status: brainstorm

Institutional Repository Importer
=================================

Want to import content from IRs. Same general workflow for CORE, SHARE, BASE,
other aggregators.

Filter input to only works with known/ingested fulltext.

Lookup file by hash. If found, skip for now. In future might do
mapping/matching.

Lookup by primary id (eg, CORE ident). If existing, can skip if it has file, or
add file/location directly.

Two indirect lookups: by external ident (DOI, PMID), or fuzzy search match. If
we get either of these, want to do release/work grouping correctly.

1. if we are certain of IR copy stage, then compare with existing release,
   and/or lookup entire work for releases with same stage. update release or
   add new release under same work.
2. not sure of IR copy stage. guess stage from sherpa/romeo color and proceed
   to insert/update.