viva_math - v1.2.100

viva_math/common

Common mathematical utilities for VIVA.

Functions not found in gleam_community_maths:

Values

clamp

</> 
pub fn clamp(value: Float, min: Float, max: Float) -> Float

Clamp a value to a range [min, max].

Examples

clamp(5.0, 0.0, 10.0)  // -> 5.0
clamp(-1.0, 0.0, 10.0) // -> 0.0
clamp(15.0, 0.0, 10.0) // -> 10.0

clamp_bipolar

</> 
pub fn clamp_bipolar(value: Float) -> Float

Clamp a value to the bipolar interval [-1, 1]. Used for PAD dimensions.

Examples

clamp_bipolar(0.5)   // -> 0.5
clamp_bipolar(-1.5)  // -> -1.0
clamp_bipolar(1.5)   // -> 1.0

clamp_unit

</> 
pub fn clamp_unit(value: Float) -> Float

Clamp a value to the unit interval [0, 1].

Examples

clamp_unit(0.5)  // -> 0.5
clamp_unit(-0.5) // -> 0.0
clamp_unit(1.5)  // -> 1.0

deterministic_noise

</> 
pub fn deterministic_noise(step: Int, seed: Int) -> Float

Deterministic pseudo-random noise generator. Uses hash-based seeding for reproducibility (no external RNG needed).

Returns value in [-1, 1] range.

Examples

deterministic_noise(0, 42)  // -> some value in [-1, 1]
deterministic_noise(0, 42)  // -> same value (deterministic)

e

</> 
pub const e: fn() -> Float

exponential_decay

</> 
pub fn exponential_decay(
  value: Float,
  rate: Float,
  time: Float,
) -> Float

Exponential decay: value * exp(-rate * time)

Examples

exponential_decay(1.0, 0.5, 1.0)  // -> ~0.606
exponential_decay(1.0, 0.0, 10.0) // -> 1.0 (no decay)

inverse_decay

</> 
pub fn inverse_decay(
  base_rate: Float,
  t: Float,
  tau: Float,
) -> Float

Inverse decay rate (pattern from viva_glyph/association.gleam). η(t) = η₀ / (1 + t/τ)

Used for learning rate decay, consolidation, etc.

inverse_lerp

</> 
pub fn inverse_lerp(
  a: Float,
  b: Float,
  value: Float,
) -> Result(Float, Nil)

Inverse linear interpolation - find t given value in range [a, b].

Examples

inverse_lerp(0.0, 10.0, 5.0)  // -> Ok(0.5)
inverse_lerp(0.0, 10.0, 2.5)  // -> Ok(0.25)
inverse_lerp(0.0, 0.0, 5.0)   // -> Error(Nil) (division by zero)

inverse_sqrt_decay

</> 
pub fn inverse_sqrt_decay(
  base_rate: Float,
  t: Float,
  tau: Float,
) -> Float

Inverse square root decay. η(t) = η₀ / √(1 + t/τ)

lerp

</> 
pub fn lerp(a: Float, b: Float, t: Float) -> Float

Linear interpolation between two values.

lerp(a, b, 0.0) = a lerp(a, b, 1.0) = b lerp(a, b, 0.5) = (a + b) / 2

Examples

lerp(0.0, 10.0, 0.5)  // -> 5.0
lerp(0.0, 10.0, 0.25) // -> 2.5

pi

</> 
pub const pi: fn() -> Float

Re-export useful constants from gleam_community_maths

safe_div

</> 
pub fn safe_div(a: Float, b: Float, default: Float) -> Float

Safe division with default value on division by zero.

Examples

safe_div(10.0, 2.0, 0.0)  // -> 5.0
safe_div(10.0, 0.0, -1.0) // -> -1.0

sigmoid

</> 
pub fn sigmoid(x: Float, k: Float) -> Float

Sigmoid function: 1 / (1 + exp(-k * x))

Maps any real number to (0, 1). k controls steepness (k=1 is standard sigmoid).

Examples

sigmoid(0.0, 1.0)   // -> 0.5
sigmoid(100.0, 1.0) // -> ~1.0
sigmoid(-100.0, 1.0) // -> ~0.0

sigmoid_standard

</> 
pub fn sigmoid_standard(x: Float) -> Float

Standard sigmoid with k=1.

smoothstep

</> 
pub fn smoothstep(edge0: Float, edge1: Float, x: Float) -> Float

Smooth step function (Hermite interpolation). Returns 0 if x < edge0, 1 if x > edge1, smooth transition otherwise.

Examples

smoothstep(0.0, 1.0, 0.5)  // -> 0.5 (roughly)
smoothstep(0.0, 1.0, 0.0)  // -> 0.0
smoothstep(0.0, 1.0, 1.0)  // -> 1.0

softmax

</> 
pub fn softmax(values: List(Float)) -> Result(List(Float), Nil)

Softmax function: converts a list of values to probabilities.

softmax([x1, x2, …]) = [exp(x1)/sum, exp(x2)/sum, …] where sum = exp(x1) + exp(x2) + …

Examples

softmax([1.0, 2.0, 3.0])  // -> [0.09, 0.24, 0.67] (approx)
softmax([0.0, 0.0])        // -> [0.5, 0.5]

tau

</> 
pub const tau: fn() -> Float

wiener_increment

</> 
pub fn wiener_increment(step: Int, seed: Int, dt: Float) -> Float

Generate Wiener process increment (discrete approximation). dW ≈ √dt × N(0,1) where N is approximated by deterministic noise.

Used for stochastic differential equations.

Parameters

  • step: Current time step (for determinism)
  • seed: Random seed
  • dt: Time step size

Examples

wiener_increment(0, 42, 0.01)  // -> small noise scaled by √dt
✨ Search Document