View on GitHub


Linked Data Regime

download .ZIPdownload .TGZ

Metablock: Semantic Publishing Tools

Metablock reads RDF data and writes a possibly modified resource description back to a RDF target.

RDF Data sources:
RDF data targets:
Compile && Configure
Indexing : write RDF data to a Solr search index

To build a Solr search index from a SPARQL service endpoint, three steps are required:

  1. Resource Enumeration: List all resources that should be indexed,

  2. Resource Dump: Query everything the triple store knows about a resource,

  3. Resource Tranformation: Transform RDF/XML to solr index format.

Step 1. and 2. need a SPARQL query, step 3 works with XSLT.
The sparql queries and xslt transformations used so far are rather general, but modelling of bibliographic resources may vary and require modification.

All configurations are done in turtle (see lib/metablock.ttl)

RDF Crawling

RDF data sources configured in lib/metablock.ttl can be tested with

java -jar metablock.jar -s [source] -t [target] -test

Copy from RDF source to RDF target

java -jar metablock.jar -s [source] -t [target]
RDF analyzers

PDF Analyzer: utilizes Grobid / CERMINE to extract metadata and bibliographical references from scientific articles

java -jar metablock.jar -s [source] -t [target] -e pdf

Reference Analysis: use external libraries to find citation context and determine citation polarity

java -jar metablock.jar -s [source] -t [target] -e sen
  1. Index a directory of PDF files, enable pdf engine to extract metadata and write to VuFind:

    java -jar metablock.jar -crawl -s files -t solr1 -e pdf Documents
  2. Write DSpace metadata from DSpace REST API to Virtuoso triplestore (experimental):

    java -jar metablock.jar -crawl -s dspace -t virt
  3. Crawl OAI sources to a jena TDB store:

    java -jar metablock.jar -crawl -s oai -t tdb
  4. Build a search index for a TDB triple store

    java -jar metablock.jar -crawl -s tdb -t solr1
Developer Documentation

Javadoc is available online.