Feature Request: BBGM SDK

[alexluong]

Hi folks, what do you think if some core functionality of BBGM can be moved into a JS package or SDK of some sort?

My understanding is most of the game logic runs in web worker. This includes both UI management and game logic such as simming games, handle trades, AI, etc.

My suggestion is moving some of the game logic which is pure JS into an SDK. The game code will use this SDK and run it in the web worker normally. The goal is to make it easier for others to explore and interact with the game from a programmatic perspective.

My motivation is both for multiplayer angle and also simming / analysis / predictions perspective. Ultimately, I want to interact more with the game programmatically. I can already do this by copying the source code and tinker with it from there, but I would like to be able to keep up with upstream if possible.

I'm 100% happy to explore this myself if this is something you're open to. Just want to understand your thoughts if this aligns with the direction of the game. Happy to elaborate further if there's anything.

[dumbmatter]

I'm open to the idea, but it's easier said than done for a big old codebase.

I think the biggest challenge for running BBGM in Node.js is that it relies on some web APIs like IndexedDB. Although it can do a lot of stuff with minimal mocking of IndexedDB, since all the "current" data (like current teams/players, not necessarily retired players and past seasons) are all stored in an in-memory cache. That's what I do for some of my tests which do run in Node.js:

zengm/src/test/helpers.ts

Lines 9 to 35 in 4538075

	const mockIDBLeague = (): any => {
	const store = {
	index() {
	return {
	getAll() {
	return [];
	},
	};
	},
	};
	const league = {
	getAll() {
	return [];
	},
	transaction() {
	return {
	store,
	objectStore() {
	return store;
	},
	};
	},
	};
	return league;
	};

(There is also https://github.com/dumbmatter/fakeIndexedDB which incidentally I am the creator of, and it just got a new release that might make it fast enough to actually run BBGM in Node.js.)

For many isolated tasks (like simulating a game between two teams) the logic like that is already fully abstracted from any web APIs. But it's still just like a function within a giant codebase, not necessarily the easiest thing for an external project to use.

Whether this aligns with the direction of the game... not really. But if it happens to improve code quality, then that's great. If it happens to do the opposite (maybe makes things too complicated/messy/slow idk), then I might not want to merge it even if it works for your purposes.

So I guess what I'd recommend is try to think about what specifically you want to achieve, what the API would look like, etc. And I can give feedback on that and see if I think it's something I actually want in the project.