gmakstutis/LightRAG
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add comments explaining chunk deduplication behavior in query context

Browse source
This commit is contained in:
yangdx 2025-08-15 02:19:01 +08:00
parent 0b45d463df
commit 3acb32f547
1 changed files with 3 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
lightrag/operate.py
View file

@ -2354,6 +2354,7 @@ async def _build_query_context(
seen_edges.add(pair) seen_edges.add(pair)
# Get text chunks based on final filtered data # Get text chunks based on final filtered data
# To preserve the influence of entity order, entiy-based chunks should not be deduplcicated by vector_chunks
if final_node_datas: if final_node_datas:
entity_chunks = await _find_related_text_unit_from_entities( entity_chunks = await _find_related_text_unit_from_entities(
final_node_datas, final_node_datas,
@ -2365,6 +2366,8 @@ async def _build_query_context(
chunk_tracking=chunk_tracking, chunk_tracking=chunk_tracking,
) )
# Find deduplcicated chunks from edge
# Deduplication cause chunks solely relation-based to be prioritized and sent to the LLM when re-ranking is disabled
if final_edge_datas: if final_edge_datas:
relation_chunks = await _find_related_text_unit_from_relations( relation_chunks = await _find_related_text_unit_from_relations(
final_edge_datas, final_edge_datas,