jac-sv-persistence

Use either a jid loop over [root-->][?:Type] comparing jid(node) to the target id, or jobj(id) with an isinstance check for O(1) lookup. jobj is required when the node lives under another user's root, and Python id() must never be used since it changes every restart.

Append bracket predicates to the traversal, such as [root -->][?:Post][?published] for boolean fields or [?author == "alice"] for value filters. Use len() for aggregation since there is no count() form.

Yes. Data persists to SQLite in .jac/data/ by default, and setting MONGODB_URI switches to MongoDB with a Redis L2 cache. The same node model and schema migration rules run identically on both backends.

Without a declared schema_alias, a rename looks like remove-plus-add: old values move to the recoverable attic and the new field reads its default. Declaring schema_alias inside __jac_schema__ flows the old value into the renamed field on next load.

This happens when stale anchors persisted by a previous run under a different schema are loaded. In development, stop the server, delete .jac/data/, and restart; in production, use the alias and quarantine machinery with jac db commands instead.

Use root traversal for listing and filtering nodes reachable from the current user's root. Use jobj when you have a specific id, need O(1) access, or the target node was granted from another user's root, but remember jobj never authorizes so you must enforce grants separately.