virtuosodump

A setup of Virtuoso which allows to dump the content of the RDFstore by a HTTP request

Getting the whole content of a virtuoso store is typically done via sql scripts. This setup configures Virtuoso so that the data is dumped on a predefined location by issuing a HTTP request.

The advantage of this approach is that the creation of a dump now can be part of any scripting environment without requiring phyisical access to the server.

Configuration of the scripts

• the names of the dumped dataset files
• the maximum size of dumped dataset files
• the location of the dumped dataset files
• selection of the graphs to be dumped

Configuration of Virtuoso

• upload the configured scripts using isql (or via the conductor isql)

• create a link in the vsp directory to the location of the dumped dataset files ln -s /data/dumps

• configure in the browsing of the dump directory via the conductor:

  • Open the Web Application Server/Virtual Domains & Directories.
  • Click on the New Directory in front of the 0.0.0.0 interface port 80.
  • Select the Type: Filesystem
  • Enter the logical path (part of the URI) fields in the Path and the relative path of the file relative to the vsp/ folder in the Physical path field ( '/dumps/' in the case of the example).
  • Check the box Allow Directory Browsing to enable the directory contents of the course.

• configure in the browsing of the dump directory via the conductor:

  • Open the Web Application Server/Virtual Domains & Directories.
  • Click on the New Directory in front of the 0.0.0.0 interface port 80.
  • Select the Type: Filesystem
  • Enter the logical path (part of the URI) fields in the Path (e.g. /gen_dumps) and the relative path of the file relative to the vsp/ folder in the Physical path field ( '/dumps/' in the case of the example).
  • Enter the dumpgeneration script in the Post-processing Function field. (This script will be executed each time before the respons is sent). for example, DB.DBA.dumps_all_graphs
  • Check the box Allow Directory Browsing to enable the directory contents of the course.

Testing

• upload RDF to a graph which is can be selected for dumping
• test browsing only: http://localhost:8890/dumps
• test generation & browsing: http://localhost:8890/gen_dumps

Improvements

Maybe the gen_dump request should return the list of dataset files, then an download iterator can be build.

VOiD

Virtuoso supports the creation of VOiD descriptions for a graph. However the implementation assumes that all information for the meta data is deriveable from the Virtuoso configuration. This is in general not the case. The script here is copied from the Virtuoso OpensSource guthub.

Changes

• bug fix: recalling will erase the existing meta data before inserting new ones
• extension: allowing to specify the SPARQL endpoint url
• change: instead to create the VOiD dataset names as namespace(graph) + "#Dataset", it is now graph/void