Files
Enterprise-Knowledge-Base/migrations/versioned/000052_models_managed_by.up.sql
T
wizardchen fdc22fd7a5 feat(builtin-models): add managed_by column to models table
Introduce a `managed_by` varchar column on `models` so future declarative
loaders can claim ownership of a subset of rows without disturbing entries
created via the UI/API or seeded by hand-written SQL.

- versioned/000052_models_managed_by.{up,down}.sql add the column with a
  default of '' and a partial index on non-empty values to keep startup
  reconciliation cheap.
- sqlite/000000_init.up.sql is updated in place (the Lite init migration
  is a single file per project convention) so fresh SQLite databases get
  the column from the start.
- Model.ManagedBy mirrors the column. Existing rows default to '' which
  the YAML loader treats as "manually managed, never touch".

Schema half of the YAML-driven built-in-model lifecycle work that follows
up on #1453; the reconciler that uses the column lands in the next commit.
2026-05-26 11:37:03 +08:00

38 lines
1.8 KiB
SQL

-- Migration: 000052_models_managed_by
-- Add a `managed_by` column to `models` so the YAML built-in models loader can
-- own a slice of the table without disturbing rows created via the UI/API or
-- seeded by hand-written SQL.
--
-- Background:
-- * 000051-era PR #1453 introduced config/builtin_models.yaml as a
-- declarative source of truth for built-in models. The first version
-- could add/update rows but had no way to remove them: deleting an entry
-- from YAML left the corresponding row in `models`, breaking the
-- "declarative" contract and forcing operators back into ad-hoc SQL.
-- * Indiscriminately deleting `is_builtin=true` rows on each startup is
-- unsafe because some deployments seed built-ins via direct SQL, and
-- those rows must not be touched.
--
-- Solution:
-- * Add `managed_by` (varchar(32), default ''). YAML loader writes "yaml";
-- anything else (UI / API / SQL seed) keeps the empty default.
-- * On every startup the loader UPSERTs YAML entries with managed_by='yaml'
-- and soft-deletes rows where (is_builtin=true AND managed_by='yaml' AND
-- id NOT IN <current YAML id set>). Manual rows are never inspected.
--
-- This migration is idempotent (IF NOT EXISTS) and safe to re-run.
DO $$ BEGIN RAISE NOTICE '[Migration 000052] Adding models.managed_by column'; END $$;
ALTER TABLE models
ADD COLUMN IF NOT EXISTS managed_by VARCHAR(32) NOT NULL DEFAULT '';
-- A partial index on the YAML-managed slice keeps the startup sweep cheap
-- even if the table grows large. Manual rows (managed_by='') are excluded
-- so the index stays small.
CREATE INDEX IF NOT EXISTS idx_models_managed_by_yaml
ON models (managed_by)
WHERE managed_by <> '';
DO $$ BEGIN RAISE NOTICE '[Migration 000052] Done'; END $$;