feat(snapshot, stages): snapshot stage

[shekhirin]

There's a scenario when we have data in the database for finalized..=tip blocks, turn off the node, turn it on again, and pipeline starts, writing to static files. The problem is that data in the database was never moved to static files.

This PR introduces a separate pipeline stage that copies all (even non-finalized, and it's fine because we can unwind from static files) data from the database to static files. The actual deletion will be done by pruner, as before.

reth/crates/stages/src/stages/snapshot.rs

Lines 7 to 10 in fb77c95

	/// The snapshot stage _copies_ all data from database to static files using [Snapshotter]. The
	/// block range for copying is determined by the current highest blocks contained in static files
	/// and [BlockNumReader::best_block_number],
	/// i.e. the range is `highest_snapshotted_block_number..=best_block_number`.

---

Tested on Holesky by taking a database generated with main branch, unwinding it by 1000 blocks, and starting the node on alexey/snapshot-stage branch. It triggers the pipeline, and the Snapshot stage moves all data from database to static files.

[mattsse]

this seems reasonable, pending @onbjerg

[onbjerg]

lgtm minus the comment @mattsse left