openai_setup.md

Interacting with the Azure OpenAI API from Slack

Azure OpenAI is a cloud-based service that provides access to OpenAI's powerful language models, including GPT-3, Codex, and DALL-E. It is designed for enterprise customers who need to use these models in a secure and compliant environment.

Prerequisites

For the corresponding terraform workspace (set via environment vairable TF_WORKSPACE), see Create Slack and OpenAI secrets in AWS Secrets Manager to create the following SSM Parameters in AWS Systems Manager for the OpenAI function to fetch and interact with Azure OpenAI,

• /azure/openai/endpoint/{TF_WORKSPACE}
• /azure/openai/key/{TF_WORKSPACE}

where {TF_WORKSPACE} will be replaced by the current terraform workspace in use.

If you are testing the function locally, set these environment variables,

• AZURE_OPENAI_ENDPOINT
• AZURE_OPENAI_KEY

Follow Azure's documentation here on how to create an OpenAI resource and deploy a model to it. See this example to understand keys and endpoints and where to retrieve them from.

Finally, replace the default environment variable AZURE_OPENAI_DEPLOYMENT_ID according to your deployment in Azure.

Interacting with OpenAI from Slack

• To stop the application from responding to every single text message in the channel, the bot will only respond to messages that start with hey openai. For example, hey openai, tell me a joke will trigger the bot to respond with a joke.
• Users can only interact with OpenAI in the channel specified in the environment variable XPLORERS_OPENAI_SLACK_CHANNEL_ID which is the channel C05G6U88QMC(#ask-openai) in Xplorers Slack Workspace. This is to prevent the bot from responding to messages in every channel it is invited to.

Deploy your feature branch to AWS

Follow the steps in feature branch deployment guide to deploy your feature branch to AWS.