Slack Consolidate: a slackbot built with Python and Supabase

Supabase is a platform-as-a-service built on top of PostgreSQL and many other amazing open-source tools. It’s a fantastic option to create data-intensive apps and tooling. Taking advantage of the community-made Python SDK and Slack's Python SDK, you can automate tasks and build apps for several use cases.

Prerequisites

Before we dive in, let’s look at some prerequisites you'll need:

Supabase Client
- The SDK only supports Python > 3.7. You can download a supported Python version from here. pip install supabase
Python dotenv to handle API keys without exposing them in the code
- This is optional, but it will avoid issues of package dependencies and version conflicts. pip install python-dotenv
Slack SDK for Python
- This is needed to create the Slack bot pip install slack-sdk

Creating the App in Slack

Now, it is time to create the bot in Slack and get the API Keys (after granting the required scopes)

https://api.slack.com/apps

Green arrow pointing to the ‘Create New App’ button.

Then, select to create an app from a manifest (this will already set the required permissions) for the app:

Green arrow pointing to the option ‘From an app manifest’ inside the “Create an app” menu.

Select the Slack workspace to deploy:

The picture shows a dropdown UI with Supabase selected as the workspace.

Manifest.yaml

display_information:
name: SlackConsolidate
features:
bot_user:
display_name: SlackConsolidate
always_online: false
oauth_config:
scopes:
user:
  - channels:history
  - channels:read
  - channels:write
  - chat:write
  - links:read
  - users:read
  - groups:history
  - groups:read
  - mpim:history
  - im:history
bot:
  - channels:history
  - channels:read
  - links:read
  - chat:write.public
  - chat:write
  - channels:join
settings:
org_deploy_enabled: false
socket_mode_enabled: false
token_rotation_enabled: false

Enter the manifest above when asked:

The picture shows the place where you have to paste the manifest file included in the repo.

Then, confirm to create the bot:

You can see a summary screen to confirm the creation of the bot and the scopes assigned.

Granting access to the bot:

Install the app on the workspace:

The picture shows the button to Install the App to the Slack Workspace.

Inviting the bot to the channels, it will post messages on:

The command to invite the bot is below:

/invite @SlackConsolidate

The picture shows the command /invite @Slackbot in the channel #team-support.

Now, we are done with Slack for now. Let's create some buffer tables in Supabase.

Creating Tables in Supabase

Create an account here (if you don't have one yet).

We will be using Supabase's database and the Python client Libraries. First, we will create one table to store the channels that are being watched and where they are going to send the message which works analogously to a multiplexer circuit. Since you may want to watch several channels but split them into a smaller buffer e.g VIP / Enterprise, etc.

Go to the SQL Editor and run this:

CREATE TABLE slack_channels (
    id SERIAL PRIMARY KEY,
    channel text,
    channel_id text,
    p_level text DEFAULT ''::text NOT NULL,
    dest_channel text,
    dest_channel_id text,
    private int DEFAULT '0'::int NOT NULL
);

CREATE TABLE slack_watcher (
    channel_name text,
    channel_id text NOT NULL,
    message text,
    ts timestamp with time zone NOT NULL,
    ts_ms text NOT NULL,
    CONSTRAINT pk_slackwatcher PRIMARY KEY (channel_id, ts, ts_ms)
);

Adding channels to the watch list and setting the destination channel:

You may want to call conversations.list to dump all the channels and channel IDs into a CSV file, then use it to populate the table slack_channels. You can also manually get the data, but copying the links to messages in the channels:

After right-clicking a message in Slack, you can see the option to select the link.

Slack links have the following format:

https://ORGANIZATION.slack.com/archives/ channel_id/pmessage_id

Organization: subdomain used in Slack

Channel ID: It is the string that you'll need to enter in slack channels as the channel id e.g C0000ABC02DE

The name of the channel is not needed. But it is recommended to set, it so you can filter and find this information later on if needed. You can ignore everything else when setting the table for slack_channels

Examples of adding data to the channel's list:

Method 1:

Using Supabase UI (easier):

The green arrow points to insert row button inside Supabase.

