87 lines
3.2 KiB
Python
87 lines
3.2 KiB
Python
import os
|
|
from pymilvus import (
|
|
connections,
|
|
utility,
|
|
FieldSchema,
|
|
CollectionSchema,
|
|
DataType,
|
|
Collection,
|
|
MilvusClient,
|
|
db,
|
|
Function,
|
|
FunctionType,
|
|
AnnSearchRequest,
|
|
WeightedRanker
|
|
)
|
|
|
|
class Milvus_Database:
|
|
def __init__(self,user=os.getenv("Milvus_USER","root"),password=os.getenv("Milvus_PASSWORD","Milvus"),uri="http://172.18.30.165:19530"):
|
|
self.client = MilvusClient(
|
|
uri="http://172.18.30.165:19530",
|
|
token="root:Milvus"
|
|
)
|
|
def get_text_by_resource(self, collection_name, database_name,resource,text):
|
|
"""
|
|
return List :
|
|
"""
|
|
self.client.using_database(db_name=database_name)
|
|
iterator = self.client.query_iterator(
|
|
collection_name=collection_name,
|
|
batch_size=5,
|
|
filter=f"text like \"{text}\" and resource like \"{resource}\"",
|
|
output_fields=["id","resource","text"],
|
|
)
|
|
|
|
results = []
|
|
|
|
while True:
|
|
result = iterator.next()
|
|
if not result:
|
|
iterator.close()
|
|
break
|
|
results += result
|
|
return results
|
|
def get_chunk_by_id(self, collection_name, database_name,id):
|
|
"""
|
|
return List :
|
|
"""
|
|
try:
|
|
self.client.using_database(db_name=database_name)
|
|
if not self.client.has_collection(collection_name):
|
|
return []
|
|
else:
|
|
self.client.load_collection(collection_name)
|
|
iterator = self.client.query_iterator(
|
|
collection_name=collection_name,
|
|
batch_size=5,
|
|
filter=f"id=={id}",
|
|
output_fields=["id","text","resource"],
|
|
)
|
|
|
|
results = []
|
|
|
|
while True:
|
|
result = iterator.next()
|
|
if not result:
|
|
iterator.close()
|
|
break
|
|
results += result
|
|
return results
|
|
finally:
|
|
# 3. 无论成功还是失败，最终释放集合
|
|
try:
|
|
self.client.release_collection(collection_name)
|
|
except Exception as e:
|
|
print(f"释放集合失败: {e}")
|
|
|
|
def get_chunk(collection_name,database_name,resource,text,id):
|
|
milvus_db=Milvus_Database()
|
|
# results = milvus_db.get_text_by_resource(collection_name=collection_name,database_name=database_name,resource=resource,text=text)
|
|
results = milvus_db.get_chunk_by_id(collection_name=collection_name,database_name=database_name,id=id)
|
|
return {"code": 200, "message": "success", "data": results}
|
|
|
|
if __name__ == "__main__":
|
|
text = "你好"
|
|
# text = "应，进一步放大风险的危害，而传播环境的深刻变革对触发和放大生成式人工智 能传播风险也具有重要影响。生成式人工智能传播风险打破了独立、单层、定向的风险路径， 通过机内交互、人机交互与人人交互的复杂网络结构形成了联动、多层、多向的风险路径， 极易形成多领域共振传导。"
|
|
result = get_chunk(database_name="XIAN",collection_name="ye85a36d295a199730340c5c755c1e2b",id="459963917155397121",text=text,resource="test.md")
|
|
print(result) |