From the course: Developing RAG Apps with LlamaIndex and Next.js

Unlock this course with a free trial

Join today to access over 25,200 courses taught by industry experts.

Querying stage: ChatEngine and querying engine full overview

Querying stage: ChatEngine and querying engine full overview

From the course: Developing RAG Apps with LlamaIndex and Next.js

Start my 1-month free trial Buy for my team

Querying stage: ChatEngine and querying engine full overview

“

Now, let's look at the building blocks at this stage, the querying stage. So here are the building blocks of the querying stage. We have the retriever and we have the response synthesizer, as you saw in the previous diagram. What do these do? So, the retriever is the one that defines how to efficiently get or retrieve relevant context from a knowledge base, i.e., index. this happens when a given query happens. Now, of course, the retrieval logic varies for different indices, with dense retrieval against a vector index being the most popular one. Now, a response synthesizer, as the name implies, it synthesizes something, it puts together something, it generates something. In this case, it generates a response from a large language model using a user query and a given set of retrieved text chunks. So we have text chunks that were retrieved from searching the database, in this case the Vec database, and then we have the query as well. So we put it all together and response synthesizers…

Contents