Is there a way to get metadata related to the clusters in which our data is stored?Like how they are stored and what type of clusters are formed, etc

General
1

I found a link here
Vector Index
Could you please provide more information on how the number of cluster vectors affects the data storage process? I am not very familiar with the technical details of Weaviate and I would appreciate your clarification.

2

Hi @SHREYAS_KULKARNI !

Not sure I quite got your question :thinking:

But putting in simple terms:

In Weaviate you have collections (or classes) that will store objects. Each object will have a vector related to it.

You can either provide those vectors yourself, or configure Weaviate properly (providing API KEYS, vectorizer config, etc) in a way that whenever needed, Weaviate will vectorize those objects or queries for you.

So consider you now have multiple objects in your class, each one with their vectors. Now you need to generate an answer for a question based on that content. This is the so called “RAG”, or Retrieval Augmented Generation.

You will query Weaviate, using for example openai generative module. What will happen under the hood:

  1. You send the question “What is a common food in Brazil?” to Weaviate, using a generative module
  2. Weaviate will vectorize the query “What is a common food in Brazil?”, and use that vector to retrieve the closest/most relevant objects to that query.
  3. Now Weaviate will send those objects content to OpenAi to generate the answer, using those documents as context.
  4. Weaviate now returns back that answer.

A great place to start ramp up your journey is:

Let me me know if this helps :slight_smile: