Evaluator

An evaluator is a trait that represents a predicate function evaluating against a value.

trait Evaluator<T> {
    fn evaluate(self, value: T) -> bool;
}

Specifically, in the context of Mímir, an evaluator checks if the value of a fact about the game's current state matches a certain condition.

You can choose to create your own implementation of the trait, or use the provided FloatEvaluator implementation (see below) that allows you to evaluate floating-point numbers (Rust's f64 type).

Real world

In the real world, an evaluator represents a condition that must be true for a contextual event to take place. However, events will typically have many evaluators that need to evaluate to true, not just one!

ℹ️ An NPC might query Mímir to ensure that they're only commenting on another NPC's behaviour if they've not exhibited the same behaviour previously (to avoid being hypocritical).

FloatEvaluator

⚠️ To use the pre-made FloatEvaluator implementation, you must enable the float feature in your project's Cargo.toml:

[dependencies]
subtale-mimir = { version = "0.5.0", features = ["float"] }

The FloatEvaluator is a built-in implementation of the Evaluator<T> trait, allowing you to define evaluators that match against floating-point numbers.

enum FloatEvaluator {
    EqualTo(f64),
    NotEqualTo(f64),
    LessThan(FloatRangeBound),
    GreaterThan(FloatRangeBound),
    InRange(FloatRangeBound, FloatRangeBound),
}

If you're interested in how we've implemented the Evaluator<f64> trait for FloatEvaluator, check out the source code on GitHub!

ℹ️ FloatRangeBound is an enum that holds a boundary value that can be inclusive (FloatRangeBound::Inclusive(f64)) or exclusive (FloatRangeBound::Exclusive(f64)).

Helper functions

Several helper functions are exposed to easily instantiate FloatEvaluator with common equality expressions:

ℹ️ FloatEvaluator::range is designed to mimic the functionality of Python's built-in range function.

Floating-point equality

Internally, Mímir's FloatEvaluator uses the float-cmp crate to perform approximate comparisons when FloatEvaluator::EqualTo or FloatEvaluator::NotEqualTo are evaluated.