Notes on Python Programming

minesweeper

Program architecture: single responsibility, one thing should do just one thing. E.g. with classes, you could have a renderer (frontend) class and a compute (backend) class.

Avoid storing mutable state inside objects when possible - it’s akin to using global variable, with the same potential drawbacks and threats for spaghetti code.

Store constants in a class within a file to reduce import statements.

Careful with storage types. E.g. grid storage should contain Cell types, not tuples - this leads to more readable code, like cell.reveal(). Storage itself could also be a separate class from any program logic, allowing for grid.get(x,y) instead of subarray look-ups.

Store assets as an object, like a dict, to allow for iterative initialisation. For example, the following looks quite repetitive:

    def _initialise_assets(self) -> None:
        """
        Initialises the assets for the Minesweeper game.
        :returns: None
        """
        self.image_1 = pygame.transform.scale(
            pygame.image.load("assets/1.png"), (BOX_WIDTH, BOX_HEIGHT)
        )
        self.image_2 = pygame.transform.scale(
            pygame.image.load("assets/2.png"), (BOX_WIDTH, BOX_HEIGHT)
        )
        self.image_3 = pygame.transform.scale(
            pygame.image.load("assets/3.png"), (BOX_WIDTH, BOX_HEIGHT)
        )
        ...

Use logger.expection("Exception Text") instead of print("Exception Text") to match the rest of my projects. For example:

    try:
        main()
    except KeyboardInterrupt:
        print("\nExiting...")

Ensure __pycache__ isn’t being committed - occasionally the .gitignore can be committed after the initial commit, leading to mismatched version control.

---

module-generation

Learn and use Poetry to create installable packages instead of requiring the user to clone the repo and run main.py.

Learn and use GitHub Actions to do things like convert TODOs into issues.

Do not use eval, ever - it leads to Remote Code Execution vulnerabilities; never let untrusted input execute code. For example, in the following example evaluating an API response (a possible alternative could be something like returning a JSON object to be parsed):

# Parse the GPT response
    try:
        gpt_response_list = ast.literal_eval(gpt_response)