retreat-challenge-slack-bot

Jira issue

Intro

Welcome to the retreat 2022 hackaton. Please follow me!

The task is to implement the logic for our Slack bot that would:

• use some simple input string command
• reach out to Athenian API to acquire and compute some data that, in your opinion, could be useful or insightful for our everyday usage
• post-process the data and prepare a Slack message in a required format

You can do it in any language, but there's still some set of limitations.

Preparation

Docker

Docker has to be installed.

Secrets

API token

To call the API let's use our personal Bearer JWT tokens.

The token expires in 24 hours.

To get your personal token you have to:

1. revert god modes if you applied them before
2. login to Athenian https://app.athenian.co
3. navigate to https://app.athenian.co/bearer
4. copy your Bearer and save it somewhere locally

To test the Bearer you can run something like this:

curl -X 'GET' \
  'https://api.athenian.co/v1/user' \
  -H 'accept: application/json' \
  -H 'Authorization: Bearer <your bearer here>'

It should return your Athenian user info.

This is exactly the same way you gonna auth to Athenian via your solution.

Slack token

Slack token is needed for debugging purposes just to validate that the message you generated looks ok in Slack itself.

To get the slack token navigate to Keeper, then Engineering/Retreat/Athenian1 - BOT token. Copy the token and save it somewhere locally.

API

You will also need API spec for obvious reasons.

Slack

You need some playground to test the messages you generated.

Regarding channels, you can use pre-created #retreat-2022-challenge-slack-bot-test, or create a separate channel for your team only:

• please create your channel with the prefix #retreat-2022-hackaton or #retreat-2022-challenge
• add your teammates
• add Athenian1 slack application

Solution

The solution can be submitted in any language, if it corresponds the following conditions.

Input

The solution has to be a CLI, that receives a flag which is a "slack command"(imagine you received this string to your slack webhook).

The format of this command depends on your solution. If you plan, let's say, to print the cycle time of a time range you will need to pass and parse the argument with an interval or timestamps etc.

Output

The output has to be a file in a JSON format. The spec is basically a simplified version of a Slack message API spec.

Here's the example:

{
  "channel": "retreat-2022-challenge-slack-bot-test",
  "title": "Test command",
  "text": "test",
  "link": "https://cutt.ly/A1it7td",
  "color": "good",
  "fields": {
    "luck": "10%",
    "skill": "20%",
    "concentrated power of will": "15%",
    "pleasure": "5%",
    "pain": "50%"
  }
}

All fields are pretty much self-explanatory.

color can be either good, warning or danger.

Message example:

Testing

To test the generated message you should just run this script:

./publish.sh ${path_to_your_message} ${slack_token}

Examples

See example-go.

Custom approach for 🐍

If you:

• want to provide a solution in Python
• feel yourself limited on functionality
• want to use more advanced Slack features

You can go with the following approach.

Navigate to Keeper Engineering/Retreat, then choose one of the bots APP, BOT and SIGN credentials of one of the apps(we have to negotiate teams and apps ownership to avoid conflicts) and run the following code:

import os

import slack_bolt
from slack_bolt.adapter.socket_mode import SocketModeHandler

app = slack_bolt.App(
    token=os.getenv("BOT_TOKEN"),
    signing_secret=os.getenv("SIGN_TOKEN"),
)


@app.message("test")
def handle_message(message, say):
    say(os.getenv("SAY"))


if __name__ == "__main__":
    SocketModeHandler(app, os.getenv("APP_TOKEN")).start()

You can use it as a template and all you need is to implement handle_message function.

Note: if you have any problems with running this approach please contact @vmarkovtsev

This will allow you to test the bot directly with input and output commands.

Speaking of which, there's a couple of docs on message builders that you could use: layouts and layout blocks.

---

That's it, GLHF!