My favorites ▼ | Sign in
Project Home Downloads Wiki Issues Source
Search
for
SesameHowto  
Howto get started with loading the data in Sesame
Updated Feb 4, 2010 by sina.hes...@gmail.com

Introduction

Sesame is an open source framework and database for working with RDF graphs. It comes with libraries for parsing xml, linking to a number of databases, including a specialised RDF database, called the Native Sail. It also integrates a SPARQL endpoint. Here we describe how to load the NetBeans bug database into your Sesame instance.

Details

Database Preparation

Here we load the data into a native database, which becomes queriable via the com.xmlns.baetle.sesame.Query program in the subversion repository.

  1. Download Sesame 2 beta 2
  2. Download the 70 MB bziped package and unzip it. You should now have a 1 GB file in N-Triples format.
  3. Check out the baetle subversion repository
  4. Compile the com.xmlns.baetle.sesame.LoadRDF class (you need to point to the Sesame jars)
  5. Run com.xmlns.baetle.sesame.LoadRDF -d datadir baetle.ntriples command

This will take some time. You will see a bunch of files appearing in your data directory.

Now, you can use the Query class to send queries to the database. It's very fast on a 2 GHz machine. (Once the indexing has been properly done) 

hjs@bblfish:0$ bin/openSesame com.xmlns.baetle.sesame.Query -d /Users/hjs/.aduna/openrdf/server/native -f json
using store in directory /Users/hjs/.aduna/openrdf/server/native
SELECT * 
WHERE { ?i a <http://xmlns.com/baetle/#Issue> . }
LIMIT 1

28139 [main] INFO org.openrdf.query.parser.QueryParserUtil - Unable to load query parser class: org.openrdf.query.parser.serqo.SeRQOParser
{
        "head": {
                "vars": [ "i" ]
        }, 
        "results": {
                "ordered": false, 
                "distinct": false, 
                "bindings": [
                        {
                                "i": { "type": "uri", "value": "http://www.netbeans.org/issues/show_bug.cgi?id=296#it" }
                        }
                ]
        }
}

Server Install

Here we create an HTTP accessible SPARQL endpoint, so that web applications can query the database. See the Sesame Docs for more detailed info (please fill out extra things)

  1. download Tomcat 6 (it's easier)
  2. unzip tomcat
  3. place the datadir in you ~/.aduna/openrdf/server/ directory. Call that dir "native" (on unix machines) for Windows see Sesame Docs.
  4. edit ~/.aduna/openrdf/server/repositories.xml 
    5. <?xml version="1.0" encoding="ISO-8859-1"?>
<openrdf-sesame>
   <admin password="admin"/>
   <repositorylist>
       <repository id="native">
         <title>Native Repository</title>
         <sailstack>
             <sail class="org.openrdf.sail.nativerdf.NativeStore">
                  <param name="triple-indexes" value="spoc,sopc,posc,psoc,opsc,ospc"/>
             </sail>
         </sailstack>
         <acl worldReadable="true" worldWriteable="true" />
       </repository>
    </repositorylist>
</openrdf-sesame>
  5. start tomcat
  6. in the manager application load the two wars from the Sesame distribution
  7. Send queries to the database:
    • go to the sesame client app and use the web interface to send the queries
    • use the com.xmlns.baetle.sesame.Query from the command line to send your queries 
hjs@bblfish:0$ bin/openSesame -Xms125m -Xmx500m com.xmlns.baetle.sesame.Query -s http://localhost:8080/openrdf/
using repository at http://localhost:8080/openrdf/
PREFIX : <http://xmlns.com/baetle/#>
CONSTRUCT { ?issue :depends_on ?other . }
WHERE { ?issue :depends_on ?other . }
LIMIT 5

@prefix : <http://xmlns.com/baetle/#> .

@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

<http://www.netbeans.org/issues/show_bug.cgi?id=2337#it> 
   :depends_on <http://www.netbeans.org/issues/show_bug.cgi?id=31362#it> ,
               <http://www.netbeans.org/issues/show_bug.cgi?id=32942#it> ,
               <http://www.netbeans.org/issues/show_bug.cgi?id=35023#it> , 
               <http://www.netbeans.org/issues/show_bug.cgi?id=46020#it> .

