chore: address followup comments for node update & re-export NetworkNotePricer

[juan518munoz]

Closes #2473

1. CHANGELOG fix: The five entries that pointed at issue Update node to 0.16 #2360/chore: update node to 0.16 #2431 now cite PR chore: update node to 0.16 #2431. The GetAccount behavior entry was dropped per your earlier decision.
2. Pass SyncedNote to observer: NoteObserver::observe now takes &SyncedNote.
3. Flatten SyncedNote: flattened to note_id, metadata, inclusion_proof, details, attachments.
4. "Is the first check covered by the second?": the validation now verifies set equality: every requested key must appear in the response, and the response must not contain unrequested keys.
5. Attachment parsing hard to follow: simplified using the ReportedAttachment enum and ReportedAttachments::from_reports.
6. Reintroduce ResolvedNoteContent: superseded by the flatten of point 3.
7. Hardcode insecure key: the runtime keygen in scripts/start-test-node.sh was replaced by three hardcoded insecure test-only keys.

Closes #2517

1. Re-export NetworkNotePricer

[igamigo]

Looks good, but I think there might be a regression here. In any case I think this warrants a bigger refactor (#2118) so maybe we fix the issue, merge this anyway and improve the whole setup afterward. I believe there are a couple of structs we could get rid of.

[igamigo]

There does not seem to be any entry here for PR 2475

[juan518munoz]

Added entries in 20dc251

[igamigo]

I think this effectively removes the attachments which means the user does not have that information during the callback

[juan518munoz]

The screener's record is now built with SyncedNote::to_committed_note() instead of CommittedNote::new(), so it carries the resolved attachments (an empty set for a note without any) rather than None.

20dc251