Master the Pool Game in Python

Master the Pool Game in Python

Table of Contents:

1. Introduction
2. Cleaning up the Code
3. Implementing Game Rules
4. Adding a Reward Calculating Function
5. Reset Function
6. Optimizing Constants
7. Separating the Eight Ball
8. End Game Condition
9. Game Rules
10. Reward Function

Introduction

In this tutorial, we will be continuing with the pool game tutorial. We will clean up some code from the previous video and implement game rules. Additionally, we will add a reward calculating function. So, let's get started!

Cleaning up the Code

First, we make some slight changes to the files, especially in the main.py file. If You want to follow along, you can visit my repository and click on the history to see all the changes I've made. To update to the latest version, open a command prompt and Type "git pull".

Next, we add a new training constant and set it to true. We also add a playing skip frame constant. In the main.py file, we add an if statement to check if We Are training. If not, we use a custom frame instead of a while loop.

Implementing Game Rules

We Create a reset function and add a time.speed to give us time to reset the environment. Inside the reset function, we remove all the shapes and bodies from the self.space. We reset some other constants as well. We add a try and accept block inside the for loop for removing the spaces in the bodies. We also add the add table function and add box function inside the reset block. Additionally, we move over the data from the collision handler and add more debugging information in the print statements.

Adding a Reward Calculating Function

We need a starting time constant in order to calculate the frames per Second (FPS). We also need a draw screen constant to determine if we need to draw the screen or not. We change the training to android screen attribute inside the class instead of using the content directly.

Next, we separate the eight ball from the other balls to create an end game condition. We create a couple of more constants and pass them inside the data attribute under the collision handler.

Reset Function

We modify the reset function so that we don't have to repeat everything in the init function. We move some of the initializing tasks from the init function to the reset function. We delete the repeated method and call the reset function in the init method to initialize additional constants.

Optimizing Constants

We change the constants inside the reset method to a dictionary format for easier maintenance. We clean up some other code as well, renaming the run function to stack function and adding an action argument and a non-stop argument. We also rename the non-stop function to run function and remove the redundancy. We move the random velocity inside the run function instead of the stack function.

Separating the Eight Ball

We create an initialize collision handler to check the ball collisions. We also create a method to check the collisions on time. We check if the first contact ball is none and assign it to the Current ball. We check if the cue ball is involved in the collision and assign the cue ball and the Shape sign to reduce typing.

End Game Condition

We set the end game condition to check whether the ball has been pocketed or the length of all the balls is less than two. We check if the agent has won or lost the game and assign the reward accordingly. We move some of the other functions inside the reset method and initialize the total pocketed balls.

Game Rules

We check what color the ball contacts first. We initialize another collision handler to check the ball collisions. We create a method named ball contact to check the collisions on time. We check if the first contact ball is solid and assign the reward accordingly.

Reward Function

We calculate the reward using a simple function that utilizes the pocket count and touch count. We add debugging print functions to check the reward. We set the done condition if a ball has been pocketed and check if we have won or lost the game. We assign the reward accordingly. We run the game with a 2 second time sleep for each episode to observe the changes.

FAQ:

1. How do I update to the latest version of the code? To update to the latest version, open a command prompt and type "git pull".

2. Can I customize the reward function? Yes, feel free to implement your own reward function that suits your needs and preferences.