The MVP diff function isn't cutting it. We need something that's easier to work with and knows more about the structure of the JSON it's looking at.

A new diff function will generate a diff based on just a unique ID (configurable: page title or ID) for each result and display everything else subordinate to that.

Desired features:

• detail is collapsible, so it is possible to compare just titles & IDs
• if a result is in both lists, perform a structure-savvy diff on their sub-parts
• make it easy to see the same item in both lists, even if they are fairly far apart in the list (e.g., 1st and 15th)

It should also be faster than the current diff, which is terribly slow on large result sets (e.g., 50 results per query; 100 is untenable for large query sets).