DESY workflows

[fschwenn]

This wrapping issue contains a list of issues which are needed to replace the workflows done at DESY outside of Invenio and which are not yet issued elsewhere (like harvesting or arXiv selection)

• journal harvesting, enrichment of records for the selection process in the holdingpen, (roughly) ordered by priority

  1. • [x] holdingpen: enrichment: fully taken journals #2643
  2. • [ ] holdingpen: enrichment: number of (core) references in INSPIRE #2644
  3. • [x] holdingpen: enrichment: run bibclassify on abstract (if not fulltext available) #2645 ---------- show-stopper threshold (KS) -------------------------
  4. • [ ] holdingpen: enrichment: number of core papers by authors in INSPIRE #2646
  5. • [ ] holdingpen: enrichment: run bibclassify with anti-HEP-ontology on abstract #2647
  6. • [ ] holdingpen: enrichment: translate PACS and count CORE PACS #2648
  7. • [ ] holdingpen: enrichment: guess field code from authors #2649
  8. • [ ] holdingpen: enrichment: guess field code from journal #2650
  9. • [ ] holdingpen: enrichment: number of core institutes #2651

  If these issues are implemented, a lot of numbers indicating the relevance are generated. These numbers shall not end up in individual proposals (whether to take or to reject something) in addition to the existing one. Instead, there shall be exactly one function making the proposal which takes the existing machine-learning-number and the new indicators to calculate one final indicator out of it.

• miscellaneous

  1. • [x] Inspirehep JSON merger: GUI #2696
  2. • [ ] holdingpen(?): workflow: manual enrichment #2657
  3. • [ ] holdingpen: workflow - general layout discussion #2658

• keywording by external experts.

  1. • [ ] expert keywording: assign papers #2652
  2. • [ ] expert keywording: send work to expert #2653
  3. • [ ] expert keywording: expert assigns keywords #2654
  4. • [ ] expert keywording: book keeping #2655

[ksachs]

If batchupload is not available any more we need integrated: manual enrichment #2657, some way to get keywords in (#2654) and do bookkeeping (#2655). This would be a show-stopper.