code cleans. (#9916)

### What problem does this PR solve?



### Type of change

- [x] Refactoring
- [x] Performance Improvement

Signed-off-by: zhanluxianshen <zhanluxianshen@163.com>
This commit is contained in:
湛露先生
2025-09-05 09:59:27 +08:00
committed by GitHub
parent ddaed541ff
commit b14052e5a2
4 changed files with 7 additions and 9 deletions

View File

@ -21,7 +21,7 @@ class NodeEmbeddings:
embeddings: np.ndarray
def embed_nod2vec(
def embed_node2vec(
graph: nx.Graph | nx.DiGraph,
dimensions: int = 1536,
num_walks: int = 10,
@ -44,13 +44,13 @@ def embed_nod2vec(
return NodeEmbeddings(embeddings=lcc_tensors[0], nodes=lcc_tensors[1])
def run(graph: nx.Graph, args: dict[str, Any]) -> NodeEmbeddings:
def run(graph: nx.Graph, args: dict[str, Any]) -> dict:
"""Run method definition."""
if args.get("use_lcc", True):
graph = stable_largest_connected_component(graph)
# create graph embedding using node2vec
embeddings = embed_nod2vec(
embeddings = embed_node2vec(
graph=graph,
dimensions=args.get("dimensions", 1536),
num_walks=args.get("num_walks", 10),

View File

@ -23,7 +23,7 @@ import trio
from api.utils import get_uuid
from graphrag.query_analyze_prompt import PROMPTS
from graphrag.utils import get_entity_type2sampels, get_llm_cache, set_llm_cache, get_relation
from graphrag.utils import get_entity_type2samples, get_llm_cache, set_llm_cache, get_relation
from rag.utils import num_tokens_from_string, get_float
from rag.utils.doc_store_conn import OrderByExpr
@ -42,7 +42,7 @@ class KGSearch(Dealer):
return response
def query_rewrite(self, llm, question, idxnms, kb_ids):
ty2ents = trio.run(lambda: get_entity_type2sampels(idxnms, kb_ids))
ty2ents = trio.run(lambda: get_entity_type2samples(idxnms, kb_ids))
hint_prompt = PROMPTS["minirag_query2kwd"].format(query=question,
TYPE_POOL=json.dumps(ty2ents, ensure_ascii=False, indent=2))
result = self._chat(llm, hint_prompt, [{"role": "user", "content": "Output:"}], {})

View File

@ -561,7 +561,7 @@ def merge_tuples(list1, list2):
return result
async def get_entity_type2sampels(idxnms, kb_ids: list):
async def get_entity_type2samples(idxnms, kb_ids: list):
es_res = await trio.to_thread.run_sync(lambda: settings.retrievaler.search({"knowledge_graph_kwd": "ty2ents", "kb_id": kb_ids, "size": 10000, "fields": ["content_with_weight"]}, idxnms, kb_ids))
res = defaultdict(list)