A base for an on-call chat bot for Slack and extends classes from lins05/slackbot. While slackbot supports Python 2 and 3, this repository is Python 3+ only.

This project was built during Adobe "Garage Week" 2019 in order to support the easy creation of bots for on-call Slack channels that listen to events and can respond intelligently.

Features

• PagerDuty integration for querying on call information
• Natural language processing (NLP) support for smarter bots using spaCy
  • Currently only supports text categorization for message routing, but could support more in the future
• Supports Slack blocks in addition to attachments

Usage

Generate the slack api token

First you need to get the slack api token for your bot. You have two options:

1. If you use a bot user integration of slack, you can get the api token on the integration page.
2. If you use a real slack user, you can generate an api token on slack web api page.

Perform lins05 setup

Follow the setup steps in the lins05/slackbot repository.

Configure PagerDuty integration

slackbot_settings.py:

PAGERDUTY_TOKEN = 'mytoken'
PAGERDUTY_SCHEDULE_ID = 'ABCDEFG'
PAGERDUTY_USERNAME_EMAIL_DOMAIN = 'adobe.com'

See the slackbot/plugins/oncall.py file for examples of using the PagerDuty integration.

Configure spaCy integration

Before spaCy can be used, it must have a model trained for text categorization or text labels. Please note that message routing is currently only capable based on labels and not any other spaCy document properties.

Training a spaCy model

If you already have a spaCy textcat model to use, you may skip this section completely.

You can use Yuri to classify slack messages, train a spaCy model, and test it. In the end you should have a directory containing the spaCy model files.

Configure spaCy model location

slackbot_settings.py:

SPACY_MODEL = '/model/dir'

This is the same directory that was generated using yuri or by manually training a spaCy model.

Use nlp responders in your plugins

See the slackbot/plugins/nlp.py file for examples of using NLP in your plugins.

Releasing new versions

To release a new version of this library, install bump2version into a virtualenv, and then call bump2version <part> where part is major, minor, or patch to update a certain part of the version number. Then push the tag and the master branch. The tagged version will be built on travis-ci.org and released to pypi and docker.