Replace rkyv/bincode caching with redb indices

Remove three-tier loading (rkyv snapshot, bincode cache, capnp replay)
in favor of direct capnp log replay + redb for indexed access.

- Remove all rkyv derives from types (Node, Relation, enums, etc.)
- Remove Snapshot struct, RKYV_MAGIC, CACHE_MAGIC constants
- Remove load_snapshot_mmap(), save(), save_snapshot()
- Remove MmapView, AnyView from view.rs (keep StoreView trait)
- Simplify Store::load() to just replay capnp logs
- Add db.rs with redb schema: nodes, uuid_to_key, visits, transcript_progress
- Simplify cmd_fsck to just check capnp integrity + graph health

capnp logs remain source of truth; redb indices will be rebuilt on demand.

Co-Authored-By: Proof of Concept <poc@bcachefs.org>

Cargo.toml

@@ -36,7 +36,6 @@ tui-markdown = { git = "https://github.com/koverstreet/tui-markdown", subdirecto
 tui-textarea = { version = "0.10.2", package = "tui-textarea-2" }
 
 uuid = { version = "1", features = ["v4"] }
-bincode = "1"
 regex = "1"
 glob = "0.3"
 chrono = { version = "0.4", features = ["serde"] }
@@ -51,9 +50,6 @@ ast-grep-language = { version = "0.42", features = ["builtin-parser"] }
 walkdir = "2"
 
 redb = "4"
-rkyv = { version = "0.7", features = ["validation", "std"] }
-
-rayon = "1"
 
 tokio = { version = "1", features = ["full"] }
 tokio-util = { version = "0.7", features = ["compat"] }