spark

v1.0.0

Initial version is a slack bot that sits in on 1/1 slack channels, records slack huddles, and categorizes the comments into performance related buckets. The intention is to showcase the ability to create structure out of unstructured data.

TODO [ ] Hello world slack bot [ ] Recall integration to sit in on slack huddles [ ] Return a transcript of the slack huddle [ ] Categorize slack huddles into performance related buckets [ ] Post back into the slack channel the performance related summary [ ] Store the summary, video recording, and the transcript in a database [ ] Admin page to see a video recording and the transcript, along with the performance related buckets [ ] Deploy with Tony Grillo, ForKeeps

Installing Deno

To install Deno, run the following command in your terminal:

curl -fsSL https://deno.land/x/install/install.sh | sh

Environment Variables

Create a .env file in the root directory of your project copy the variables in the .env.example file into it. They include:

RECALL_API_KEY=your_recall_api_key
SLACK_BOT_TOKEN=your_slack_bot_token
PYTHON_BACKEND_API=your_python_backend_api_key

Dependencies

Deno manages dependencies through direct URL imports. Ensure that all necessary dependencies are correctly imported in your code. For example, to import the dotenv package, add the following import statement to your code:

import "https://deno.land/x/dotenv/load.ts";

For more information on Deno packages, visit the Deno third-party modules page.

Database

The database url is in the .env.example file. Copy this value to your .env file. Then run this command to generate Prisma client:

deno task prisma-generate

Running the App

• Run deno task dev in the terminal

Scripts

There are two scripts in the application- one posts a conversation transcript to Slack while the other posts a summary gotten from OpenAI.

• Run deno task transcribe $botId to get a conversation transcript in the huddle's private Slack channel
• Run deno task summarize $botId to get a summary of the conversation in buckets posted to the huddle's private channel

Production Server

• The Deno server is deployed here