feat(katana): getStorageProof endpoint

[kariy]

This PR introduces a new endpoint, starknet_getStorageProof for querying state proofs, including for historical states.

We are already computing all the state tries (ie classes, contracts, and storages) in #2609, but we don't store the historical trie nodes. Therefore, it's only possible to get the latest state proofs. To support historical tries requires, there are at least two approaches I have could thought of; (1) storing the entire trie including historical trie nodes, or (2) recomputing the trie at runtime.

Approach (2) may possibly be the simplest considering we already have historical states and fetching them is pretty easy with the provider abstractions. We also don't need to create new tables for storing the historical nodes. But computing the trie is a very expensive operation, and depending on the block for which we want to recompute the trie, if the state update is very large, it might consume a lot of resources (iterating over db entries to fetch all historical states, and then computing the trie itself).

We also have to take into account the primary consumer for this new endpoint, which is saya. It'll be primarily used to fetch the state proofs required to run SNOS for generating the proof of a block for settlement. It is understood very well that saya's proof generations will inevitably lag behind katana in block productions. So, saya could be asking katana for the state proofs of block 5 while katana is already producing block 10. To prevent overloading katana runtime performance (possibly impacting its perceived performance when executing transactions), we have opted for approach (1).

We introduced 6 new tables; ClassesTrieHistory, ContractsTrieHistory, StoragesTrieHistory, for storing the trie node at a particular block, and ClassesTrieChangeSet, ContractsTrieChangeSet, StoragesTrieChangeSet for indexing the list of blocks where a trie node has changed. These tables serve similar purposes to the already existing tables that are also suffixed by *History and *ChangeSet. This mean for each node that exist in a trie, we require 2 more entries stored in the aforementioned tables respectively.

We don't yet implement any pruning mechanism. This shall be implemented in the future.

---

The getStorageProof endpoint is also part of the RPC 0.8.0 spec. But we don't bump the RPC version as this PR doesn't include the other changes as defined in the new specification.

[codecov]

Codecov Report

Attention: Patch coverage is 46.99248% with 423 lines in your changes missing coverage. Please review.

Project coverage is 55.31%. Comparing base (8111b6d) to head (52941db).
Report is 1 commits behind head on katana/dev.

Files with missing lines	Patch %	Lines
crates/katana/storage/db/src/trie/mod.rs	42.37%	68 Missing ⚠️
...atana/storage/provider/src/providers/fork/state.rs	0.00%	66 Missing ⚠️
.../katana/storage/provider/src/providers/db/state.rs	20.98%	64 Missing ⚠️
crates/katana/storage/db/src/trie/snapshot.rs	0.00%	54 Missing ⚠️
crates/katana/executor/src/implementation/noop.rs	0.00%	26 Missing ⚠️
...na/executor/src/implementation/blockifier/state.rs	0.00%	24 Missing ⚠️
crates/katana/executor/src/abstraction/mod.rs	0.00%	22 Missing ⚠️
...ana/storage/provider/src/providers/fork/backend.rs	0.00%	21 Missing ⚠️
...s/katana/storage/db/src/abstraction/transaction.rs	39.39%	20 Missing ⚠️
crates/katana/rpc/rpc/src/starknet/mod.rs	70.45%	13 Missing ⚠️
... and 8 more

Additional details and impacted files

@@              Coverage Diff               @@
##           katana/dev    #2809      +/-   ##
==============================================
- Coverage       55.71%   55.31%   -0.41%     
==============================================
  Files             439      444       +5     
  Lines           55657    56230     +573     
==============================================
+ Hits            31010    31104      +94     
- Misses          24647    25126     +479     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.