Search System Technology
From Dryad wiki
The Dryad search system is based on the DSpace Discovery system.
SOLR indexes
- authority -- terms for autocompletion using controlled vocabularies, including HIVE
- dataoneMNlog -- log of accesses through the DataONE API
- dryad -- local storage of DOIs (this index needs to be renamed)
- search -- primary search index
- statistics -- log of accesses to item pages and bitstream downloads
Configuration
- dspace.cfg -- specifies the URLs for the various solr indexes, specifies fields that are used within the search system
- dspace-solr-search.cfg -- specifies the parameters that are added to queries, both automatic queries and user-generated queries
- solr/(index name)/conf -- specifies the fields stored in an index, and how those fields are processed
Maintenance
Primary maintenance of the index is performed with:
/opt/dryad/bin/dspace update-discovery-index
Options available:
- -i <internal_item_id>
- re-index a single item
- -f
- forced every item to be re-indexed, even if it is up-to-date
- -b
- rebuild by dropping the index and reindexing everything
- -r <handle>
- remove an item from the index
- -o
- optimize the solr indexes on disk
For convenience, the dryad-utils package has a script that will update the index for items that were archived during a particular date range.
dryad-utils/reindex-discovery.py --date_from 2017-01-01 --date_to 2017-01-31