This commit is contained in:
wangmeihua 2025-08-12 18:26:28 +08:00
parent 821bac3d37
commit de76d6a261

View File

@ -1,5 +1,3 @@
from urllib.request import Request
from appPublic.timeUtils import curDateString from appPublic.timeUtils import curDateString
from filemgr.filemgr import FileMgr from filemgr.filemgr import FileMgr
from rag.uapi_service import APIService from rag.uapi_service import APIService
@ -191,7 +189,7 @@ where a.orgid = b.orgid
for i in range(0, len(texts), 10): # 每次处理 10 个文本块 for i in range(0, len(texts), 10): # 每次处理 10 个文本块
batch_texts = texts[i:i + 10] batch_texts = texts[i:i + 10]
batch_embeddings = await api_service.get_embeddings( batch_embeddings = await api_service.get_embeddings(
request=Request, request=request,
texts=batch_texts, texts=batch_texts,
upappid=service_params['embedding'], upappid=service_params['embedding'],
apiname="BAAI/bge-m3", apiname="BAAI/bge-m3",
@ -246,7 +244,7 @@ where a.orgid = b.orgid
debug(f"处理 {len(chunk_texts)} 个分片进行三元组抽取") debug(f"处理 {len(chunk_texts)} 个分片进行三元组抽取")
tasks = [ tasks = [
api_service.extract_triples( api_service.extract_triples(
request=Request, request=request,
text=chunk, text=chunk,
upappid=service_params['triples'], upappid=service_params['triples'],
apiname="Babelscape/mrebel-large", # 固定 apiname apiname="Babelscape/mrebel-large", # 固定 apiname
@ -289,7 +287,7 @@ where a.orgid = b.orgid
for i in range(0, len(unique_triples), 30): # 每次插入 30 个三元组 for i in range(0, len(unique_triples), 30): # 每次插入 30 个三元组
batch_triples = unique_triples[i:i + 30] batch_triples = unique_triples[i:i + 30]
neo4j_result = await api_service.neo4j_insert_triples( neo4j_result = await api_service.neo4j_insert_triples(
request=Request, request=request,
triples=batch_triples, triples=batch_triples,
document_id=id, document_id=id,
knowledge_base_id=fiid, knowledge_base_id=fiid,