How to represent inputs and outputs bound to a workflow?

[svenvanderburg]

• (www.example.com/step1, hasOutputVar, www.example.com/output1)
• (www.example.com/output1, RDF:type, 'int')
• (www.example.com/step2, hasInputVar, www.example.com/input1)
• (www.example.com/input1, RDF:type, 'int')

Saying that (www.example.com/output1, RDF:type, www.example.com/input1) is conceptually wrong, because they are not semantically the same. Step 2 does not have to be connected to step1, it's just that in this particular workflow they are.

How to represent that connection within a specific workflow?

Related: This validation of inputs and outputs is wrong:

fairworkflows/fairworkflows/fairworkflow.py

Line 345 in f62db49

def _validate_inputs_and_outputs(self):

We need to update it based on what comes out of this discussion.

[svenvanderburg]

@tkuhn

[tkuhn]

I think this is covered in our Plex ontology profile. @rcelebi how is that done again?

[svenvanderburg]

@raar1 mentioned in #129:
Fair enough, as we discussed in the meeting earlier. Although I think we will need this back eventually, with some sort of 'connector' triples, perhaps. Maybe:

(workflow, has:varbinding, bindinguri)
(bindinguri, rdf:type, somenamespace:VarBinding)
(bindinguri, from, fairvariableuriA)
(bindinguri, from, fairvariableuriB)

(if you'll forgive the horrific abuse of notation above). I'm thinking of how e.g. CWL does it, where the workflow specifies that input A of step 2 comes from output B of step 2. I can check with Tobias/Remzi on this matter though since they're the semantic experts.