Add rpc wallet example crates

[rajarshimaitra]

Description

This PR adds the RPC wallet syncing codes. This includes the rpcwallet library, keychain_tracker_rpc command line app, and wallet_rpc BDK wallet example.

This works on the existing structure of bdk_chain and might require update after #895.

Notes to the reviewers

This still has many rough ages.

• The RpcClient currently initiates a legacy wallet, into which we import all the tracker addresses. This still causes the problem reported in RPC sync failure  #859. The work-around for now added in the code is to ignore the resource unavailable error and carry on. Looking for more suggestions. One future hope is, using the core's descriptor wallet and work with descriptor imports. But still might not fix this RPC issue.

• I added a new Wallet API dump_spks which is useful for getting all the existing spks from the tracker without rederiving them.

• Basic docs are added. If you see anything missing or inappropriate let me know.

• What should be the approach of testing? Should we think of recreating the previous blockchian tests?

Changelog notice

• Add the new rpc-wallet syncing examples.

Checklists

All Submissions:

• I've signed all my commits
• I followed the contribution guidelines
• I ran cargo fmt and cargo clippy before committing

[notmandatory]

I seem to recall a discussion about syncing a new 1.0 wallet with RPC using full blocks, or I could be conflating with a different discussion. Either way.. in theory would it be possible to sync a bdk 1.0 wallet by scanning all the blocks from the birth date of the wallet, and then updating with new blocks every time the chain tip changes? This question came up in a discussion I was having with @bodymindarts about improving sync performance for very large (200K+) address wallets.

[evanlinjin]

@notmandatory That would be a separate example!

[evanlinjin]

Closing as it is stale and is based on codebase from a while back. In the future we may want this, however we had no requests so far, and I would assume block-by-block would be the way to go.