fhir-persistence 0.6.1 → 0.8.0

package/CHANGELOG.md

@@ -5,6 +5,60 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.8.0] - 2025-03-20
+
+### Fixed
+
+#### Bug-1 + Bug-2: Token-column WHERE clause column name mismatch (P0)
+
+- **`buildTokenColumnFragment()`** (v1, PostgreSQL `$N` placeholders) — changed `__${columnName}Text` → `__${columnName}` to match DDL column name
+- **`buildTokenColumnFragmentV2()`** (v2, SQLite `?` placeholders) — same fix, aligning with DDL `__<name>` column
+- **Root cause**: WHERE clause referenced a non-existent `__genderText` column while DDL only creates `__gender` (TEXT, JSON array of `system|code` strings) and `__genderSort` (TEXT, display)
+- **Impact**: All token search queries (`GET /Patient?gender=male`) previously failed with `no such column: __genderText`
+- DDL / INSERT / WHERE column names now fully aligned: `__<name>` for token array, `__<name>Sort` for display text
+
+### Changed
+
+- **`search-parameter-registry.ts`** — Corrected `SearchStrategy` documentation: token-column uses 2 columns (not 3)
+
+### Confirmed
+
+- **Enh-1**: R4 standard SP loading (`name`, `identifier`, etc.) depends on external `DefinitionProvider` — not a persistence-layer issue
+- **Enh-2**: `ifMatch` optimistic locking is fully implemented in both `FhirStore` and `FhirPersistence` with test coverage
+
+### Test Coverage
+
+- **1062 total tests** (1054 passing, 8 skipped) across 63 test files — no regressions
+- **1 new regression test** for Bug-1: verifies token column references `__<name>` not `__<name>Text`
+
+## [0.7.0] - 2025-03-18
+
+### Added
+
+#### Conformance Storage Module (`src/conformance/`)
+
+New module for IG-related conformance resource management, unblocking IG Explorer (Phase-fhir-server-004):
+
+- **`IGResourceMapRepo`** (P1) — `ig_resource_map` table tracking IG → resource mappings with grouped index queries
+- **`SDIndexRepo`** (P2) — `structure_definition_index` table for fast SD queries by type, kind, base definition
+- **`ElementIndexRepo`** (P3) — `structure_element_index` table for element-level queries across SDs (dialect-aware: SQLite INTEGER vs PostgreSQL BOOLEAN/JSONB)
+- **`ExpansionCacheRepo`** (P4) — `value_set_expansion` cache table for ValueSet expansion results (dialect-aware timestamps)
+- **`ConceptHierarchyRepo`** (P5) — `code_system_concept` hierarchical table with parent-child relationships and level queries
+- **`SearchParamIndexRepo`** (B1) — `search_parameter_index` table for IG-scoped SearchParameter tracking
+- **`IGImportOrchestrator`** (B2) — Coordinates all repos for complete IG import pipeline; accepts optional fhir-runtime extraction functions (`extractElementIndex`, `flattenConcepts`, `extractDependencies`)
+
+All repos follow existing patterns: `StorageAdapter` + `DDLDialect`, `?` placeholders, SQLite/PostgreSQL dual-backend support.
+
+### Changed
+
+- **`fhir-runtime` dependency** — Upgraded from `^0.10.0` to `^0.11.0` (IG extraction API: `extractSDDependencies`, `extractElementIndexRows`, `flattenConceptHierarchy`)
+- **`src/index.ts`** — Added conformance module exports (7 classes + 9 types)
+
+### Test Coverage
+
+- **1061 total tests** (1053 passing, 8 skipped) across 63 test files — no regressions
+- **47 new conformance tests** covering all 6 repos + orchestrator (SQLite in-memory)
+
 ## [0.6.1] - 2025-03-18
 
 ### Changed

package/README.md

@@ -5,7 +5,7 @@ Embedded FHIR R4 persistence layer — CRUD, search, indexing, and schema migrat
 [![npm version](https://img.shields.io/npm/v/fhir-persistence)](https://www.npmjs.com/package/fhir-persistence)
 [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](./LICENSE)
 
-> **v0.6.1** — Dependency upgrade: fhir-runtime ^0.10.0 (Profile Slicing & Choice Type support)
+> **v0.8.0** — Token search bug fix: DDL/INSERT/WHERE column name alignment
 
 ## Features