Various DCAT tools for harvesting metadata from Belgian open data portals, converting metadata to DCAT-AP files and updating the Belgian data.gov.be portal.
The portal itself is a Drupal 7 website, based on Fedict's Openfed distribution plus two extra modules RestWS and RestWS i18n.
Only interested in the result ? The N-Triples and XML files (DCAT-AP) used to update data.gov.be can be found in the dcat repository
These tools can be used with a Java runtime 11 or newer, on a headless machine, i.e. there is no fancy GUI.
Internet connection is obviously required, although a proxy can be used.
- Helper classes: for storing scraped pages locally, conversion tools etc.
- Various scrapers: getting metadata from various repositories and websites, and turning the metadata into DCAT files
- Also part of the scrapers are a series of SPARQL scripts to turn DCAT into DCAT-AP: e.g. map site-specific themes, add missing properties and prepare the files for updating data.gov.be
- Data.gov.be updater: update the data.gov.be (currently Drupal 7) website using the enhanced DCAT files
- Some tools: link checker, EDP converter tool
There is also separate, stand-alone RDF validator project which can be used to validate DCAT metadata, regardless if the metadata is to be published on data.gov.be or not.
- The various portals (except
all) should be harvested using the scrapers. - The enhanced files can be uploaded to the data.gov.be portal using the updater
- Then use
allenhancer to merge all the files from the various portals into one filedatagovbe.nt - Convert the merged file using the EDP tool to an XML file called
datagovbe_edp.xml - Upload both the
datagovbe.ntanddatagovbe_edp.xmlto github - This will be used as input for the European Data Portal (scheduled Thursday morning, every week)
All configuration is done using Java (plain text) properties files. Some examples can be found in dist/cfg
See also the Notes