Support both Dataset.files and Dataset.ingest_metadata.files

[shirey]

Only do this after search-api #857 Remove ancetor.* and descendant.* fields has been completed

We would like to stop adding file information under the Dataset.ingest_metadata field. To this end we'll ask all groups who use this file information to start using it directly from Dataset.files. In the interim we need to make sure that

• if data comes in (PUT or POST) at Dataset.files or Dataset.ingest_metadata.files that a copy of it ends up in both places.
  • In search-api make sure the files object isn't copied to the top level any longer- BUT because we now send it from entity-api in both places, make sure that it is still in the ES indices in both places.
• make sure any existing Dataset.files or Dataset.ingest_metadata.files is retrofitted (copied) to both places for existing Datasets

NOTE: After this has been completed we will then need to:

• Ask consumers of files (Harvard, IU, other?) to start using it from Dataset.files
• Ask PSC to only send files information to Dataset.ingest_metadata.files
• Remove this patch
• Remove all Dataset.ingest_metadata.files (Neo4j and ES)

[shirey]

Won't do because this handled in #748