Skip to main content

as_f64_literal

Function as_f64_literal 

Source
pub fn as_f64_literal(value: &Value) -> Option<f64>
Expand description

Resolve a Value into an f64 if numeric-shaped.

Same UNKNOWN-coercion behavior as as_usize_literal but for floating-point contexts (fusion weights, scoring thresholds, confidence intervals).

§Bounds

f64::from_str accepts NaN, inf, subnormals, and values that overflow to ±inf via the IEEE-754 rules. Callers that need to reject those (e.g. fusion weights outside [0, 1]) must validate the returned value themselves — this helper is purely a literal extractor, not a domain validator.