Slackbots — from commands to conversations

This is the codebase for 2019 DevUp conference talk.

The slides for the talk can be found here.

Talk Information

Abstract

Have you ever tried building a Slackbot, but couldn't move past single word commands? Processing natural language is complex and daunting for anyone especially beginners. What if there was a shortcut for diving into natural languages processing (NLP). That is where classifiers come in. They can streamline your introduction into NLP. In this session we will build a NodeJS Slackbot leveraging a regression classifier. My hope is that you will be able to turn around and use this knowledge. Lets make your next Slackbot capable of understanding natural language together.

Setup

Environment Variables

The following environment variables are needed to run this application.

# For Mac or Linux
export SLACKEDEX_SLACK_TOKEN="slack_token"

You will get the Slack token from an application created via the Slack API site.

Repo Setup

You will need to have NodeJS v10+ to run this application.

1. Clone the repository
2. Setup up the Environment Variables
3. run npm install to install all the third party packages.

How to Run (Package commands)

Run the following command to start the completed solution.

npm run completed

To run the non-completed version without hot-reloading you can run the following command.

npm start

Questions

If you have any questions, feel free to reach out to @Chris Marfia on the STL Tech Slack.