Building and deploying the TIMBUS Context Acquisition tool

miguel gama nunes

Installing the TIMBUS Context Acquisition tool

Installing Virgo Jetty Server:

  • Download "Virgo Jetty Server 3.6.2.RELEASE": from http://www.eclipse.org/virgo/download
  • Extract Virgo Jetty Server and move it to directory /opt:
    $> unzip virgo-jetty-server-3.6.2.RELEASE.zip
    $> sudo mv virgo-jetty-server-3.6.2.RELEASE /opt
  • check if Virgo's folder doesn't require root privileges to perform changes

 

Virgo Repository:

  • Download virgo-repo-usr-1.tgz and virgo-repo-usr-2.tgz
  • Extract Virgo Repository files and move them to Virgo directory /repository/usr:
    $> mkdir temp-repository
    $> cd temp-repository
    $> tar -zxvf ../virgo-repo-usr-1.tgz
    $> tar -zxvf ../virgo-repo-usr-2.tgz
    $> mv * /opt/virgo-jetty-server-3.6.2.RELEASE/repository/usr/

 

Jackson:

  • Download jackson-xc-1.9.13.jar and move it to Virgo directory /pickup:
    $> mv jackson-xc-1.9.13.jar /opt/virgo-jetty-server-3.6.2.RELEASE/pickup/

 

Configure the Virgo port:

  • The default port is 8080. To change it, edit the configuration file on Virgo directory /jetty/etc/jetty.xml.

 

$> ./startup.sh@ or @> ./startup.sh -clean@

 

Logging:

  • All Virgo logs are stored in Virgo directory /serviceability/logs/log.log

 

Managing applications/bundles:
- Browse to localhost:8080 (or another port if you have configured so before). You should now see the Virgo splash screen:

alt text

  • To manage applications/bundles, go to "Admin Console" (the default authentication is: user "admin" and password "springsource")
  • Applications are deployed in the "Artifacts" tab and displayed under the "bundle" tree node

 

Deploying Extractors and Context Populators:

alt text

  • Two new bundles should have appeared now:

alt text

Now that the core bundles are deployed, we can deploy the business-specific bundles (extractors and populators):

 

Operating with the Context Acquisition tool

To use the APIs, POST requests must be sent with @Content-Type@ header set to application/json and raw body

Extractors API (GET if not mentioned):
- /extractors/api/extractors (lists all available extractors)
- /extractors/api/requests (lists all requests sent)
- /extractors/api/requests/{id} (shows a specific request, by id)
- /extractors/api/extract (POST)
Context Populator API (GET if not mentioned):
/populator/api/converters (lists all available converters)
/populator/api/requests (lists requests sent)
/populator/api/requests/{id} (shows a specific request, by id)
/populator/api/convert (POST)
/populator/api/reset (resets DIO and CUDF datasets in testbed's Jena Fuseki)

Extractors "extract" input:

for DSpace:

1
2
3
4
{"extractions":[{
    "fqdn":"...","port":80,"module":"DSpace Extractor"
    }
]}

for software on Intel's testbed using timbus-cms' private key:

1
2
3
4
{"extractions":[{
    "user":"timbus-cms","fqdn":"testbed.timbusproject.net",
    "port":22,"module":"Debian Software Extractor"
}]}

for other software requests:

1
2
3
{"extractions":[{"user":"...","fqdn":"...","password":"...",
    "port":22,"module":"Debian Software Extractor"
}]}

Context Populators "convert" input:

1
{"url":"http://localhost:8080/extractors/api/requests/{id}","requestId":"{id}"}

("url" is the full URI path to the request intended to be converted. Same ID should be specified in both_ url and requestId keys)