FHIR RDF Tutorial

harold: I made some corrections

https://github.com/BD2KOnFHIR/BLENDINGFHIRandRDF

Meeting with Stanford about Protege issues

harold: Met with them on Friday about ontology URIs.
... Surprising discovery: Michel Dumontier asked if we can put it into RDF/XML, Eric and I mentioned that Turtle is becoming the de facto standard, they had not realized that.
... Protege was using Turtle 1.0 parser instead of 1.1.
... They will fix that.
... Don't know where they are on unnamed individuals (bnodes), but they were interested to know about it.
... But they also brought to my attention: the URLs used as ont URIs don't deref against the FHIR server.

dbooth: Not set up yet in the FHIR build process?

<hsolbrig> http://build.fhir.org/diagnosticreport-example-f001-bloodexam.xml.html

<hsolbrig> http://<server>/DiagnosticReport/f201

harold: In the XML rendering of f001 . . . my understanding is that the URI pattern should be server/DiagnosticReport/f201

<hsolbrig> http://build.fhir.org/diagnosticreport-example-f001-bloodexam.ttl.html

harold: I've said that that URI follows the FHIR pattern.

<hsolbrig> http://hl7.org/fhir/

harold: Stanford pointed out that neither of those URIs dereferences correctly.

# - ontology header ------------------------------------------------------------

<http://hl7.org/fhir/DiagnosticReport/f001.ttl> a owl:Ontology;

owl: imports fhir:fhir.ttl;
... versionIRI <http://build.fhir.org/DiagnosticReport/f001.ttl> .

harold: without the .ttl extension it redirects: http://hl7.org/fhir/DiagnosticReport/f001

[[

$ curl -I http://hl7.org/fhir/DiagnosticReport/f001

HTTP/1.1 301 Moved Permanently

Content-Length: 165

Content-Type: text/html

Location: http://hl7.org/fhir/DiagnosticReport/f001/

Server: Microsoft-IIS/6.0

X-Powered-By: ASP.NET

Date: Tue, 28 Mar 2017 15:14:30 GMT

]]

ISSUE: FHIR RDF ontology URLs do not resolve

<trackbot> Created ISSUE-40 - Fhir rdf ontology urls do not resolve. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/40/edit>.

harold: Eric presented yesterday at stanford on the work we're doing, and will present at AMIA conference on shex.

Protege and bnodes

eric: They listened but do not know how to solve the problem. May need to brainstorm with them.

harold: The problem is that protege does not cope with bnodes in individual records. It stops at the first bnode and says its an invalid name. It's still there but you cannot see it.

eric: I think it's the OWL API, and the reason we're seeing it is that we're doing something slightly unusual with protege -- something SPARQLy.
... Usually people want to name everything.

dbooth: How to follow up with them on fixing this?

eric: Email thread next week.

ISSUE: Protege does not show bnodes

<trackbot> Created ISSUE-41 - Protege does not show bnodes. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/41/edit>.

harold: It would be a huge gain to get this done.

eric: My plan was to update the turtle parser to 1.1, and ask them to fix the bnode problem.

ShEx update

harold: I will be presenting to Univ of Munich on FHIR RDF STU3 and shex tools, on Apr 3.

dbooth: Please announce it to W3C HCLS list if others are welcome to join.

Versioned references in FHIR/RDF (Harold)

harold: Grahame made the URIs the way that I envisioned them.

[[

# - ontology header ------------------------------------------------------------

<http://hl7.org/fhir/DiagnosticReport/f001.ttl> a owl:Ontology;

owl: imports fhir:fhir.ttl;
... versionIRI <http://build.fhir.org/DiagnosticReport/f001.ttl> .

]]

<hsolbrig> http://hl7.org/fhir/diagnosticreport-example-f001-bloodexam.xml.html

<hsolbrig> http://hl7.org/fhir/diagnosticreport-example-f001-bloodexam.ttl.html

harold: That link specifies STU3 in the URL:

[[

# - ontology header ------------------------------------------------------------

<http://hl7.org/fhir/DiagnosticReport/f001.ttl> a owl:Ontology;

owl: imports fhir:fhir.ttl;
... versionIRI <http://hl7.org/fhir/STU3/DiagnosticReport/f001.ttl> .

]]

harold: They all have a permanent URI, plus multiple version URIs.
... So the URI on the build site will diverge from the STU3 URI as we make updates.
... Grahame noticed this issue 10 minutes before publishing STU3, and put this in.
... I think this is the right way to do it.
... Next step is to meet with grahame and decide whether the keep the .ttl extension on the URI, or use conneg, and understand any objections in FHIR community to using these versioned URIs.

eric: We could work up an example that shows the benefit of versioned URIs.

harold: I had approached grahame about including a version variable in the build, and he pushed back.

<scribe> ACTION: Harold to initiate discussion with Grahame about adding conneg and a variable for ont version. [recorded in http://www.w3.org/2017/03/28-hcls-minutes.html#action01]

<trackbot> Created ACTION-79 - Initiate discussion with grahame about adding conneg and a variable for ont version. [on Harold Solbrig - due 2017-04-04].

ShEx status

eric: ShEx primer and test suite were published last night. Feedback is invited.

https://lists.w3.org/Archives/Public/semantic-web/2017Mar/0083.html

ADJOURNED

This is scribe.perl Revision: 1.152  of Date: 2017/02/06 11:04:15  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Present: Tomas_Montgomery Al_Pivonka Harold_Solbrig David_Booth Julia_Chan EricP
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth
Found Date: 28 Mar 2017
Guessing minutes URL: http://www.w3.org/2017/03/28-hcls-minutes.html
People with action items: harold