adding unique internal identifier to records, `combine_id`

[ghukill]

Related to #122, something that will help address the problem of identifiers will be slightly changing how they are handled internally. This issue outlines the process of adding a new column combine_id, a randomly generated UUID hash at a record's creation.

It will keep the record_id as a string that is semantically meaningful, and will be used for publishing. What this allows for, is altering the record_id if desirable, perhaps even a dedicated task in Combine.

Current progress:

Done

• adding combine_id to Record model
• updated combine_tables_prime.sql
  • BUILD: need to update this?
  • both tables
• have linkage between records look for combine_id not record_id
• make sure ES indexes with combine_id, and not record_id
• having Spark jobs write random uuid to combine_id
  • still writing record_id as before
  • using withColumn likely...
• ES doc retrieval
• use record_id for OAI publishing

ToDo

• create workflows for updating record_id
  • adding as "tack on" like validation scenarios and index mapping for jobs
    • with live preview as well (consider moving test_record_id_transform() view to models?)
• find and address where Records were selected or disambiguated with record_id and change to combine_id if it makes sense
• add combine_id to search in tables?

*moving considerations about uniqueness to new issue: #125

[ghukill]

Done!