This feature is available in the Fedora 4 Alpha 2 release.
1. Finding Objects by Datastream Name or MIME Type
Create a few sample objects with either a PDF or a text file, or both attached as datastreams (these commands assume you have two files named test.pdf and test.txt in your current working directory, please substitute working filenames as needed):
PREFIX dc: <http://purl.org/dc/elements/1.1/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> PREFIX indexing: <http://fedora.info/definitions/v4/indexing#> DELETE { } INSERT { <> indexing:hasIndexingTransformation "default"; rdf:type indexing:indexable } WHERE { }
curl -X PUT "http://localhost:8080/rest/objects/101" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/101" curl -X PUT -H "Content-type: application/pdf" --data-binary "@test.pdf" "http://localhost:8080/rest/objects/101/master" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/101/master/fcr:metadata" curl -X PUT "http://localhost:8080/rest/objects/102" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/102" curl -X PUT -H "Content-type: text/plain" --data-binary "@test.txt" "http://localhost:8080/rest/objects/102/master" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/102/master/fcr:metadata" curl -X PUT "http://localhost:8080/rest/objects/103" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/103" curl -X PUT -H "Content-type: application/pdf" --data-binary "@test.pdf" "http://localhost:8080/rest/objects/103/master" curl -X PUT -H "Content-type: text/plain" --data-binary "@test.txt" "http://localhost:8080/rest/objects/103/text" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/103/master/fcr:metadata" curl -X PATCH -H "Content-type: application/sparql-update" --data-binary "@object-indexing.rdf" "http://localhost:8080/rest/objects/103/text/fcr:metadata"
Find objects with a datastream named "text":
prefix fcrepo: <http://fedora.info/definitions/v4/repository#> select ?object where { ?ds fcrepo:mixinTypes "fedora:datastream" . ?ds fcrepo:hasParent ?object . filter(str(?ds)=concat(str(?object),'/text/fcr:metadata')) }
Expected Results
object |
---|
<http://localhost:8080/rest/objects/103> |
Find objects with a PDF datastream:
prefix fcrepo: <http://fedora.info/definitions/v4/repository#> select ?object where { ?ds fcrepo:mixinTypes "fedora:datastream" . ?ds fcrepo:hasParent ?object . ?ds fcrepo:hasContent ?content . ?content fcrepo:mimeType "application/pdf" }
Expected Results
object |
---|
<http://localhost:8080/rest/objects/101> |
<http://localhost:8080/rest/objects/103> |
2. Attach metadata properties directly to objects
Create an object and attach the dc:title property:
curl -X PUT http://localhost:8080/rest/objects/201 echo "prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> prefix index: <http://fedora.info/definitions/v4/indexing#> insert data { <> rdf:type index:indexable . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/201 echo "prefix dc: <http://purl.org/dc/elements/1.1/> insert data { <http://localhost:8080/rest/objects/201> dc:title 'Foo' . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/201
Find objects with titles:
prefix dc: <http://purl.org/dc/elements/1.1/> select ?object ?title where { ?object dc:title ?title }
Expected Results
object | title |
---|---|
<http://localhost:8080/rest/objects/201> | "Foo" |
3. Find Objects By Collection Membership
Create three objects and three collections:
curl -X PUT http://localhost:8080/rest/objects/col1 curl -X PUT http://localhost:8080/rest/objects/col2 curl -X PUT http://localhost:8080/rest/objects/col3 curl -X PUT http://localhost:8080/rest/objects/obj1 curl -X PUT http://localhost:8080/rest/objects/obj2 curl -X PUT http://localhost:8080/rest/objects/obj3
Mark the objects indexable:
echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj1 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj2 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj3 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/col1 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/col2 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/col3
Link each object to a collection:
echo 'insert data { <http://localhost:8080/rest/objects/obj1> <http://fedora.info/definitions/v4/rels-ext#isMemberOfCollection> <http://localhost:8080/rest/objects/col1> . }' | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj1 echo 'insert data { <http://localhost:8080/rest/objects/obj2> <http://fedora.info/definitions/v4/rels-ext#isMemberOfCollection> <http://localhost:8080/rest/objects/col2> . }' | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj2 echo 'insert data { <http://localhost:8080/rest/objects/obj3> <http://fedora.info/definitions/v4/rels-ext#isMemberOfCollection> <http://localhost:8080/rest/objects/col3> . }' | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj3
Link the collections in a hierarchy:
echo 'insert data { <http://localhost:8080/rest/objects/col1> <http://fedora.info/definitions/v4/rels-ext#hasPart> <http://localhost:8080/rest/objects/col2> . }' | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/col1 echo 'insert data { <http://localhost:8080/rest/objects/col2> <http://fedora.info/definitions/v4/rels-ext#hasPart> <http://localhost:8080/rest/objects/col3> . }' | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/col2
Find objects directly attached to a collection:
select ?obj ?col where { ?obj <http://fedora.info/definitions/v4/rels-ext#isMemberOfCollection> ?col }
Expected Results
obj | col |
---|---|
<http://localhost:8080/rest/objects/obj1> | <http://localhost:8080/rest/objects/col1> |
<http://localhost:8080/rest/objects/obj2> | <http://localhost:8080/rest/objects/col2> |
<http://localhost:8080/rest/objects/obj3> | <http://localhost:8080/rest/objects/col3> |
Find objects directly or indirectly attached (recursive retrieval along the rels:hasPart chain). Applications developed using Fedora 3 may use Mulgara's walk() function for queries that navigate a hierarchy. That functionality is now part of standard SPARQL 1.1 (Jena's property paths documentation provides some good examples).
prefix rels: <http://fedora.info/definitions/v4/rels-ext#> select ?obj where { <http://localhost:8080/rest/objects/col1> rels:hasPart* ?col . ?obj rels:isMemberOfCollection ?col }
Expected Results
obj |
---|
<http://localhost:8080/rest/objects/obj1> |
<http://localhost:8080/rest/objects/obj2> |
<http://localhost:8080/rest/objects/obj3> |
Also link an object to a project:
curl -X PUT http://localhost:8080/rest/objects/proj1 echo "insert data { <> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://fedora.info/definitions/v4/indexing#indexable> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/proj1 echo "prefix ex: <http://example.org/> insert data { <http://localhost:8080/rest/objects/obj1> ex:project <http://localhost:8080/rest/objects/proj1> . }" \ | curl -X PATCH --upload-file - http://localhost:8080/rest/objects/obj1
Find objects in a collection or linked to a project:
prefix rels: <http://fedora.info/definitions/v4/rels-ext#> prefix ex: <http://example.org/> select ?obj where { { ?obj ex:project <http://localhost:8080/rest/objects/proj1> } UNION { ?obj rels:isMemberOfCollection <http://localhost:8080/rest/objects/col2> } }
Expected Results
obj |
---|
<http://localhost:8080/rest/objects/obj1> |
<http://localhost:8080/rest/objects/obj2> |
Count objects instead of listing them:
prefix rels: <http://fedora.info/definitions/v4/rels-ext#> prefix ex: <http://example.org/> select (count(distinct ?obj) as ?count) where { { ?obj ex:project <http://localhost:8080/rest/objects/proj1> } UNION { ?obj rels:isMemberOfCollection <http://localhost:8080/rest/objects/col2> } }
Expected Results
count |
---|
"2"^^<http://www.w3.org/2001/XMLSchema#integer> |