Memoizee

Memoizee is a Rust library designed to simplify caching for both synchronous and asynchronous functions. With just a procedural macro, you can avoid redundant computations and improve performance effortlessly.

Features

• Synchronous memoization: Automatically cache results for synchronous functions.
• Asynchronous memoization: Cache results for async functions with ease.
• Simple to use: Add the #[memoize] attribute to your functions, and you're done!

Installation

Add memoizee to your Cargo.toml:

[dependencies]
memoizee = "0.2.0"

Usage

Here's how you can use memoizee to cache the results of your synchronous and asynchronous functions:

use memoizee::memoize;

#[memoize]
fn compute_sync(x: u32) -> u32 {
    println!("Computing sync...");
    x * x
}

#[memoize]
async fn compute_async(x: u32) -> u32 {
    println!("Computing async...");
    x * x
}

#[tokio::main]
async fn main() {
    // Test async memoization
    assert_eq!(compute_async(3).await, 9); // Prints: "Computing async..."
    assert_eq!(compute_async(3).await, 9); // Cached, no print

    // Test sync memoization
    assert_eq!(compute_sync(3), 9); // Prints: "Computing sync..."
    assert_eq!(compute_sync(3), 9); // Cached, no print
}

How It Works

1. Add the #[memoize] attribute to your function.
2. The macro generates a static memoizer that stores previously computed results.
3. For synchronous functions, results are cached using a lightweight in-memory store.
4. For asynchronous functions, memoizee handles futures and caches their resolved values.

Benefits

• Performance: Avoid redundant computations for frequently called functions.
• Ease of use: Requires no boilerplate—just annotate your functions with #[memoize].
• Flexibility: Works seamlessly with both sync and async functions.

Limitations

• Currently, #[memoize] only supports functions with a single argument.
• The argument type must implement Clone, Eq, and Hash.
• Return values must implement Clone.

Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request if you have ideas for improvements or bug fixes.

License

This project is licensed under the MIT License.