Grandmasters Slack-Based Chess Game

Setup

• Ensure you have Python3.7.5 and PIP installed
• RUN ALL commands in Setup.md file in the root directory of this REPO
  
  
  

Running the game in command line

• open python interpreter - e.g. python3 on Mac from the src directory
• import ui
• ui = ui.UI()
• ui.start()
  
  
  

Running the game in Slack

• If you want to run in slack, then ensure you have slackclient installed (python3 -m pip install slackclient) and the env variable passed when starting the python interpreter in src directory (e.g. SLACK_API_TOKEN="token" python3 slack_ui.py)

• Open up your workspace and invite Chessy! /invite @Chessy
  
  
  

Running Tests

• Once pytest is installed, as per Setup MD, you can run tests from your project directory in CMD using python3 -m pytest
• To include coverage, use python3 -m pytest ./

Techstack

---

Tech used	What does it do?	How we used it
Python 3	Backend, (MVC)	In MVC layers
Pytest	Python Testing Framework	TDD, TDD, TDD!
Pytest-Cov	Measures test coverage	Measure test coverage
Slack API	Allows you to interface with Slack	Created an API with our Python Chess Logic
SlackClient	Launches your Slack App from CMD with an AUTH token	Resurrecting Chessy the Chess Bot
CairoSVG	Converts SVG format images to other image types	Converting SVG to PNG and rendering the current board
Python-Chess	Python Chess Library	ONLY used for front-end board-rendering
Pillow	Python Image Library	Manipulating Images/saving files in board rendering
Travis	CI Tool	To make sure pull requests weren't breaking our functionality

---

Our Process

• We focused on a feature lead design with a strict focus on TDD using Pytest. We made sure that we built each module so it was easily abstractable, and could be replaced easily (e.g. Rulesets, pieces etc.)

• We worked in 2 day sprints, and paired on the more technically difficult parts of the script. Working on modular design with clear separate responsibilities for each team member meant that there were barely any merge conflicts throughout

• By the end of the second day we had arrived at our MVP:

Command-line version using unicode chars

First Slack Version

• Once we had a working MVP, we moved on to creating a module for each piece type, and making sure they each inherited from a master Piece class

• This is where the headaches started. Programming Check/Checkmate/Stalemate conditions proved tricky. We prevailed!

• We extracted rules to a StandardRules class, from which weird and wonderful new game variants could inherit

• Next, we moved on to implementing in Slack. We set up the slack UI to accept commands and print the state of the board in a slack channel

• Finally, we wanted to compete against an AI, and see how complex we could build a chess AI using a minimax algorithm. This required a few different versions and lots of reading/understanding Python iterative methods

• We finally arrived, by the end of the second week, with a product that:

  • Has a Python backend that can run a command line UI but also interact with the Slack API to run in slack channels
  • Adheres to all of the rules of chess (legal/illegal moves, check/checkmate/stalemate etc.)
  • Multiple games can be played in parallel by any users that invite the Chessy Bot.
  • Has many different rulesets that can be injected into the game at various points (Bishops excommunicating, Rooks being sold off etc.)
  • Has an AI option, so that you can play vs DominationBot3000
  • This AI option can be modified to increase/decrease the depth (though increasing the depth has an exponential performance trade-off)
  • Renders a PNG image of the board after each legal turn is made, so that each player can see the state of the board before they make their turn.
  • Is great fun!
    
    
    

Minimax Algorithm Iterations

Kamikaze bot

Bot only wants to move its Knights

Bot works!