Module effect_system 

Effect system for tracking computational effects in TensorLogic expressions.

This module provides an effect system that tracks various kinds of computational effects in logical expressions and tensor operations, enabling:

• Effect tracking: Know which operations have side effects
• Differentiability: Track which operations support gradient computation
• Probabilistic reasoning: Distinguish deterministic from stochastic operations
• Memory safety: Track memory access patterns
• Effect polymorphism: Functions parametric over effects

Examples

use tensorlogic_ir::effect_system::{Effect, EffectSet, ComputationalEffect};

// Pure computation (no side effects)
let pure_effect = EffectSet::pure();
assert!(pure_effect.is_pure());

// Differentiable operation
let diff_effect = EffectSet::new()
    .with(Effect::Computational(ComputationalEffect::Pure))
    .with(Effect::Differentiable);

// Combine effects
let combined = pure_effect.union(&diff_effect);
assert!(combined.contains(&Effect::Differentiable));

Structs

EffectAnnotation

Effect annotation for expressions

EffectSet

Set of effects for an expression or operation.

EffectVar

Effect variable for effect polymorphism

Enums

ComputationalEffect

Computational purity effects.

Effect

Individual effect kinds.

EffectScheme

Effect scheme for effect polymorphism (analogous to type schemes)

MemoryEffect

Memory access effects.

ProbabilisticEffect

Probabilistic effects.

Functions

infer_operation_effects

Infer effects for common operations

Type Aliases

EffectSubstitution

Substitution mapping effect variables to effect sets