Function observeSelectDeep

Deeply observe entity data for an archetype query.

⚠️ NOT FOR PRODUCTION USE. This has poor runtime performance by construction: it attaches a separate per-entity observer to every matched entity, rebuilds a fresh data object on every change, and deep-equals compares to dedupe. Cost scales with the size of the result set times the churn rate. It exists for debugging and ad-hoc inspection only — reach for db.observe.select (entity ids + membership changes) and read components on demand in real code.

The result rows include id deliberately: unlike read(entity) (where you already hold the id), these are otherwise-anonymous query results, so id is the only thing identifying which entity each row is.

Unlike db.observe.select (which returns entity IDs and only triggers on set membership changes), observeSelectDeep returns full typed entity data and re-emits whenever ANY included component value changes on any matched entity.

Composes observe.select for set membership with per-entity observers for value changes. Per-entity deduplication avoids re-emissions when unrelated components change on a tracked entity.

Equivalent parameters return the same cached Observe factory, and Observe.withCache shares a single underlying subscription across all simultaneous subscribers.