feat: read-requests generated by app circuit, checked in private kernel

[dbanks12]

Description

When the app circuit reads a note, it should output a read request for that note. The RPC client will then generate membership witness (leaf index & sibling path) for that read request and pass that to the private kernel. The private kernel will silo the read request by contract address and perform a membership check to ensure that the read request corresponds to an existing commitment in the private data tree. All read requests must hash to matching private data roots. App circuit no longer needs to get any membership info for notes when it makes a get oracle call.

Outlined in these discourse threads:

• https://discourse.aztec.network/t/to-read-or-not-to-read/178
• https://discourse.aztec.network/t/proposal-spending-notes-which-havent-yet-been-inserted/180

Blocks #512
Closes #513

Followup tasks:

• SCARY BUG UNCOVERED: bug: simulator oracle getNotes' leaf index does not match private data tree #788
• Remove historic private data root from app circuit public inputs - add to PrivateCallData #778
• Spike: what if app circuit outputs different #read-requests vs what the RPC client gets from the simulator? #779
• bug: get_2 in set.nr only pushes one read_request even when reading two notes #780
• Spending pending commitments [tracking ticket] #512

Checklist:

• I have reviewed my diff in github, line by line.
• Every change is related to the PR description.
• I have linked this pull request to the issue(s) that it resolves.
• There are no unexpected formatting changes, superfluous debug logs, or commented-out code.
• The branch has been merged or rebased against the head of its merge target.
• I'm happy for the PR to be merged at the reviewer's next convenience.

[Maddiaa0]

looks pretty good to me! just some comments, all nits

[suyash67]

Circuits/cpp changes look correct to me. Just added minor comments.

[jeanmon]

LGTM. My comments are very minor.