Pinecone retriever [AttributeError: 'Pinecone' object has no attribute 'query']

[mllife]

I want to access an available index using llamaindex and query it using a retriver object.
"""
import os
from pinecone import Pinecone

api_key = os.environ['PINECONE_API_KEY']
pc = Pinecone(api_key=api_key)
index_name = 'multi-ext'
index = pc.Index(index_name)
index.describe_index_stats()
"""

{'dimension': 1536,
'index_fullness': 0.0,
'namespaces': {'sample_li': {'vector_count': 107}},
'total_vector_count': 107}

The vector store have nodes added already, how can I query it.
sample node:
"""
ID
db6e4cc2-d82b-4185-80d2-e96717046427

VALUES
0.0369355902, 0.0153047238, 0.0251910929, ...

METADATA
node_content: "{"id": "db6e4cc2-d82b-4185-80d2-e96717046427", "embedding": null, "metadata": {"file_path": "JNJ-US-20240123__2861917.txt", "file_name": "JNJ-US-20240123.txt", "file_type": "text/plain", "file_size": 58296, "creation_date": "2024-05-03", "last_modified_date": "2024-04-15", "company": "JNJ", "story_date": "20240123", "source": "pinecone_external"}, "excluded_embed_metadata_keys": [], "excluded_llm_metadata_keys": [], "relationships": {}, "text": "node text", "start_char_idx": null, "end_char_idx": null, "text_template": "{metadata_str}\n\n{content}", "metadata_template": "{key}: {value}", "metadata_seperator": "\n", "class_name": "TextNode"}"
_node_type: "TextNode"
creation_date: "2024-05-03"
doc_id: "None"
document_id: "None"
file_id: "2861917"
file_name: "JNJ-US-20240123.txt"
file_path: "JNJ-US-20240123.txt"
file_size: 58296
file_type: "text/plain"
last_modified_date: "2024-04-15"
ref_doc_id: "None"
source: "pinecone_external"
story_date: "20240123"
"""

The below code, is not working

from llama_index.vector_stores.pinecone import PineconeVectorStore
from llama_index.core import VectorStoreIndex
from llama_index.core.response.pprint_utils import pprint_source_node
pc = Pinecone(api_key=os.environ.get("PINECONE_API_KEY"))
vector_store = PineconeVectorStore(
    pinecone_index=pc, text_key="_node_content"
)
pc_retriever = VectorStoreIndex.from_vector_store(vector_store).as_retriever(
    similarity_top_k=2
)
# retriever = index.as_retriever(filters=filters)
nodes = pc_retriever.retrieve("some query")

error -
File ~/Projects/multi_index_poc/.venv/lib/python3.11/site-packages/llama_index/vector_stores/pinecone/base.py:465, in PineconeVectorStore.query(self, query, **kwargs)
462 else:
463 filter = kwargs.pop("filter", {})
--> 465 response = self._pinecone_index.query(
466 vector=query_embedding,
467 sparse_vector=sparse_vector,
468 top_k=query.similarity_top_k,
469 include_values=kwargs.pop("include_values", True),
470 include_metadata=kwargs.pop("include_metadata", True),
471 namespace=self.namespace,
472 filter=filter,
473 **kwargs,
474 )
476 top_k_nodes = []
477 top_k_ids = []

AttributeError: 'Pinecone' object has no attribute 'query'

library versions I am using,

!pip show llama-index
Name: llama-index
Version: 0.10.33

!pip show llama-index-vector-stores-pinecone
Name: llama-index-vector-stores-pinecone
Version: 0.1.6

!pip show pinecone-client
Name: pinecone-client
Version: 3.2.2

[dosubot[bot]]

Hey there, @mllife! 😊 Great to see you back here with another intriguing puzzle for us to solve together. Hang tight while I dive into this and I'll get back to you with some insights in a jiffy!

[mllife]

This seems to me a potential bug in the library for now, but need more opinions.

[logan-markewich]

@mllife this isn't a bug, here is the proper usage

pc = Pinecone(api_key=api_key)

# create if not already created
pc.create_index(
    name="quickstart",
    dimension=1536,
    metric="euclidean",
    spec=ServerlessSpec(cloud="aws", region="us-west-2"),
)

pinecone_index = pc.Index("quickstart")

vector_store = PineconeVectorStore(pinecone_index=pinecone_index)

[dosubot[bot]]

The issue you're encountering with the AttributeError: 'Pinecone' object has no attribute 'query' is due to an incorrect use of the Pinecone client in conjunction with the Llama Index library. The error arises because the Pinecone client object is being used where a Pinecone.Index object is expected.

To resolve this, you need to adjust your code to correctly initialize and use the Pinecone index. Here's a step-by-step guide to fix the issue:

1. Initialize the Pinecone client with your API key.
2. Create or ensure the existence of your index using the Pinecone client. This step is crucial if the index doesn't already exist.
3. Retrieve the Pinecone.Index object for your specific index using the pc.Index("index_name") method, where "index_name" is the name of your index.
4. Pass the Pinecone.Index object to the PineconeVectorStore constructor.

Here's the corrected code snippet:

from pinecone import Pinecone
from llama_index.vector_stores.pinecone import PineconeVectorStore
from llama_index.core import VectorStoreIndex

# Initialize Pinecone client
pc = Pinecone(api_key=os.environ.get("PINECONE_API_KEY"))

# Ensure the index exists and get the Index object
pinecone_index = pc.Index("quickstart")  # Replace "quickstart" with your index name

# Initialize the vector store with the correct Pinecone Index object
vector_store = PineconeVectorStore(pinecone_index=pinecone_index)

# Use the vector store to create a retriever
pc_retriever = VectorStoreIndex.from_vector_store(vector_store).as_retriever(
    similarity_top_k=2
)

# Now you can use the retriever to query
nodes = pc_retriever.retrieve("some query")

Ensure that the index name you use with pc.Index("index_name") matches the name of the index you've created or intend to use. This approach should eliminate the AttributeError and enable you to query your Pinecone index as intended.

Details

Sources

• [Bug]: AttributeError: 'str' object has no attribute 'upsert'
• [Feature Request]: Support Pinecone 3.0.0 rc2
• AttributeError: 'GPTVectorStoreIndex' object has no attribute 'query'
• llama-index-integrations/vector_stores/llama-index-vector-stores-pinecone/llama_index/vector_stores/pinecone/base.py
• llama-index-integrations/vector_stores/llama-index-vector-stores-pinecone/pyproject.toml

About Dosu
This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

To continue the conversation, mention @dosu.