addAll([
  rdf.quad(ns.ex.bob, ns.ex.exoticProperty, rdf.literal('Bob')),
  rdf.quad(ns.ex.exoticProperty, ns.ex.label, rdf.literal('First name')),
  rdf.quad(ns.ex.bob, ns.ex.age, rdf.literal(99)),
])

const bob = clownface({ dataset: dataset })
   .namedNode(ns.ex.bob)

console.log(bob.outE().values)
// Outputs
// [ 'http://example.org/exoticProperty', 'http://example.org/age' ]

[bergos]

In the RDF world, the term edge is not used that much. predicate and property are more common, where predicate is used in the context of triples and property in a higher level context. property is therefore the better choice for clownface. prop has already become a synonym for property for software developers. That would be in balance of having a short method name, but still one with a speaking name.

-> E should be replaced with prop

[cristianvasquez]

@bergos, the property is renamed. Now that you mention it, I'm a bit unsure about prop instead of predicate (or pred), because in general, the API is using 'predicate'

[bergos]

The predicate is used only for indexed argument variables. It's somehow in the docs but not really part of the interface. For sure, I didn't spend much time thinking about the naming. But you are right, that's a conflict. We should align the internal variables to the public interface, but let's do that in a separate PR to keep this one small.