File size: 756 Bytes
8cf4b8e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
import os
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
# Hàm khởi tạo retriever
def load_retriever():
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
retriever = FAISS.load_local("vectorstore", embeddings, allow_dangerous_deserialization=True).as_retriever(search_kwargs={"k": 5})
return retriever
# Lần đầu load retriever
retriever = load_retriever()
# Hàm reload retriever khi thêm tài liệu
def reload_retriever():
global retriever
retriever = load_retriever()
# Hàm retrieve_docs để lấy tài liệu
def retrieve_docs(query):
return retriever.get_relevant_documents(query)
|