검색 파이프라인
- 입력 쿼리를 필요 시 임베딩 후, 엣지/노드/에피소드/커뮤니티 각 레이어에 대해
- 검색: BM25(풀텍스트), 코사인 유사도(벡터), BFS
- 리랭킹: RRF, MMR, Cross-Encoder, Node Distance, Episode Mentions 를 병렬 실행하여 통합 결과를 만듭니다.
graphiti_core/search/search.py(68:86)
async def search(
clients: GraphitiClients,
query: str,
group_ids: list[str] | None,
config: SearchConfig,
search_filter: SearchFilters,
center_node_uuid: str | None = None,
bfs_origin_node_uuids: list[str] | None = None,
query_vector: list[float] | None = None,
) -> SearchResults:
start = time()
driver = clients.driver
embedder = clients.embedder
cross_encoder = clients.cross_encoder
if query.strip() == '':
return SearchResults()graphiti_core/search/search.py(110:129)
# if group_ids is empty, set it to None
group_ids = group_ids if group_ids and group_ids != [''] else None
(
(edges, edge_reranker_scores),
(nodes, node_reranker_scores),
(episodes, episode_reranker_scores),
(communities, community_reranker_scores),
) = await semaphore_gather(
edge_search(...),
node_search(...),
episode_search(...),
community_search(...),
)Graphiti.search는 간단 레시피(엣지 중심 하이브리드 + RRF 또는 Node Distance)로 엣지 리스트를 반환.Graphiti.search_는 고급 설정(SearchConfig)을 받아SearchResults(노드/엣지/에피소드/커뮤니티+점수)를 반환.
graphiti_core/graphiti.py(910:967)
async def search(
self,
query: str,
center_node_uuid: str | None = None,
group_ids: list[str] | None = None,
num_results=DEFAULT_SEARCH_LIMIT,
search_filter: SearchFilters | None = None,
) -> list[EntityEdge]:
...
edges = (
await search(
self.clients,
query,
group_ids,
search_config,
search_filter if search_filter is not None else SearchFilters(),
center_node_uuid,
)
).edges
return edges