feat: hash-based JSON circuit

[Autoparallel]

THis one hurt my head

[Autoparallel]

This is based on #40 , so that should 100% be merged first.

This needs SERIOUS review and further testing.

[Autoparallel]

The basic idea is this:

We want to show that a given sequence of keys for retrieving from a JSON :

  {
     "data": {
         "items": [
             {
                 "data": "Artist",
                 "profile": {
                     "name": "Taylor Swift"
                 }
             }
         ]
     }
  }

determines a unique value inside of the JSON, and furthermore the value is as expected.

For example, we may want to ultimately supply to this circuit:

["data", "items", 0, "profile", "name", "Taylor Swift"]

---

Ultimately what I have achieved at this point is getting unique hash values at points in the JSON that take into account all the data at that point that we need.

So, for example, we know we need to get:

• "data" as a KEY at depth 1, so this can only be the case if we sequentially hash the stack [1,0] with the byte sequence [100,97,116,97] (over the next machine iterations). So this uniquely defines this key. (TODO: This does not rollover the previous hash into the next for this exact part yet, I almost have it though).
• "items" as a KEY follows the same paradigm, but we should completely exit rewrite the hasher state before we get to a new key. This is done. Hitting a comma , while not parsing a string will wipe the hasher state.
• 0 as an ARRAY INDEX also follows the same paradigm, but there is now just repeated hashing with "[2,n]at indexn. Comma characters ,` are consistently wiping the state as with the key.
• This continues down until we hit "Taylor Swift" in our example which will be a VALUE, so the only difference is a hashing with stack [1,1] at that point. Once again guaranteeing uniqueness.