Create Annif corpus

[acka47]

• [ ] create a Short text document corpus (TSV file)
• [ ] document the process
• [ ] publish the corpus data (could also be useful as a test set for others who learn/teach Annif with German-language data)

Based on this, we'd be able to approach further steps in the long run:

• train Annif with the data & provide a suggest API
• use Annif for adding subjects to NWBib predecessor bibliographies
• Use/build an Annif plugin for Alma
• enable suggestions in Alma for NWBib editors

[fsteeg]

Here is what I did with NWBib data in/after the Annif workshop at SWIB19:

https://github.com/fsteeg/python-data-analysis/tree/master/annif

[acka47]

See also https://github.com/hbz/lobid-resources/issues/681