<http://www.netbeans.org/issues/show_bug.cgi?id=2419#it> 
   :depends_on <http://www.netbeans.org/issues/show_bug.cgi?id=7065#it> .

Sending queries from JSON

Note: this is now documented in "HTTP communication protocol for Sesame 2"

The requests between the client and the server are simple HTTP requests. I think these are following the standard SPARQL Application Protocol. Using a tool such as tcpflow (which suddenly stopped working on my intel core 2 duo with OSX 10.4.9) or tshark you can get a good idea of how to call this directly.

So I started tshark like this 

sudo tshark -i lo0 -w - -R "tcp.port == 8080" > out

and ran the following Query 

hjs@bblfish:0$ bin/openSesame com.xmlns.baetle.sesame.Query -s http://localhost:8080/openrdf/ 
using repository at http://localhost:8080/openrdf/
SELECT *
WHERE { ?s ?r ?b . }
LIMIT 1

The tshark output showed the above command to have sent the following request over the wire 

GET /openrdf/repositories/native?queryLn=SPARQL&query=SELECT+*WHERE+%7B+%3Fs+%3Fr+%3Fb+.+%7DLIMIT+1&infer=true HTTP/1.1
Accept: application/sparql-results+xml
Accept: application/x-binary-rdf-results-table
Accept: application/sparql-results+json
User-Agent: Jakarta Commons-HttpClient/3.0.1
Host: localhost:8080

(notice how the SPARQL endpoint is really at http://localhost:8080/openrdf/repositories/native)

The above request was followed by the response 

HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Disposition: attachment; filename=queryresult.srx
Content-Type: application/sparql-results+xml
Transfer-Encoding: chunked
Date: Sun, 01 Apr 2007 11:16:18 GMT

<?xml version='1.0' encoding='UTF-8'?>
<sparql xmlns='http://www.w3.org/2005/sparql-results#'>
        <head>
                <variable name='s'/>
                <variable name='r'/>
                <variable name='b'/>
        </head>
        <results ordered='false' distinct='false'>
                <result>
                        <binding name='b'>
                                <uri>http://rdfs.org/sioc/ns#User</uri>
                        </binding>
                        <binding name='r'>
                                <uri>http://www.w3.org/1999/02/22-rdf-syntax-ns#type</uri>
                        </binding>
                        <binding name='s'>
                                <uri>http://netbeans.org/people/10</uri>
                        </binding>
                </result>
        </results>
</sparql>

So to get back a JSON response I ran the query directly from the command line but only asking for the json representation 

hjs@bblfish:0$ telnet localhost 8080
Trying ::1...
Connected to localhost.
Escape character is '^]'.
GET /openrdf/repositories/native?queryLn=SPARQL&query=SELECT+*WHERE+%7B+%3Fs+%3Fr+%3Fb+.+%7DLIMIT+1&infer=true HTTP/1.1
Accept: application/sparql-results+json
User-Agent: Jakarta Commons-HttpClient/3.0.1
Host: localhost:8080

HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Disposition: attachment; filename=queryresult.srj
Content-Type: application/sparql-results+json
Transfer-Encoding: chunked
Date: Sun, 01 Apr 2007 11:20:20 GMT

164
{
        "head": {
                "vars": [ "s", "r", "b" ]
        }, 
        "results": {
                "ordered": false, 
                "distinct": false, 
                "bindings": [
                        {
                                "b": { "type": "uri", "value": "http://rdfs.org/sioc/ns#User" }, 
                                "r": { "type": "uri", "value": "http://www.w3.org/1999/02/22-rdf-syntax-ns#type" }, 
                                "s": { "type": "uri", "value": "http://netbeans.org/people/10" }
                        }
9

                ]
        }
}
0

(I don't know why I get those extra numbers in there... answer: those are probably due to chunked transfer encoding, which a good http library should take care of. )

JavaScript experts should find it easy to build up such calls and get their preferred response.


► Sign in to add a comment
Powered by Google Project Hosting