A nice little Telegram Bot to help you be more active!
This bot has been implemented for a Seminar about Chatbots and Conversational Agents @ University of Applied Sciences of Western Switzerland The topic of this project is User Engagement in Chatbots. This prototype implements a User Engagement Analyzer which analyses the engagement and deploys different strategies for communicating with the user.
Currently the bot has two different strategies/personalities:
- supportive
- disciplined
If you want to build this project, the following things are required:
- Ruby 2.5.1
- PostgreSQL Database
- Wit.ai Account and API token
- Telegram Bot API token
Additional Requirements:
- docker
- docker-compose
Steps:
- Check out project
docker-compose build- Set Telegram and Wit API Keys in
docker-compose.yml(See below for Wit.ai and Telegram setup) docker-compose up
- Check out project
- Install Ruby (Need help? --> https://www.ruby-lang.org/en/downloads/)
- Install bundler
gem install bundler
- Install dependencies with bundler
bundle install
- Generate and populate a new
credentials.yml(Need help? --> https://medium.com/cedarcode/rails-5-2-credentials-9b3324851336)- A sample credentials file is under
config/credentials.sample.yml
- A sample credentials file is under
- Prepare the DB
bin/rails db:setup
- Start the bot
bin/rake telegram:bot:poller
To extract intents from user messages, this bot relies on wit.ai. To build a copy of this project for yourself you need to create a project on wit.ai.
Create a new Project on wit.ai from the backed up available in the file ActivityBuddy-##########.zip
Copy the API token and save it to config/credentials.yml or docker-compose.yml if using docker.
Create a new Bot on Telegram. How? https://core.telegram.org/bots
Copy the API token and save it to config/credentials.yml or docker-compose.yml if using docker.
There is a rake task to reset the database to its initial state:
bin/rails reset:full
Or with docker:
docker-compose run web bin/rails reset:full