Re: Input and output

FWIW in our (Jena-based) implementation, the input is a Dataset and a 
shapes graph and the output is an RDF Graph containing instances of 
sh:Error, sh:Warning etc. If someone activates validation in TopBraid 
Composer, we also send the currently selected resource (as the focus 
node) to the engine, and only report the validation triggered by that. 
The UI then walks through the sh:Error objects and highlights the 
affected properties on the form, based on sh:subject, sh:predicate, 
sh:object.

Do you think your group would like to get a copy of the API? We plan to 
make that open source (and I have prepared a github repository already), 
but I am hesitating to open this up while the spec is unstable and 
people waste time.

Holger


On 7/11/15 12:50 AM, Karen Coyle wrote:
> Thanks, Peter, for the description. However, we were hoping for actual 
> examples. The TopBraid implementation uses forms for input and I don't 
> know if one can capture what the form actually sends to the SHACL code 
> or the raw output from that. But that's the kind of thing we're 
> looking for.
>
> kc
>
> On 7/9/15 10:51 AM, Peter F. Patel-Schneider wrote:
>> My view is that SHACL validation takes two inputs
>> 1/ a SHACL shapes graph
>> 2/ an RDF data graph or dataset
>>
>> The output of SHACL validation is a set of constraint violations.
>> http://w3c.github.io/data-shapes/shacl/#violations states that these 
>> can be
>> encoded into an RDF graph and augmented with other information. 
>> Alternatively
>> you could think of these as just the results of the top-level SPARQL 
>> queries
>> corresponding to the shapes in the SHACL shapes graph.
>>
>> My test implementation of my proposal takes two URLs - for SPARQL 
>> endpoints
>> for the shape and data graphs - and prints the violations (i.e., the 
>> results
>> of the generated SPARQL queries).
>>
>> peter
>>
>>
>> On 07/09/2015 10:24 AM, Karen Coyle wrote:
>>> There are folks in my area who are interested in attempting to code 
>>> some SHACL
>>> experimentally -- in part as a way to see if it works for the Cultural
>>> Heritage data and situation. The sticking point appears to be a lack of
>>> description of inputs and outputs to SHACL.
>>>
>>> Since some of you have already done coding, could you provide some
>>> input/output examples that could help these folks get started?
>>>
>>> Thanks,
>>> kc
>>
>

Received on Saturday, 11 July 2015 01:54:37 UTC