Then, enter the information as needed:

The picture shows the UI in Supabase to insert a new row to the database table.

Method 2:

Go to SQL Editor and run insert commands:

Inserting a public channel named #support-channel to be monitored:

INSERT INTO slack_channels(channel,channel_id,p_level,dest_channel, dest_channel_id, private)
VALUES ('support-channel','C0000ABC02DE','Support msgs', 'all_them_messages','C0000ABC02DF',0);

Inserting a private channel named #support-enterprise to be monitored:

INSERT INTO slack_channels(channel,channel_id,p_level,dest_channel, dest_channel_id, private)
VALUES ('support-enterprise','C0000ABC02DC','Enterprise Support msgs', 'all_them_messages','C0000ABC02DF',1);

Notes:

p_level is an optional message that will be included in with the message e.g VIP customer.

private is an integer field that should be set to 1 if the channel has a 🔒padlock before the name (not a public channel).

channel_id is the source channel ID and dest_channel_id is the ID of the channel where the bot will post the message.

Setting up the environment File:

You need to get the supabase URL and API keys from here:

https://app.supabase.com/project/_/settings/api

The picture illustrates where to get the URL and the service key to setup the environment variables.

Slack environment variables:

Then, copy the bot and person API Tokens for the bot:

The picture has arrows pointing to the tokens that will be used as environment variables from Slack.

Now, you have everything needed to set up the environment file. Please note that Slack ORG is the subdomain of your slack organization i.e supabase for supabase.slack.com

The environment file:

SUPABASE_URL=https://XXXX.supabase.co
SUPABASE_KEY=eyJhbGc_SUP4N4CH0_IkpXVCJ9.SUPAKEY_*
SLACK_TOKEN=xoxp-Slack_PERSON_TOKEN
SLACK_BOT_TOKEN=xoxb-Slack_BOT_TOKEN
SLACK_ORG=slack_sub_domain

Now, we can run the bot:

It works in a similar fashion to Arduino / PIC processors of an infinity loop looking for new data and performing tasks.

import time
import logging
from slack_sdk import WebClient
from slack_sdk.errors import SlackApiError
from supabase import create_client, Client
from datetime import datetime
from dotenv import dotenv_values
logger = logging.getLogger(__name__)
config = dotenv_values(".env")
bot_client = WebClient(token=config['SLACK_BOT_TOKEN'])
client = WebClient(token=config['SLACK_TOKEN'])
SUPABASE_URL=config['SUPABASE_URL']
SUPABASE_KEY=config['SUPABASE_KEY']
supabase: Client = create_client(SUPABASE_URL, SUPABASE_KEY)
SLACK_ORG = config['SLACK_ORG']
SLACK_ORG_LINK = f"https://{SLACK_ORG}.slack.com/archives/"
############
# Config:
############
# 1.4 seconds should be the minimum to avoid passing Slack API limits.
# https://api.slack.com/docs/rate-limits#tier_t3
POOLING_DELAY = 1.4
# Check if new channels were added each hour.
SCAN_CHANNELS_DELAY = 3600.0
BUFFER_SIZE = 20

class SlackChannel:
  def __init__(self, id, name, p_level, dest_channel_id, dest_channel, private):
    """_summary_

