pub struct EngineConfig {
pub min_clone_tokens: usize,
pub winnow_window: usize,
pub literals: LiteralNorm,
pub posting_cap: usize,
pub pair_budget: usize,
pub max_statement_window: usize,
pub entropy_ratio_floor: f64,
pub degree_cap: usize,
}Expand description
Engine tuning. The defaults are the evaluated configuration.
Fields§
§min_clone_tokens: usizeSmallest clone length in tokens; also the k-gram length.
winnow_window: usizeWinnowing window: every run of at least winnow_window + min_clone_tokens - 1 shared tokens is guaranteed a shared fingerprint.
literals: LiteralNormLiteral-normalization strategy for the Type-2 pass.
posting_cap: usizeLongest posting list (raw pass) or fragment class (Type-2 pass) that still enters pairing; longer ones are dropped and counted.
pair_budget: usizeUpper bound on candidate pairs examined by each pass. Pairing is rarest-first, so exhaustion sacrifices the lowest-signal candidates. The allowance is per pass rather than shared: the raw pass runs first and would otherwise be able to spend the whole of it, which stops the renamed-copy pass finding anything at all.
max_statement_window: usizeLargest number of consecutive statements cut as one candidate fragment.
entropy_ratio_floor: f64Groups whose normalized content-entropy ratio is below this value are marked suppressed as degenerate repetition.
degree_cap: usizeGroups with more members than this are marked suppressed as recurring boilerplate.
Trait Implementations§
Source§impl Clone for EngineConfig
impl Clone for EngineConfig
Source§fn clone(&self) -> EngineConfig
fn clone(&self) -> EngineConfig
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more