Prototype deployment details

What's first

• Getting started
• Sample FAqT deployment

What we'll cover

This page covers the details for how to set up the FAqT Brick at http://sparql.tw.rpi.edu/datafaqs/dump (and http://sparql.tw.rpi.edu:3030/datafaqs/query) and the FAqT Brick Explorer at http://aquarius.tw.rpi.edu/datafaqs. It can be used as a template for setting up other FAqT Bricks and explorers. This DOES NOT cover how to deploy FAqT Services; it assumes that it has already been done.

Let's get to it

sparql.tw

Get the DataFAQs utilities:

• cd /opt
• sudo git clone git://github.com/timrdf/DataFAQs.git

Get some dependencies:

csv2rdf4lod-automation is needed for a couple of RDF-handling utilities.

• cd /opt
• sudo git clone git://github.com/timrdf/csv2rdf4lod-automation.git
• ./install.sh

tdb is needed to create a triple store of the RDF files accumulated in the FAqT Brick

• sudo mkdir /opt/tdb
• cd /opt/tdb
• curl -L http://sourceforge.net/projects/jena/files/TDB/TDB-0.8.10/tdb-0.8.10.zip/download > tdb-0.8.10.zip
• gunzip

fuseki is needed to expose the tdb triple store as a SPARQL endpoint

• sudo mkdir /opt/fuseki
• cd /opt/fuseki
• wget http://openjena.org/repo/org/openjena/fuseki/0.2.0/fuseki-0.2.0.zip
• gunzip

virtuoso can also be used in addition to -- or in place of -- fuseki.

• TODO describe after we've migrated to datafaqs.aquarius.tw

rapper is needed to retrieve and reserialize RDF (and count triples)

• see "rapper" here

Set up a FAqT Brick:

• sudo mkdir /srv/DataFAQs
• sudo chown user:group /srv/DataFAQs
• mkdir -p /srv/DataFAQs/default/faqt-brick

Set the following [environment variables](DATAFAQS environment variables) in /srv/DataFAQs/default/faqt-brick/datafaqs-source-me.sh:

export DATAFAQS_HOME="/opt/DataFAQs"
export DATAFAQS_BASE_URI="http://sparql.tw.rpi.edu"
export DATAFAQS_PUBLISH_TDB="true"
export DATAFAQS_PUBLISH_TDB_DIR="/srv/DataFAQs/default/faqt-brick/tdb"
export DATAFAQS_PUBLISH_THROUGHOUT_EPOCH="true"
export DATAFAQS_PUBLISH_METADATA_GRAPH_NAME="http://www.w3.org/ns/sparql-service-description#NamedGraph"
export CSV2RDF4LOD_HOME="/opt/csv2rdf4lod-automation"
export TDBROOT="/opt/tdb/TDB-0.8.10/"
export PATH=$PATH`/opt/DataFAQs/bin/df-situate-paths.sh`

Create an epoch (see FAqT Brick):

• cd /srv/DataFAQs/default/faqt-brick
• /opt/DataFAQs/bin/df-epoch.sh

Link the FAqT Brick to the web:

• sudo ln -s /srv/DataFAQs/default/faqt-brick /var/www/datafaqs/dump
• Browse around http://sparql.tw.rpi.edu/datafaqs/dump/__PIVOT_epoch

If tdb didn't load during epoch creation, load the triple store:

• df-load-triple-store.sh --recursive-by-sd-name

Turn on the SPARQL endpoint:

• cd /opt/fuseki/Fuseki-0.2.0
• ./fuseki-server --loc=/srv/DataFAQs/default/faqt-brick/tdb /datafaqs
• Check that http://sparql.tw.rpi.edu:3030/datafaqs/query says "Error 404: Service Description"
• See what graphs are in the endpoint as xml or json

Deploy lodspeakr:

(aquarius.tw)

• mkdir /var/www/datafaqs
• cd /var/www/datafaqs
• bash < <(curl -Ls http://purl.org/twc/datafaqs-lodspeakr/install)
  • Answer 1/3: http://aquarius.tw.rpi.edu/datafaqs/
  • Answer 2/3: (hit enter for default)
  • Answer 3/3: http://sparql.tw.rpi.edu:3030/datafaqs/query
• http://aquarius.tw.rpi.edu/datafaqs/DataFAQs-lodspeakr
• http://aquarius.tw.rpi.edu/datafaqs-view/special/index showed classes in default graph (we aren't using the default graph)

What's next

• FAqT Brick to walk through the creation of an evaluation epoch.
• DATAFAQS environment variables needed when an FAqT Brick evaluation epoch is created.