Allow extraction methods to run independently

[rhysrevans3]

Could we allow extraction methods to be run independently/asynchronously this would allow asset/items/collection to be generated by the faster methods (regex) and then updated by slower methods (header) when ready.

Issues:

• post extraction methods (vocabs)
  • could run after every extraction method
  • could be triggered when all extraction methods have finished
• id extraction methods
  • what happens when id facets aren't yet present
    • allow for bigger items that will split as the relevant metadata is added
    • wait for all id facets before generating the item
• collection descriptions
  • the generator spits up the extraction methods
  • we split the extraction methods into separate collection descriptions (pointing at the same collection)
  • we split the extraction methods in the collection description
• race conditions
  • how do we know if the asset/item/collection already exists (do we care?)
  • elasticsearch concurrency control