Find answers from the community

Updated 4 months ago

Trying To Get The Nodes From The Retrieve Function

At a glance

The community member is trying to retrieve nodes using the VectorContextRetriever function, but is getting an empty list as the result. Another community member suggests that the issue might be with the VectorContextRetriever and that the community member might be missing the vector store. The solution provided by another community member is to add the vector_store=index.vector_store parameter to the VectorContextRetriever function, which seems to have resolved the issue.

I am trying to get the nodes from the retrieve funcion
but I am getting empty list as 'nodes' results.

import openai
#from openai import OpenAI
import os
import sys
from collections import Counter
import json
sys.path.append('../..')

from dotenv import load_dotenv, finddotenv = load_dotenv(find_dotenv()) # read local .env file

openai.api_key = os.environ['OPENAI_API_KEY']
client = OpenAI()
import nest_asyncio
nest_asyncio.apply()

embed_model = OpenAIEmbedding(model='text-embedding-3-small')
llm = OpenAI(temperature=0, model="gpt-4o-mini")

documents = SimpleDirectoryReader(
"......../small_docs"
).load_data()

index = PropertyGraphIndex.from_documents(
documents,
)

from llama_index.core.indices.property_graph import VectorContextRetriever

vector_retriever = VectorContextRetriever(
index.property_graph_store,
embed_model=embed_model,
include_text=False,
similarity_top_k=2,
path_depth=1,
)

retriever2 = index.as_retriever(sub_retrievers=[vector_retriever])
i
L
6 comments
Thanks for your response. how can i fix that?
this code (without vectorcontexretriever) works fine: # use
retriever = index.as_retriever(
include_text=False, # include source chunk with matching paths
similarity_top_k=5, # top k for vector kg node retrieval
)
nodes = retriever.retrieve("who is bob?")
I think somehow the problem is VectorContextRetriever
I think you are missing the vector store
Plain Text
vector_retriever = VectorContextRetriever(
    index.property_graph_store,
    embed_model=embed_model,
    include_text=False,
    similarity_top_k=2,
    path_depth=1,
    vector_store=index.vector_store,
)
thnks bro, it worked.
Add a reply
Sign up and join the conversation on Discord