Creating our agent for complex interactions
Our agent is going to be powered by tools that provide the functionality to meet our requirements. The main concept is that the agent will be able to take our question and use an LLM call to reason and split this task down into smaller tasks, while deciding which tools are appropriate to use to satisfy our request and deliver an answer to our question. The reasoning processes the agent uses areillustrated in the following diagram:
Figure 8.1 – The agent’s reasoning process
Begin by creating a dedicated folder that will contain all the code for our agent. This will help organize our project files and make the code base easier to navigate. I’m going to name this folder streamlit, as it’s where we’ll create our Streamlit app later on.
Copy over the Chroma database we’ve created to store the vectors in this folder: path to your project\streamlit\chapter8db\.
Next...
