Page Comparison

• Julie will give us a quick run down and demo of the work that has been done around Bulkrax, a bulk import-export engine for Hyrax

SLIDES

• Bulkrax is a ruby gem for bulk import-export (DCE's Zizia is another option in this area
• Current functionality
  • CSV import (creates Work for each line)
  • CSV export
  • import from OAI-PMH
• Going forward, eyeing bagit export/import, improved UI and better documentation
• import metadata & files
• provides flexibility with field mappings → map source data to Hyrax fields; can specify some rules for data clean-up and simple parsing of source data 
  • on roadmap to integrate the metadata mapping functionality into the dashboard
• OAI-PMH importer: can schedule updates, harvest/import particular subsets of source data

DEMO

• demo running on nurax instance, added the gem & ran installer, no custom code
• Imports:
  • OAI-PMH importer, set becomes a collection
  • CSV
    • provide path to the CSV manifest
    • looks for files in a specified directory (part of the path to the manifest), expects files to be referenced in the manifest file
  • can rerun imports to update existing records
  • Looking forward, hoping to add MODS, ProQuest XML formats to import, will be soliciting community feedback for formats
  • relatively straightforward developer task to add local parsers, formats; hoping to have that better documented in near future
• Exports:
  • metadata only or with files; export from collections or an existing import (currently only works from collections?)
  • downloads a zip file that is dumped into folder (can be configured)
  • current development plans to do exports from collections, Work types, and exporting particular search results

QUESTIONS

• are errors encountered during export being tracked?
  
  • "records failed" currently showing in the interface, but would have to look at the jobs queue to know exactly what failed
    • hoping soon to be able to click through to individual imports to see a summary of errors
• can this be used to make metadata updates?
  • round-tripping of metadata should be possible soon (in the next week or two)