Refactor mhld_display

[jcoyne]

Currently we do a lot of processing to get the folio data to look like the data from SIRSI. However, this could be greatly simplified if we just index the data as it appears in folio. This would require a coordinated change to the code in Searchworks that expects mhld_display to look a certain way.

[cbeer]

I wonder if a good intermediate step would be creating an mhld_struct with some JSON in it.

{
"mhld_struct": { 
  "GREEN": {
    "STACKS": {
      "note": [],
      "holdings": [
         { "statement": "", "note": "" }
      ],
      "indexHoldings": [],
      "supplementalHoldings": [],
      "latestReceived": { "an array or a hash? not sure what data we need": "??" }
    }
  }
}
}