intfloat

Fast and simple hashable float alternative using ints

Installing

$ cargo add intfloat

Alternatively, you can edit your Cargo.toml directly and run cargo update:

[dependencies]
intfloat = "0.1.0"

Getting started

IntFloat can be used as a fast hashable float alternative, with a limited accuracy compared to f32/f64. To use it, call it with IntFloat::from(x, y), with x the float you want to convert and y the number of decimal points you want to keep.

use intfloat::IntFloat;
let a = IntFloat::from(10 as f32, 0);
let b = IntFloat::from(5.2, 0);
assert_eq!(a, b+b);

let c = IntFloat::from(5.2, 1);
assert_ne!(b, c);

Internally, the float is represented by two integers z and y, such that x = z * 10^(-y), so for example 5.2 = 52 * 10 ^ -1.

See the docs for more functions and details. Feel free to open an issue in case some things are unclear.

Motivation & goal

This package is implemented in the context of (and usable by) the package intervalues. That package is used to track (combinations of) intervals between points, for which an internal algorithm uses HashMaps. To support float as interval bounds as well, the input floats need to be converted to a hashable float. An alternative to IntFloat is rust_decimal::Decimal, which is more mature and feature-rich, but performs about 30% slower when used with intervalues::combine_intervals on large datasets, likely due to the extra accuracy/granularity that Decimal can take into account.

This package aims to add a faster alternative when small discrepancies don't matter. For example, in many cases it is fine to round a float to 4 decimal places, which can accurately be represented by IntFloat(..., 4).

License

Licensed under either of Apache License, Version 2.0 or MIT license at your option.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this crate by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.