Exclude next/previous contents from indexes

[benbalter]

A follow up to #138, #138 removed the content fields from documents when retrieving an index... but we still included the content of the next/previous document, and did so recursively.

So while the document itself had no content, for each document we loaded, we loaded up to four document's contents in the next/previous fields.

This PR updates the to_api logic to not include the content of nor the next/previous document for other documents when generating a given document's next/previous fields.

Before

{
  "path": "whatever",
  "next" {
    "path": "something",
    "content": "foo",
    "next": { 
     ...
    } 
    ...
  }
}

After

{
  "path": "whatever",
  "next" {
    "path": "something"
    ...
  }
  ...
}

I don't believe next/previous was used by the front end, and thus no front end change should be necessary. This does however, make document lists load significantly faster on large sites.

[mertkahyaoglu]

👍🏻