Struct PrefetchModel 

pub struct PrefetchModel { /* private fields */ }

Tracks prediction accuracy for model self-evaluation.

Implementations

impl PrefetchModel

pub fn new() -> Self

pub fn observe(&mut self, path_hash: u64)

Record a file access and learn transition patterns.

pub fn predict( &self, current_hash: u64, active_hashes: &[u64], ) -> Vec<(u64, f64)>

Predict which files will be accessed next, based on current state. Returns (path_hash, confidence) pairs sorted by confidence descending.

pub fn report_hit(&mut self, predicted_hash: u64, was_accessed: bool)

Report whether a predicted file was actually accessed (feedback loop).

pub fn accuracy(&self) -> f64

Current prediction accuracy (0.0 - 1.0).

pub fn free_energy(&self) -> f64

Free Energy = surprise metric. High value means predictions are poor.

pub fn should_prefetch(&self) -> bool

Should we actively prefetch? Only when model has learned enough and prediction accuracy is reasonable.

Trait Implementations

impl Default for PrefetchModel

fn default() -> Self

Returns the “default value” for a type.