Pipeline for EAGER

[marcverhagen]

Probably create a new repository for this. Current EAGER index has entire text of Pubmed central, but scoring focusses on abstract and title.

Tasks:

• [ ] Pipeline documentation should not depend on bash code.
• [ ] Keigh to run Stanford dependencies and NER on all data.
• [ ] Marc to run TTK (Evita only), Techwatch and Topic model
• [ ] Extract a limited set of relations from dependencies

[marcverhagen]

New repository created at https://github.com/lappsgrid-incubator/eager-elk.