chore: migrate GRACE-Poly anchors to hierarchical dotted naming

Systematic rename of all semantic anchors (#region, [DEF], @RELATION)
across 1400+ files — backend Python, frontend Svelte/TS, specs, docs:
- Flat anchors become Namespace.Module.Entity
- @RELATION references updated to match new anchor paths
- Zero business logic changes

docs/adr/ADR-0018-rejected-dataset-review.md

@@ -1,8 +1,8 @@
-# [DEF:ADR-0018:ADR]
+# [DEF:Doc.Adr.ADR0018:ADR]
 # @STATUS REJECTED
 # @PURPOSE Record the permanent rejection of the Dataset Review feature and prohibit restoring its workflow, UI, APIs, session model, or assistant commands without a replacement ADR.
-# @RELATION DEPENDS_ON -> [ADR-0006]
-# @RELATION DEPENDS_ON -> [ADR-0008]
+# @RELATION DEPENDS_ON -> [Doc.Adr.ADR0006]
+# @RELATION DEPENDS_ON -> [Doc.Adr.ADR0008]
 # @RATIONALE Dataset Review was a cross-cutting orchestration proposal spanning persistent review sessions, semantic enrichment, guided assistant interaction, SQL preview, and SQL Lab launch. Its large coupled surface would preserve unsupported product behavior and maintenance obligations after the feature has been removed from the active product.
 # @RATIONALE The existing `specs/027-dataset-llm-orchestration/` artifacts remain useful as historical design evidence. They are retained in place and explicitly deactivated rather than deleted, so prior decisions and scope can be audited without implying that the feature is active.
 # @REJECTED Restoring Dataset Review from its retained spec, routes, components, APIs, stores, session entities, or assistant commands without a new ADR and a newly accepted specification - rejected because the retained materials are historical only and do not authorize implementation.
@@ -64,4 +64,4 @@ this decision.
 - ADR-0006 - Frontend Architecture
 - ADR-0008 - Assistant Tool Registry
 
-# [/DEF:ADR-0018:ADR]
+# [/DEF:Doc.Adr.ADR0018:ADR]