Update SPARQL in functions.py by dr-shorthair · Pull Request #47 · CSIRO-enviro-informatics/loci-integration-api

dr-shorthair · 2020-03-11T04:33:48Z

Suppress references to unused namespaces and deprecated predicates

benjaminleighton

This looks like a good start. We will need to build the new cache and test this code pretty thoroughly before we can merge it.

dr-shorthair · 2020-03-11T04:43:55Z

Yes I wonder if qb4st:crs --> geox:inCRS may have fallen through the cracks.
I would expect everything else to be OK.

benjaminleighton · 2020-03-11T04:52:09Z

-            rdf:predicate rdf:type ;
-            rdf:object prov:Location .
-    } .
+#    UNION


My best guess as to why this was here in the first place is because "overlaps" objects are treated as geo:Features but we want to exclude them from being returned as locations. I'm not sure though, @ashleysommer ?

Are these computed ahead and cached? If so, I would prefer to see an explicit feature-type for overlaps. Maybe loci:Overlap ? Then these can be explicitly excluded, rather than borrowing an unrelated feature-type.

Yes I agree it shouldn't be a feature but I can't remember all the reasoning around it from our older discussions. I believe we can actually get rid of overlaps objects altogether because all that information should be elsewhere but @ashleysommer has more history and probably more insight on this. We need to break up scope here as well. It would be good to merge whatever is the minimum set of changes to support the new LDAPIs and deal with larger refactors, like overlaps, later.

benjaminleighton · 2020-03-11T04:52:47Z

 SELECT <SELECTS>
 WHERE {
    {
        { 


Note this bracket might not be closed and might need commenting

Which bracket? there is nothing in the change.

benjaminleighton · 2020-03-11T04:54:36Z

-           rdf:predicate geox:transitiveSfOverlap;
-           rdf:object ?o  .
-        } UNION {
+#           ?s1 rdf:subject <URI> ;


I don't think we can take this out until we've rebuilt the linksets and that is beyond the scope of the LDAPI refactors currently underway. AFAIK we use geox:transitiveSfOverlap still and can't depend on geo:sfOverlaps yet

Lets verify.

so in tests with excelerator use cases, we either have to:
a) change the linksets and/or triples in the cache so that this API endpoint works with your proposed query @dr-shorthair
or
b) defer removing the union statement (i.e. keep it) so it works with the current linkset triples

b) is more expedient; but a) is probably what we should do in the longer term

benjaminleighton · 2020-03-11T05:11:16Z

-           rdf:predicate geox:transitiveSfOverlap;
-           rdf:object ?o  .
-        } UNION {
+#           ?s1 rdf:subject <URI> ;


Sorry it is the bracket above this comment I think it also needs commenting. Not sure though it just looks like an extra line needs commenting.

benjaminleighton · 2020-03-11T05:17:07Z

-            rdf:predicate rdf:type ;
-            rdf:object prov:Location .
-    } .
+#    UNION


Yes I agree it shouldn't be a feature but I can't remember all the reasoning around it from our older discussions. I believe we can actually get rid of overlaps objects altogether because all that information should be elsewhere but @ashleysommer has more history and probably more insight on this. We need to break up scope here as well. It would be good to merge whatever is the minimum set of changes to support the new LDAPIs and deal with larger refactors, like overlaps, later.

dr-shorthair · 2020-03-13T02:48:19Z

So far I have only looked at one file: functions.py
Is there SPARQL anywhere else (I can't find any).

jyucsiro · 2020-04-20T00:54:29Z

@dr-shorthair i think all the SPARQL in the loci-integration-api is in functions.py

dr-shorthair · 2020-04-22T23:05:54Z

I've simplified the geo:sfOverlaps query to match all sub-properties as well. See lines 645-655.
This should catch the geox:transitiveSfOverlap cases now, and ignore them later, without actually mentioning them.

# Conflicts: # functions.py

jyucsiro · 2020-04-23T00:07:30Z

 PREFIX dt: <http://linked.data.gov.au/def/datatype/>
 SELECT <SELECTS>
 WHERE {
+    ?p rdfs:subPropertyOf* geo:sfOverlaps .


this would work, except we haven't loaded the geox ontology in the cache yet.

See http://3.104.29.122/sparql?name=&infer=true&sameAs=true&query=PREFIX+geox%3A+%3Chttp%3A%2F%2Flinked.data.gov.au%2Fdef%2Fgeox%23%3E%0APREFIX+rdfs%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E%0Aselect+*%0Awhere+%7B%0A++++geox%3AtransitiveSfOverlap++++rdfs%3AsubPropertyOf+%3Fs+%3B++++%0A%7D

question is should the cache load the geox ontology also?

geox ontology here: https://github.com/CSIRO-enviro-informatics/geosparql-ext-ont/blob/master/geox-adds.ttl

dr-shorthair · 2020-04-23T00:20:31Z

Yes - it is a very small ontology. But we'll have to make sure that the deprecated elements are back in there. I'll let you know when that's done.

dr-shorthair · 2020-04-23T00:26:26Z

Shoudl be good now.

Update SPARQL in functions.py

06df70e

Suppress references to unused namespaces and deprecated predicates

dr-shorthair added bug Something isn't working enhancement New feature or request labels Mar 11, 2020

dr-shorthair requested review from ashleysommer and benjaminleighton March 11, 2020 04:33

benjaminleighton reviewed Mar 11, 2020

View reviewed changes

dr-shorthair requested a review from jyucsiro March 11, 2020 05:10

benjaminleighton reviewed Mar 11, 2020

View reviewed changes

dr-shorthair mentioned this pull request Mar 13, 2020

Update SPARQL queries to match new ontology descriptions when new cache available CSIRO-enviro-informatics/loci.cat#49

Open

Simon Cox added 2 commits April 23, 2020 09:03

Simplify query to match all sub-properties of geo:sfOverlaps

4ebb25d

Merge branch 'master' into simplify-queries-to-core-ontology

4464491

Merge branch 'master' into simplify-queries-to-core-ontology

03b1618

# Conflicts: # functions.py

jyucsiro reviewed Apr 23, 2020

View reviewed changes

Uh oh!

Conversation

dr-shorthair commented Mar 11, 2020

Uh oh!

benjaminleighton left a comment

Choose a reason for hiding this comment

Uh oh!

dr-shorthair commented Mar 11, 2020

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dr-shorthair Mar 11, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dr-shorthair commented Mar 13, 2020

Uh oh!

jyucsiro commented Apr 20, 2020

Uh oh!

dr-shorthair commented Apr 22, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jyucsiro Apr 23, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dr-shorthair commented Apr 23, 2020

Uh oh!

dr-shorthair commented Apr 23, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

dr-shorthair Mar 11, 2020 •

edited

Loading

dr-shorthair commented Apr 22, 2020 •

edited

Loading

jyucsiro Apr 23, 2020 •

edited

Loading