Struct CompressionPredictorConfig 

pub struct CompressionPredictorConfig {
    pub enabled: bool,
    pub min_samples: u64,
    pub candidate_ratios: Vec<f32>,
    pub retrain_interval: u64,
    pub max_training_samples: usize,
}

Configuration for the performance-floor compression ratio predictor (#2460).

When enabled = true, before hard compaction the predictor selects the most aggressive compression ratio that keeps the predicted probe score above probe.hard_fail_threshold. Requires enough training data (min_samples) before activating — during cold start the predictor returns None and default behavior applies.

Fields

enabled: bool

Enable the adaptive compression ratio predictor. Default: false.

min_samples: u64

Minimum training samples before the predictor activates. Default: 10.

candidate_ratios: Vec<f32>

Candidate compression ratios evaluated from most to least aggressive. Default: [0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9].

retrain_interval: u64

Retrain the model after this many new samples. Default: 5.

max_training_samples: usize

Maximum training samples to retain (sliding window). Default: 200.

Trait Implementations

impl Clone for CompressionPredictorConfig

fn clone(&self) -> CompressionPredictorConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CompressionPredictorConfig

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

Formats the value using the given formatter.

impl Default for CompressionPredictorConfig

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for CompressionPredictorConfig

where CompressionPredictorConfig: Default,

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for CompressionPredictorConfig

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.