Configure custom slack alerts

Set up custom slack alerts

  1. Follow the Database Access with Python instructions.
  2. In the U19_pipeline_python repository, open the u19_pipeline/alert_system/custom_alerts directory.
  3. Create a new Python file with a meaningful name for the alert (e.g. subject_bias.py).
  4. Copy the skeleton code from u19_pipeline/alert_system/alert_code_skeleton.py.
    • All Slack alert code has two parts: a Slack channel configuration and a main function, described in the next sections.

main function guide

  • This function should return a pandas DataFrame in which each row is a Slack alert message on the configured channels.
  • You can use DataJoint to get data for the alert (e.g. custom_alerts/rig_bias.py) or simply call OS scripts (e.g. custom_alerts/braininit_storage.py).
  • All columns of the DataFrame are included in the alert. (Don't add too many!)
  • DataFrame example with a Slack notification message:
  • You can check examples of some alerts in the u19_pipeline/alert_system/custom_alerts directory.

Slack channel dictionary configuration

  • The Slack channel configuration is a dictionary that links the corresponding Slack channels and conversations with a specific alert.
  • The dictionary has two keys: 'slack_notification_channel' and 'slack_users_channel'.
  • slack_notification_channel General channel names to send notifications to.
  • slack_users_channel Private direct messages to send notifications to.
  • You can add a list of channels to each of the keys:
  • slack_notification_channel Any webhook_name (see the next section).
  • slack_users_channel Any user_id with a configured slack_webhook (see the next section).

Check available notification channels:

MATLAB
  1. Execute fetch(lab.SlackWebhooks,'*').
Python
  1. Execute:
    • lab = dj.create_virtual_module('lab', 'u19_lab')
    • lab.SlackWebhooks.fetch(as_dict=True)

Check available user channels:

MATLAB
  1. Execute:
fetch(lab.User & "slack_webhook <> ''",'slack_webhook')
Python
  1. Execute:
    lab = dj.create_virtual_module('lab', 'u19_lab')
(lab.User & "slack_webhook <> ''").fetch('KEY', 'slack_webhook', as_dict=True)

Create and register new webhooks for alerts:

  1. Create a new Slack channel if needed (for notification channels).
  2. Follow the instructions to create webhooks from the Slack documentation.
  3. Copy the Slack webhook from the Slack API web page.

Add notification channels:

MATLAB
new_slack_webhook = struct
new_slack_webhook.webhook_name = (notification channel name)
new_slack_webhook.webhook_url  = (webhook url from slack API)
insert(lab.SlackWebhooks,new_slack_webhook)
Python
lab = dj.create_virtual_module('lab', 'u19_lab')
new_slack_webhook = dict()
new_slack_webhook['webhook_name'] = (notification channel name)
new_slack_webhook['webhook_url']  = (webhook url from slack API)
lab.SlackWebhooks.insert1(new_slack_webhook)

Update user channel webhook notification channels:

MATLAB
user = struct
user.user_id = (NETID of user)
update(lab.User & user,'slack_webhook', (webhook url from slack API))
Python
lab = dj.create_virtual_module('lab', 'u19_lab')
user = dict()
user['user_id'] = (NETID of user)
user['slack_webhook'] = (webhook url from slack API)
lab.User.update1(user)