fix cache bug

lightrag/__init__.py

@@ -1,5 +1,5 @@
 from .lightrag import LightRAG as LightRAG, QueryParam as QueryParam
 
-__version__ = "1.0.0"
+__version__ = "1.0.1"
 __author__ = "Zirui Guo"
 __url__ = "https://github.com/HKUDS/LightRAG"

lightrag/operate.py

@@ -630,10 +630,16 @@ async def _find_most_related_edges_from_entities(
     all_related_edges = await asyncio.gather(
         *[knowledge_graph_inst.get_node_edges(dp["entity_name"]) for dp in node_datas]
     )
-    all_edges = set()
+    all_edges = []
+    seen = set()
+
     for this_edges in all_related_edges:
-        all_edges.update([tuple(sorted(e)) for e in this_edges])
+        for e in this_edges:
-    all_edges = list(all_edges)
+            sorted_edge = tuple(sorted(e))
+            if sorted_edge not in seen:
+                seen.add(sorted_edge)
+                all_edges.append(sorted_edge)
+
     all_edges_pack = await asyncio.gather(
         *[knowledge_graph_inst.get_edge(e[0], e[1]) for e in all_edges]
     )
@@ -833,10 +839,16 @@ async def _find_most_related_entities_from_relationships(
     query_param: QueryParam,
     knowledge_graph_inst: BaseGraphStorage,
 ):
-    entity_names = set()
+    entity_names = []
+    seen = set()
+
     for e in edge_datas:
-        entity_names.add(e["src_id"])
+        if e["src_id"] not in seen:
-        entity_names.add(e["tgt_id"])
+            entity_names.append(e["src_id"])
+            seen.add(e["src_id"])
+        if e["tgt_id"] not in seen:
+            entity_names.append(e["tgt_id"])
+            seen.add(e["tgt_id"])
 
     node_datas = await asyncio.gather(
         *[knowledge_graph_inst.get_node(entity_name) for entity_name in entity_names]

lightrag/utils.py

@@ -274,13 +274,19 @@ def process_combine_contexts(hl, ll):
     if list_ll:
         list_ll = [",".join(item[1:]) for item in list_ll if item]
 
-    combined_sources_set = set(filter(None, list_hl + list_ll))
+    combined_sources = []
+    seen = set()
 
-    combined_sources = [",\t".join(header)]
+    for item in list_hl + list_ll:
+        if item and item not in seen:
+            combined_sources.append(item)
+            seen.add(item)
 
-    for i, item in enumerate(combined_sources_set, start=1):
+    combined_sources_result = [",\t".join(header)]
-        combined_sources.append(f"{i},\t{item}")
 
-    combined_sources = "\n".join(combined_sources)
+    for i, item in enumerate(combined_sources, start=1):
+        combined_sources_result.append(f"{i},\t{item}")
 
-    return combined_sources
+    combined_sources_result = "\n".join(combined_sources_result)
+
+    return combined_sources_result