Creating an LLM Instance

Obtain an API Key

To obtain an API key, you will need to log into platform.openai.com. Use the left-hand menu to navigate to the API keys section and click + Create new secret key. Give the key the name GraphAcademy Course and click Create secret key.

If you have successfully followed these steps, you will be presented with a new API key. Click the Copy icon to the right of the key to copy to your clipboard.

Setting Streamlit Secrets

The Streamlit documentation outlines four approaches to handling secrets and credentials in your application. For simplicity, we recommend creating a secrets.toml file in .streamlit/ folder.

Copy the following text into a newly created .streamlit/secrets.toml file, replacing the value in quotes with your API key.

Below the API key, add another entry called OPENAI_MODEL. We have found that using gpt-4 yields the best results, but you can also select a different model if you are concerned about costs.

OPENAI_API_KEY = "sk-..."
OPENAI_MODEL = "gpt-3.5-turbo"

You can access this value in your Streamlit application using the st.secrets map.

import streamlit as st

openai_api_key = st.secrets['OPENAI_API_KEY']
openai_model = st.secrets['OPENAI_MODEL']

Initializing an OpenAI LLM

As the LLM will be used across the application, it makes sense to create an LLM instance that can be shared across the application.

Create a new llm.py file in the project root. Create a new variable called llm assigned to a new instance of the ChatOpenAI class.

The LLM should be initialized with an openai_api_key keyword argument. This should be set to the secret defined in the previous step.

Unresolved directive in lesson.adoc - include::https://raw.githubusercontent.com/neo4j-graphacademy/app-python/main/solutions/llm.py[tag=llm]

Initializing an Embedding Model

To use the Vector Search Index, you will also need to create an instance of the OpenAIEmbeddings model. This will be used by Langchain to create an embedding of the user’s input which will be used to find similar documents using Neo4j’s vector index.

Unresolved directive in lesson.adoc - include::https://raw.githubusercontent.com/neo4j-graphacademy/app-python/main/solutions/llm.py[tag=embedding]

Using the LLM

Once you have completed the steps, you will be able to import the llm and embeddings objects into other modules within the project.

from llm import llm, embeddings

That’s it!

Once you have completed the steps above, click the button to mark the lesson as completed.