I have tried reading this a few times and am still a little lost... 

1)  I dont see how the context  "If a process ontology is used to control 
the Semantic XML mapping"  relates the to arguments for 
sml:ConvertXMLToRDF (sm:Module)

Converts an arbitrary XML input document into an RDF graph using the 
Semantic XML mapping approach. The input graph of this module may contain 
class definitions that have sxml: declarations attached to them and these 
will be used for the instances. For more, see Help > Import and Export > 
Creating, Importing, Querying, Saving XML documents with Semantic XML.

 
  [image: Template Module]See Also
   
   - 
   
http://composing-the-semantic-web.blogspot.com/2007/11/xmap-mapping-arbitrary-xml-documents-to.html
 
   - sml:ConvertRDFToXML

Arguments
*sml:baseURI*: The base URI of the new RDF (for the creation of the new 
class and property names).
*sml:replace* (xsd:boolean): [Optional] If true then the resulting output 
graph will not include the input graph, i.e. only the new triples will be 
returned.
*sml:xml*: The XML document that shall be converted to RDF. To avoid 
character encoding issues, we strongly recommend this value to be a 
reference to an already parsed XML document, and not a literal. In other 
words, use "Add SPARQL expression" from the drop down menu and enter 
?varName and do not use a string value such as {?varName}. The actual 
document parsing should be handled by predecessing modules such as 
sml:ImportXMLFromURL.
*sml:xmlType* (xsd:string): [Optional] An (optional) type indicator for the 
Semantic XML conversion. Current supported values are "XHTML" (treats the 
input as HTML source, and may run a tidy algorithm in case the HTML is not 
well-formed XHTML).

2) i think I can follow how to manually construct a rdfs class model with 
sxml:annotations - but an example would be really helpful here! 

3) I dont really see why you couldnt take an XSD and map it to a namespace 
rather than having to do this manually - am trying to minimise the number 
of 'unnatural acts' for someone who knows XML to see an equivalent RDF 
model.

So with the current state - if there really is a way to use the SXML 
annotations in ConvertXMLtoRDF, then I think I'd need to build a 
pre-processing step to read the XSD and generate the equivalent process 
model in a target namespace and inject these annotations...  not TQ is 
inspecting the XSD building those classes anyway - so i guess this is where 
you suggest throwing away the instances to get these classes.. 

I get the feeling all the pieces i need are there, but cant quite get a 
handle on how to access them individually in the right order :-)



On Thursday, 15 August 2019 10:41:59 UTC+10, Holger Knublauch wrote:
>
> Hi Rob,
>
> if you have already existing classes and properties with sxml: annotations 
> then the algorithm should reuse those instead of creating new classes. See 
> comment at sml:ConvertXMLToRDF:
>
> Converts an arbitrary XML input document into an RDF graph using the 
> Semantic XML mapping approach. The input graph of this module may contain 
> class definitions that have sxml: declarations attached to them and these 
> will be used for the instances. For more, see Help > Import and Export > 
> Creating, Importing, Querying, Saving XML documents with Semantic XML.
>
> One approach to produce a suitably annotated ontology is to import an XSD, 
> another is to import an XML instance file, then delete the instances and 
> adjust the namespaces. Use the resulting file as input to the 
> sml:ConvertXMLToRDF step.
>
> Holger
>
>
> On 15/08/2019 10:07, Rob Atkinson wrote:
>
> Checking I'm not missing something here.. can i import XML and have the 
> model (classes and properties) in a different namespace than the instance 
> data?
>
> when using the <sml:ConvertXMLtoRDF> tag (NB not able to use the XML 
> import wizard as i need a repeatable process under program control)
>
> you have the option of specifying an XSD..
>
> however AFAICT it forces both the model (elements => Class and property 
> mappings) and instances into the same namespace, in the same output graph..
>
> this is somewhat inconsistent with improvements elsewhere to separate 
> model and instances.
>
> I can deal with this by post-processing to strip model triples into a 
> separate graph, and update all the namespaces .. but if there was already a 
> way to control different namespaces for model and instances it would be 
> much better...
>
> Rob Atkinson
> -- 
> You received this message because you are subscribed to the Google Groups 
> "TopBraid Suite Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected] <javascript:>.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/topbraid-users/c5e86483-2869-4696-96a4-fb2991ec80db%40googlegroups.com
>  
> <https://groups.google.com/d/msgid/topbraid-users/c5e86483-2869-4696-96a4-fb2991ec80db%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"TopBraid Suite Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/topbraid-users/f9e46257-fc4c-4f85-92ea-e93f648839d5%40googlegroups.com.

Reply via email to