Clone records recursively before caching

[tokebe]

(fixes biothings/biothings_explorer#234)

This change replaces current cache_handler._copyRecord behavior to recursively clone each record, 'freezing' values from getters into standard data properties (also recursively -- if a getter returns an object, that object is recursively cloned).

Unlike #25, this change is slightly more future-proof in that no values to be saved are hard-coded. This may result in some bloating of the data being cached, however.

The following example query returns kg_nodes with biolink:undefined categories, but should work as expected with this change.

{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "categories": [
            "biolink:Disease"
          ]
        },
        "n1": {
          "ids": [
            "HGNC:6284"
          ],
          "categories": [
            "biolink:Gene"
          ]
        }
      },
      "edges": {
        "e0": {
          "subject": "n0",
          "object": "n1"
        }
      }
    }
  }
}

[andrewsu]

I was able to reproduce the issue locally and confirm that this PR fixes it for two additional queries (aside from the one in the first PR comment). Also confirmed that the initial response and the cached response are identical (except for the logs).