      Args:
          id (str): Slack channel ID from the source channel
          name (str): Name of the source channel (used in logging)
          p_level (str): (Optional message) added when posting
          dest_channel_id (str): Slack channel ID for the destination channel
          dest_channel (str): Name of the destination channel (used in logging)
          private (int): Integer to check if the channel is private (private==1) or public channel
      Returns:
          SlackChannel: object
    """
    self.id = id
    self.name = name
    self.p_level = p_level
    self.dest_channel = dest_channel
    self.dest_channel_id  = dest_channel_id
    self.private  = private

def setup():
    """_summary_
        Fetches the list of channels from Supabase and returns them in a dict()
    Returns:
        dict: Dictionary with SlackChannel objects.
    """
    channels = dict()
    data = supabase.from_("slack_channels").select("channel_id, channel, p_level, dest_channel, dest_channel_id, private").execute().data
    data_dic = data
    for row in data_dic:
        channels[row['channel_id']] = SlackChannel(id = row['channel_id'],
        name = row['channel'],
        p_level = row['p_level'],
        dest_channel = row['dest_channel'],
        dest_channel_id = row['dest_channel_id'],
        private = row['private'])
    return channels

def post(src_channel, link, message):
    """_summary_
        Post a message from a source channel into the destination channel
    Args:
        src_channel (SlackChannel): SlackChannel object
        link (_type_): The link of the message in slack
        message (_type_): _description_

    Returns:
        _type_: _description_
    """
    try:
        aux_text = ""
        if src_channel.private != 1:
            aux_text = ("Message on <#"+src_channel.id+
                ">. "+src_channel.p_level+" \n"+link)
        else:
            aux_text = ("Message on <#"+src_channel.id+
                ">."+src_channel.p_level+" \n"+message+" \n"+link)
        result = bot_client.chat_postMessage(
                channel= src_channel.dest_channel_id,
                text=aux_text
                )
        logger.info(result)
    except SlackApiError as e:
        logger.error(f"Error posting message: {e}")

def ts_to_strtime(ts):
    """_summary_
        Converts the UNIX time in timestamp to ISO format.
    Args:
        ts (int): TS datetime

    Returns:
        str: ISO format datetime string for compatibility with Postgres.
    """
    aux_ts = int(ts)
    return str(datetime.utcfromtimestamp(aux_ts).isoformat())

def loop_through_channels(channels):
    """_summary_
        Loop through the channels and post messages on postgres if they aren't cached.
    Args:
        channels (dict): dict() with SlackChannel objects
    """
    for channel_id in channels:
        channel = channels[channel_id]
        conversation_history = []
        try:
            result = client.conversations_history(channel=channel.id, limit = BUFFER_SIZE)
            conversation_history = result["messages"]
            logger.info("{} messages found in {}".format(len(conversation_history), id))
        except SlackApiError as e:
            logger.error("Error creating conversation: {}".format(e))
        for message in conversation_history:
            try:
                msg_dic = dict()
                msg_dic['channel_name'] = channel.name
                msg_dic['channel_id'] = channel.id
                aux_msg = "<@"+message['user']+"> wrote: \n"
                msg_dic['message'] = aux_msg + message['text']
                ts_aux  = message['ts'].split(".")
                msg_dic['ts'] = ts_to_strtime(ts_aux[0])
                msg_dic['ts_ms'] = ts_aux[1]
                supabase.table("slack_watcher").insert(msg_dic).execute()
                auxint = ts_aux[0]+ts_aux[1]
                auxint = auxint.replace(".","")
                link = SLACK_ORG_LINK+channel.id+"/p"+auxint
                post(channel, link, msg_dic['message'])
            except Exception as e:
                pass
        time.sleep(POOLING_DELAY)

def main():
    """_summary_
        Main loop to infinitely keep pooling data from channels and posting on Slack.
        It also checks for new channels every hour.
    """
    channels = setup()
    start = time.time()
    while True:
        end = time.time()
        if ((end - start) > SCAN_CHANNELS_DELAY):
            start = time.time()
            channels = setup()
        else:
            loop_through_channels(channels)

if __name__ == '__main__':
    main()

Of course, we aren't doing proper Python example if we don't make a test notebook available:

Conclusion

Using Supabase and Slack SDK, it is very easy to create a bot that consolidates data according to the rules set. It just takes some steps to get started with the Python SDK and you can even run a demo directly in Google Colab.

If you have any questions please reach out via Twitter or join our Discord.