Here we go! Thanks ```pythonimport qdrant

Here we go! Thanks

Plain Text

import qdrant_client
from gpt_index import GPTQdrantIndex

# Creating a Qdrant vector store
client = qdrant_client.QdrantClient(
    host="<qdrant-host>",
    api_key="<qdrant-api-key>",
    https=True
)
collection_name = "paul_graham"

# Initialize the GPTQdrantIndex with the existing Qdrant vector store
index = GPTQdrantIndex(client=client, collection_name=collection_name)

# Query index
response = index.query("What did the author do growing up?")

17 comments

kkapa.ai

Please stand by @pikachu8887867 as I search the knowledge base. We are experiencing especially high traffic right now and might take longer than usual. :wave:

kkapa.ai

Sorry @pikachu8887867, I seem to be having a bit of trouble. Don't worry though, I've pinged my supprt and they will try to fix me up! :technologist:

Perhaps try again in 15mins?

ppikachu8887867

@kapa.ai no worries bro. we good 👌

kkapa.ai

Let me check the knowledge base for you. One moment! :books:

kkapa.ai

----------------------
I'm glad everything is good! If you have any questions about GPT index in the future, feel free to ask. Have a great day!
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------

ppikachu8887867

@kapa.ai index = GPTQdrantIndex(client=client, collection_name=collection_name) throws :

{ValueError}One of documents or index_struct must be provided.

kkapa.ai

Let me check the knowledge base for you. One moment! :books:

kkapa.ai

----------------------
It seems you're trying to create a GPTQdrantIndex without providing the required documents or index_struct parameter. To fix this issue, you need to load the documents using SimpleDirectoryReader and pass them to the GPTQdrantIndex constructor. Here's an example from the Qdrant Index Demo (https://github.com/jerryjliu/gpt_index/tree/main/examples/vector_indices/QdrantIndexDemo.ipynb):

Plain Text

from gpt_index import GPTQdrantIndex, SimpleDirectoryReader

# Load documents
documents = SimpleDirectoryReader('../paul_graham_essay/data').load_data()

# Create the GPTQdrantIndex
index = GPTQdrantIndex(documents, client=client, collection_name="paul_graham")

Make sure to load the documents and pass them to the GPTQdrantIndex constructor to avoid the ValueError.
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------

ppikachu8887867

@kapa.ai I do not want to do it, because I've already uploaded my documents. Now I want to use the index for inference

kkapa.ai

Let me check the knowledge base for you. One moment! :books:

kkapa.ai

----------------------
If you have already uploaded your documents and created an index, you can save the index to disk and then load it later for inference. Here's how you can do that using the information from the Qdrant Index Demo (https://github.com/jerryjliu/gpt_index/tree/main/examples/vector_indices/QdrantIndexDemo.ipynb):

Save the index to disk:

Plain Text

index.save_to_disk("index_qdrant.json")

Load the index from disk when you want to use it for inference:

Plain Text

index = GPTQdrantIndex.load_from_disk("index_qdrant.json", client=client)

After loading the index, you can query it:

Plain Text

response = index.query("<query_text>")

Make sure to replace <query_text> with your actual query.
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------

ppikachu8887867

@kapa.ai I don't want to save the index in disk

kkapa.ai

Let me check the knowledge base for you. One moment! :books:

kkapa.ai

----------------------
If you don't want to save the index to disk, you can directly use the GPTQdrantIndex object you created earlier for inference. However, you need to make sure that the client and collection_name are still available when you want to use the index for inference.

Here's an example of how to query the index:

Plain Text

response = index.query("<query_text>")

Replace <query_text> with your actual query.

Keep in mind that if you don't save the index to disk and your script or session ends, you will lose the index object and need to recreate it.
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------

ppikachu8887867

@kapa.ai Ok. So let's say I created the index in one computer. Now I'm using another computer and all I have is an api_key and url for qdrant vector store (I don't have index.json).

How can an instance of GPTQdrantIndex ?

kkapa.ai

Let me check the knowledge base for you. One moment! :books:

kkapa.ai

----------------------
If you have the API key and URL for the Qdrant vector store, you can create a Qdrant client and use the GPTQdrantIndex with the existing collection. However, you'll need to use a custom method to retrieve the documents from the Qdrant collection, as the provided examples don't cover this specific use case.

You can create a Qdrant client like this:

Plain Text

import qdrant_client

client = qdrant_client.QdrantClient(
    host="<qdrant-host>",
    api_key="<qdrant-api-key>",
    https=True
)

Replace <qdrant-host> and <qdrant-api-key> with your actual Qdrant host and API key.

Unfortunately, the provided information extracts don't include a direct method to create a GPTQdrantIndex instance using only the API key and URL. You may need to explore the Qdrant API and the qdrant_client package to retrieve the documents and create the index accordingly.
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------

Add a reply

Find answers from the community

Here we go! Thanks ```pythonimport qdrant_clientfrom gpt_index import GPTQdrantIndex