From 45c0edc632c667766ae8640a4715e4eb64e447d0 Mon Sep 17 00:00:00 2001 From: chenzihong-gavin Date: Wed, 14 Jan 2026 16:56:31 +0800 Subject: [PATCH 1/2] fix: ignore missing nodes --- .../models/kg_builder/light_rag_kg_builder.py | 28 +++++++++++-------- 1 file changed, 16 insertions(+), 12 deletions(-) diff --git a/graphgen/models/kg_builder/light_rag_kg_builder.py b/graphgen/models/kg_builder/light_rag_kg_builder.py index 53fe1d66..642d3398 100644 --- a/graphgen/models/kg_builder/light_rag_kg_builder.py +++ b/graphgen/models/kg_builder/light_rag_kg_builder.py @@ -164,18 +164,22 @@ async def merge_edges( set([dp["source_id"] for dp in edge_data] + source_ids) ) - for insert_id in [src_id, tgt_id]: - if not kg_instance.has_node(insert_id): - kg_instance.upsert_node( - insert_id, - node_data={ - "entity_type": "UNKNOWN", - "entity_name": insert_id, - "description": "", - "source_id": source_id, - "length": self.tokenizer.count_tokens(description), - }, - ) + if not kg_instance.has_node(src_id) or not kg_instance.has_node(tgt_id): + logger.warning("Edge (%s, %s) has missing nodes.", src_id, tgt_id) + return {} + + # for insert_id in [src_id, tgt_id]: + # if not kg_instance.has_node(insert_id): + # kg_instance.upsert_node( + # insert_id, + # node_data={ + # "entity_type": "UNKNOWN", + # "entity_name": insert_id, + # "description": "", + # "source_id": source_id, + # "length": self.tokenizer.count_tokens(description), + # }, + # ) description = await self._handle_kg_summary( f"({src_id}, {tgt_id})", description From 72df736abfd7576bef214d8e26d22bfdf51eb783 Mon Sep 17 00:00:00 2001 From: chenzihong-gavin Date: Wed, 14 Jan 2026 16:57:20 +0800 Subject: [PATCH 2/2] fix: delete useless code --- graphgen/models/kg_builder/light_rag_kg_builder.py | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/graphgen/models/kg_builder/light_rag_kg_builder.py b/graphgen/models/kg_builder/light_rag_kg_builder.py index 642d3398..460dcea0 100644 --- a/graphgen/models/kg_builder/light_rag_kg_builder.py +++ b/graphgen/models/kg_builder/light_rag_kg_builder.py @@ -168,19 +168,6 @@ async def merge_edges( logger.warning("Edge (%s, %s) has missing nodes.", src_id, tgt_id) return {} - # for insert_id in [src_id, tgt_id]: - # if not kg_instance.has_node(insert_id): - # kg_instance.upsert_node( - # insert_id, - # node_data={ - # "entity_type": "UNKNOWN", - # "entity_name": insert_id, - # "description": "", - # "source_id": source_id, - # "length": self.tokenizer.count_tokens(description), - # }, - # ) - description = await self._handle_kg_summary( f"({src_id}, {tgt_id})", description )