tree(ics23): add bounding path calculation proptests

[erwanor]

This PR adds a proptest setup that exercises the bounding path logic used to generate ics23 non-existence proofs.

Context
In #100, we fixed a bug located deep in the logic determining the adjacency of two subtrees in the jmt. As part of review, it was pointed out that writing property tests that exercise the different "types" of nonexistence setups/adjacency would have high ROI. As a result, we filed #104 and outlined a testing strategy in this comment.

Proptest strategy

/// Assert that the Ics23 bonding path calculations are correct.
/// To achieve this, the uses a strategy that consists in:
/// 1. generating a sorted vector of key preimages
/// 2. instantiating a JMT over a `TransparentHasher`
///
/// The last point allow us to easily test that for a given key
/// that is *in* the JMT, we can generate two non-existent keys
/// that are "neighbor" to `k`: (k-1, k+1).
///
/// To recap, we generate a vector of sorted key <k_1, ... k_n>;
/// then, we iterate over each existing key `k_i` and compute a
///     tuple of neighbors (`k_i - 1`, `k_i + 1`) which are *not*
///     in the tree.
/// Equipped with those nonexisting neighbors, we check for exclusion
/// in the tree, and specifically assert that the generated proof contains:
/// 1. the initial key we requested (i.e. `k_i + 1` or `k_i - 1`)
/// 2. the correct left neighbor (i.e. `k_{i-1}`, or `k_{i+1}`, or none`)
/// 2. the correct right neighbor (i.e. `k_{i-1}`, or `k_{i+1}`, or none`)
/// across configurations e.g. bounding path for a leftmost or rightmost subtree.
/// More context available in #104.

[cratelyn]

magnificent!

[TalDerei]

@erwanor I'm curious, how does a proof of non-existence work in the JMT? Is it the same non-membership proof like other sparse merkle tree constructions where you build a merkle proof that the leaf at specific index is empty?

referencing https://hackmd.io/@aztec-network/ryJ8wxfKK#Sparse-Merkle-Trees-and-Non-Membership-Proofs