tanliyon / gym-xiangqi

This repo sets up the environment to play Xiang Qi (chinese chess) following the OpenAI Gym framework.
GNU Lesser General Public License v3.0
32 stars 6 forks source link

Dkim/modify game screen #55

Closed dooki114 closed 3 years ago

dooki114 commented 3 years ago

Description

Screen Shot 2021-04-06 at 4 20 19 AM

Type of change

How has this been tested?

hojoungjang commented 3 years ago

@tanliyon Actually, I'll do it here.

So basically my idea was to expose just the step_user() method and users/developers can create whatever custom game mode loop they want just like how step() method is used in every gym environment. Developers create their own training loop with the environment's exposed methods and let their agents interact with the environment.

So that was my idea with game mode implementation and with this we do not have a clear way to insert "GAME OVER" message because the program execution will immediately jump back to the users' code and not our code once the game is finished. Hope this part was clear.

But then, in our last meeting, you brought up the idea of implementing the game mode on our end and wrapping and exposing the game mode itself as an executable application with command line arguments right? In that case I think we can easily add the "GAME OVER" message.

dooki114 commented 3 years ago

As we agreed in our previous discussion, Timer has been temporarily deactivated, and is not being displayed on the screen. Once our discussion on "Game Over" settles, I'll make more revisions!

tanliyon commented 3 years ago

Ahh I see what you mean @hojoung97, yea I agree with your idea. So for now I guess we can remove it, and later on we can add it into the game_mode executable.

hojoungjang commented 3 years ago

Yup sounds good, so let's leave the game_over() method in xiangqi_game.py for now since we might use it later