Struct PagedMarkovModel 

pub struct PagedMarkovModel { /* private fields */ }

Memory-paged Markov model for shell command prediction.

Uses aprender’s bundle module to store n-gram data on disk and load segments on-demand, enabling handling of large shell histories without exhausting RAM.

Implementations

impl PagedMarkovModel

pub fn new(n: usize, memory_limit_mb: usize) -> Self

Create a new paged model with given n-gram size and memory limit.

Arguments

• n - N-gram size (2-5)
• memory_limit_mb - Maximum memory usage in megabytes

pub fn memory_limit(&self) -> usize

Get memory limit in bytes.

pub fn train(&mut self, commands: &[String])

Train on a list of commands.

pub fn save(&self, path: &Path) -> Result<()>

Save model to a paged bundle file.

pub fn load(path: &Path, memory_limit_mb: usize) -> Result<Self>

Load model from a paged bundle file with memory limit.

pub fn suggest(&mut self, prefix: &str, count: usize) -> Vec<(String, f32)>

Suggest completions for a prefix.

pub fn stats(&self) -> PagedModelStats

Get model statistics.

pub fn paging_stats(&self) -> Option<PagingStats>

Get paging statistics from the bundle.

pub fn prefetch_hint(&mut self, prefix: &str)

Hint that a segment will be needed soon (for prefetching).

pub fn total_commands(&self) -> usize

Total commands trained on.

pub fn ngram_size(&self) -> usize

N-gram size.

pub fn vocab_size(&self) -> usize

Vocabulary size.

pub fn top_commands(&self, count: usize) -> Vec<(String, u32)>

Top commands by frequency.