What apis would be used between express and react?

1. Here is a rough list:

Serve the app: Send GET @ "/" return { Built React App }
To choose mode: Send Post { Multiplayer or Singleplayer } @ "/:gameid" returns { Gamestate, Status: Complete or Error}
To play a move: Send POST { move:[x, y] } @ "/:gameid" returns { Gamestate, Status: Complete or Error }

License are must.

Check whether the default branch of yours is what you expected it to be.
module is the current default branch.
I think you intended bot to be your default branch?!

List of AI agents we could potentially use (Please note, the author of this issue had limited knowledge about the vast variety of AI solutions, hence the list above would not reflect the actual optimized solution yet)

1. Minimax
2. Heuristic based
3. Re-inforcement learning.

The complexity of each solution increase by the order.

The actual game would be hosted on NodeJS Express server.
For temporary hosting, we can use REPL to store and run the backend server.
React will be used to create and update the front-end.
The choice between using FIrebase or MongoDB is debatable. Firebase could provide real-time updates which is a plus, so currently the priority would be Firebase.

Setting up workflows is another art one has to deal with, using workflows we could reduce much needed human hours and also build some cool looking dashboards right inside the repo.

On the backend the real game would either functional or class based implemented.
We could store the game objects created by a single class in mongodb or firebase easily as both support json.

Bot type: Heuristic weights

Modules till now:

1. Game setup: uttt.js
2. Testing setup: uttt.test.js extends uttt.js
3. Bot: uttt.bot.js extends uttt.test.js // Work in progress

Currently the bot that I am personally experimenting with is but static weights that I manually add based on few heuristics I have devised.
The heuristic weights and the coefficients for these weights are stored in the setup module.

But given that heuristic can fall short of performance when they encounter unexpected cases, the more robust solution would be to design neural-networks using re-inforcement learning. Which is quite a challenge in it's own. Hence this part of the issue will be taken care of in: #6 What bot?

• Programs are not well documented.
• README.md is not precise, a reader might be confused as what features are pending, and what is to added.
  That is the current state, and the state which we want.
  Improve this.
• Make it clear, understandable and precise.
• Can't work unless, well documented.
• Also, make a license for this repo.
• Make a contributing.md.
• Also, add images, explaining different things.
• Also, how to install and uninstall, and run the program.

This, is a repo, which is does not count to open source standards.

We divide the front-end into two variety.

1. User mode: Only keeps the current state of the game and general options.
2. Research mode: Have full access and special panel to run simulations on back-end. This UI will be more dynamic in nature.

Specification for user frontend:

1. It will not contain the game logic but will only reflect the game state.
2. It cannot directly update the game state. but rather have controlled user inputs which will trigger API requests to update the game state in the backend.

That is it.

Game state:

• currentPlayer,
• currentMove,
• movesArray, etc.

Controlled user inputs:

• clickable cells,
• non-clickable cells,
• start game,
• undo, etc.

API request:

• send move for currentPlayer,
• read reply,
• undo move, etc.

ROUTINE for frontend development:

1. Specify the general options for user on game setup changes in backend.
2. Track bot experimental features and update the special panel for researchers to interact with these newly designed features.

UI DESIGNs for frontend:
Figma file for user and researcher interface design

TODOs (based on routines defined):

1. Setup boards and cells that can be clicked. (Game setup)
2. Setup api requests that sends user inputs like cell location and other metadata (player, name, etc.)
3. When receiving an update, accordingly change user interface.

Firebase is known for it's realtime database which is the reason why it could be considered the only viable database for a good UX in multiplayer games. Further details of implementation and integration with the rest of the backend is open to discussion.