release: v0.6.0 — 자동 온톨로지 + 하이브리드 검색 + 벤치마크

## v0.6.0 주요 변경

### 새 기능
- 자동 온톨로지 구축 3단계
  - RuleBasedClassifier: 키워드 기반 NodeKind 자동 분류 (zero-dep)
  - RuleBasedRelationDetector: 역인덱스 + 임베딩 유사도 자동 Edge
  - LLMClassifier: few-shot 프롬프트로 kind + tags + search_keywords 생성 (86% 정확도)
  - LLMRelationDetector: LLM 기반 의미적 관계 추출
- LLMProvider: Ollama/OpenAI 지원, JSON structured output
- FTS+embedding 하이브리드 점수 (alpha×FTS + (1-alpha)×vector + cosine similarity)

### 검색 개선
- spreading activation 감쇠 0.5→0.25, depth 2→1
- ResonanceWeights relevance 0.4→0.55
- agent_search kind 필터 fallback
- kind-intent 부스트, tag-query 부스트
- _search_keywords FTS 매칭

### 벤치마크
- 외부 데이터셋 8종 (HotPotQA, Allganize, Ko-StrategyQA 등)
- 7단계 Ablation Study
- S7 Auto+Embed: MRR 0.83, R@10 1.00 (Allganize-ko)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: SonAIengine

pyproject.toml

@@ -4,18 +4,20 @@ build-backend = "hatchling.build"
 
 [project]
 name = "synaptic-memory"
-version = "0.5.0"
-description = "Brain-inspired knowledge graph for LLM agents — ontology, Hebbian learning, memory consolidation, intent-based search."
+version = "0.6.0"
+description = "Brain-inspired knowledge graph for LLM agents — auto-ontology (rule/embedding/LLM), Hebbian learning, memory consolidation, hybrid search."
 license = "MIT"
 requires-python = ">=3.12"
 readme = "README.md"
 authors = [{ name = "Son Seongjun" }]
 keywords = [
     "knowledge-graph",
     "agent-memory",
+    "auto-ontology",
     "hebbian-learning",
     "spreading-activation",
     "memory-consolidation",
+    "hybrid-search",
     "ontology",
     "mcp",
     "neo4j",

src/synaptic/__init__.py

@@ -36,7 +36,7 @@
 )
 from synaptic.resonance import ResonanceWeights
 
-__version__ = "0.5.0"
+__version__ = "0.6.0"
 
 __all__ = [
     "ActivatedNode",