Struct Tolerance 

pub struct Tolerance<F> { /* private fields */ }

The tolerance used to compare to floating point numbers.

Generally, two numbers x and y are approximately equal if

|x - y| < max(R * (|x + y|), A)

where R is the relative tolerance and A is the absolute tolerance.

The most common way to initialize this struct is to use Tolerance::<F>::default(). In that case, the relative and absolute tolerances are computed using an heuristic based on the EPSILON and MIN_POSITIVE values of the given floating point type F.

Another common initialization is Tolerance::<F>::rel_abs(1e-4, 1e-5).set_half_precision_relative(1e-2). This will use a sane default to manage values too close to 0.0 and use different relative tolerances depending on the floating point precision.

Implementations

impl<F: Float> Tolerance<F>

pub fn strict() -> Self

Create a tolerance with strict precision setting.

pub fn balanced() -> Self

Create a tolerance with balanced precision setting.

pub fn permissive() -> Self

Create a tolerance with permissive precision setting.

pub fn rel_abs<FF: ToPrimitive>(relative: FF, absolute: FF) -> Self

When comparing two numbers, this uses both the relative and absolute differences.

That is, x and y are approximately equal if

|x - y| < max(R * (|x + y|), A)

where R is the relative tolerance and A is the absolute tolerance.

pub fn relative<FF: ToPrimitive>(tolerance: FF) -> Self

When comparing two numbers, this uses only the relative difference.

That is, x and y are approximately equal if

|x - y| < R * max(|x|, |y|)

where R is the relative tolerance.

pub fn absolute<FF: ToPrimitive>(tolerance: FF) -> Self

When comparing two numbers, this uses only the absolute difference.

That is, x and y are approximately equal if

|x - y| < A

where A is the absolute tolerance.

pub fn set_relative<FF: ToPrimitive>(self, tolerance: FF) -> Self

Change the relative tolerance to the given one.

pub fn set_half_precision_relative<FF: ToPrimitive>(self, tolerance: FF) -> Self

Change the relative tolerance to the given one only if F is half precision.

pub fn set_single_precision_relative<FF: ToPrimitive>( self, tolerance: FF, ) -> Self

Change the relative tolerance to the given one only if F is single precision.

pub fn set_double_precision_relative<FF: ToPrimitive>( self, tolerance: FF, ) -> Self

Change the relative tolerance to the given one only if F is double precision.

pub fn set_absolute<FF: ToPrimitive>(self, tolerance: FF) -> Self

Change the absolute tolerance to the given one.

pub fn set_half_precision_absolute<FF: ToPrimitive>(self, tolerance: FF) -> Self

Change the absolute tolerance to the given one only if F is half precision.

pub fn set_single_precision_absolute<FF: ToPrimitive>( self, tolerance: FF, ) -> Self

Change the absolute tolerance to the given one only if F is single precision.

pub fn set_double_precision_absolute<FF: ToPrimitive>( self, tolerance: FF, ) -> Self

Change the absolute tolerance to the given one only if F is double precision.

pub fn approx_eq(&self, x: F, y: F) -> bool

Checks if x and y are approximately equal given the tolerance.

Trait Implementations

impl<F: Clone> Clone for Tolerance<F>

fn clone(&self) -> Tolerance<F>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<F: Debug> Debug for Tolerance<F>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<F: Float> Default for Tolerance<F>

fn default() -> Self

Returns the “default value” for a type.

impl<F: Copy> Copy for Tolerance<F>