This post is about using Apache Velocity ( a Java-based template engine ) and the Jena RDF library. My aim was to use Velocity to handle the content of one or more RDF store without compiling, just by using a custom velocity template. This idea was much inspired by Egon Willighagen's posts where the RDF was handled with a scripting engine embedded in bioclipse. It also seems that I'm not the first who had this idea of using Velocity+RDF: see [here].
OK, my experimental source code for the program JenaVelocity is available here:
On the command line, one (or more) RDF dataset is described as a JSON document. In the following example this is a remote file, but it could also be the description of a persistent database, a N3 file, etc... This RDF file will also be used later for resolving some names from the bio2rdf repository, this is why I also added a table for the prefix mappings. This RDF model will be inserted in the VelocityContext under the name "$store1"
After running JenaVelocity, I got the following result:
Here, I've inserted an object called $sparql in the VelocityContext. This object is used to send a SPARQL query to the bio2rdf sparql endpoint and the Statements related to the rdf:type
After running JenaVelocity, I got the following result:
Conclusion: Velocity templates allow to handle and render some RDF data without compiling anything. However, a prior knowledge of the Jena API is required.
That's it.
Pierre
OK, my experimental source code for the program JenaVelocity is available here:
Describing the RDFstores
On the command line, one (or more) RDF dataset is described as a JSON document. In the following example this is a remote file, but it could also be the description of a persistent database, a N3 file, etc... This RDF file will also be used later for resolving some names from the bio2rdf repository, this is why I also added a table for the prefix mappings. This RDF model will be inserted in the VelocityContext under the name "$store1"
[
{
"name": "store1",
"url":"http://www.lri.fr/~pietriga/foaf.rdf",
"prefix-mapping":{
"uniprot":"http://bio2rdf.org/uniprot:"
}
}
]
{
"name": "store1",
"url":"http://www.lri.fr/~pietriga/foaf.rdf",
"prefix-mapping":{
"uniprot":"http://bio2rdf.org/uniprot:"
}
}
]
Example 1
In the following example: all the RDFstores are inserted in the VelocityContext as $rdfstores. For each rdfstore a HTML list is created. The list contains the names of all the infividuals of the FOAF file described previously.<html><body>
#set($RDF="http://www.w3.org/1999/02/22-rdf-syntax-ns#")
#set($FOAF="http://xmlns.com/foaf/0.1/")
<h1>Staff</h1>
<ul>
#foreach($store in $rdfstores)
#set($pred = ${store.model.createProperty("${FOAF}","name")})
#foreach($stmt in
${store.model.listStatements(null,${store.model.createProperty("${FOAF}","name")},null,null)})
<li>${stmt.object.string}</li>
#end
</ul></body></html>
#set($RDF="http://www.w3.org/1999/02/22-rdf-syntax-ns#")
#set($FOAF="http://xmlns.com/foaf/0.1/")
<h1>Staff</h1>
<ul>
#foreach($store in $rdfstores)
#set($pred = ${store.model.createProperty("${FOAF}","name")})
#foreach($stmt in
${store.model.listStatements(null,${store.model.createProperty("${FOAF}","name")},null,null)})
<li>${stmt.object.string}</li>
#end
</ul></body></html>
After running JenaVelocity, I got the following result:
Staff
- Jean-Daniel Fekete
- Chris Bizer
- Caroline Appert
- Ralph Swick
- Vincent Quint
- Jean-Yves Vion-Dury
- Yves Guiard
- Eric Miller
- Renaud Blanch
- Emmanuel Pietriga
- Jose Kahan
- Eric Prud'hommeaux
- Catherine Letondal
- Olivier Chapuis
- Michel Beaudouin-Lafon
- Nicolas Roussel
- Ryan Lee
- Wendy Mackay
Example 2
Here, I've inserted an object called $sparql in the VelocityContext. This object is used to send a SPARQL query to the bio2rdf sparql endpoint and the Statements related to the rdf:type
http://bio2rdf.org/ns/uniprot:Strain are fetched and displayed in a HTML table. For each Resource, we try to get a short form of its URI using our previously defined $store1. It the object of a statement is a literal, the quoted string is printed.<html><body>
#set($RDF="http://www.w3.org/1999/02/22-rdf-syntax-ns#")
#set($FOAF="http://xmlns.com/foaf/0.1/")
<h1>Strains</h1>
<table>
#foreach($row in
$sparql.select("http://quebec.bio2rdf.org/sparql","select distinct ?s
?p ?o where { ?s a <http://bio2rdf.org/ns/uniprot:Strain> . ?s ?p ?o}
LIMIT 100"))
<tr>
<td><a href="${row.get("s").getURI()}">${store1.shortForm(${row.get("s").getURI()})}</a></td>
<td><a href="${row.get("p").getURI()}">${store1.shortForm(${row.get("p").getURI()})}</a></td>
<td>#if(${row.get("o").isResource()})
<a href="${row.get("o").getURI()}">${store1.shortForm(${row.get("o").getURI()})}</a>
#else
<span>"$row.get("o").string"</span>
#end</td>
</tr>
#end
</table>
#end
</body></html>
#set($RDF="http://www.w3.org/1999/02/22-rdf-syntax-ns#")
#set($FOAF="http://xmlns.com/foaf/0.1/")
<h1>Strains</h1>
<table>
#foreach($row in
$sparql.select("http://quebec.bio2rdf.org/sparql","select distinct ?s
?p ?o where { ?s a <http://bio2rdf.org/ns/uniprot:Strain> . ?s ?p ?o}
LIMIT 100"))
<tr>
<td><a href="${row.get("s").getURI()}">${store1.shortForm(${row.get("s").getURI()})}</a></td>
<td><a href="${row.get("p").getURI()}">${store1.shortForm(${row.get("p").getURI()})}</a></td>
<td>#if(${row.get("o").isResource()})
<a href="${row.get("o").getURI()}">${store1.shortForm(${row.get("o").getURI()})}</a>
#else
<span>"$row.get("o").string"</span>
#end</td>
</tr>
#end
</table>
#end
</body></html>
After running JenaVelocity, I got the following result:
Strains
Conclusion: Velocity templates allow to handle and render some RDF data without compiling anything. However, a prior knowledge of the Jena API is required.
That's it.
Pierre
Nice!
ReplyDelete