Add new DocIndexRetriever example

[xuechendi]

Description

Add a new example for only add data to vector store and retrieve from vector store

related AIComps change:
opea-project/GenAIComps#314

Issues

n/a.

Type of change

List the type of change like below. Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds new functionality)
• Breaking change (fix or feature that would break existing design and interface)

Dependencies

No

Tests

No

[xuechendi]

@minmin-intel @ftian1 @lvliang-intel , please take a look

[minmin-intel]

Looks good.

[hshen14]

@xuechendi It seems the new GenAI example could be part of ChatQnA (RAG). Does it make sense to improve ChatQnA to add this example?

[xuechendi]

@xuechendi It seems the new GenAI example could be part of ChatQnA (RAG). Does it make sense to improve ChatQnA to add this example?

@hshen14 , thanks for the review.
The reason we extract RAG from ChatQnA is that we will not need LLM inference part. And we are planning to call this RAG megaservice by Agent.

If we use ChatQnA directly, we will need to provide different Spec from ChatQnA with different chatQnA megaflow to only enable RAG part.

Either way is working for me, but I felt adding a new example is clearer. Also, we can support more than one optional vectorstore backend in RAG megaservice for different scenarios.

[hshen14]

@xuechendi It seems the new GenAI example could be part of ChatQnA (RAG). Does it make sense to improve ChatQnA to add this example?

@hshen14 , thanks for the review. The reason we extract RAG from ChatQnA is that we will not need LLM inference part. And we are planning to call this RAG megaservice by Agent.

If we use ChatQnA directly, we will need to provide different Spec from ChatQnA with different chatQnA megaflow to only enable RAG part.

Either way is working for me, but I felt adding a new example is clearer. Also, we can support more than one optional vectorstore backend in RAG megaservice for different scenarios.

We are now having Agent example. Since it's designed to be used by Agent, how about moving under Agent?

[xuechendi]

@kevinintel , can you help to '@' the right person who should review/approve this PR.
So we can provide it to RagAgent and InsightAgent who will treat DocRetriever as Tool.