Add an idempotent, per-dataset migrator that synthesizes a Pandera DataFrameSchema from each v1 Dataset's Field rows, uploads it as the first schema version, and copies records (deriving reference), questions (mapped to column bindings), responses, and suggestions into v2, building the LanceDB table. Once a workspace's datasets are migrated and verified, delete the corresponding v1 tables/handlers and work through the server/frontend retirement ledger — back-compat is explicitly not a goal, so v1 exists only as a migration source. See spec §8–9.
Part of #218.
Add an idempotent, per-dataset migrator that synthesizes a Pandera
DataFrameSchemafrom each v1Dataset'sFieldrows, uploads it as the first schema version, and copies records (derivingreference), questions (mapped to column bindings), responses, and suggestions into v2, building the LanceDB table. Once a workspace's datasets are migrated and verified, delete the corresponding v1 tables/handlers and work through the server/frontend retirement ledger — back-compat is explicitly not a goal, so v1 exists only as a migration source. See spec §8–9.Part of #218.