Function dist_to_sim

nodedb-vector/src/multivec/scoring.rs:27–33

Convert a distance value produced by `scalar_distance` into a similarity score where higher is better. | Metric | Conversion | |----------------|---------------------| | L2 | `−distance` | | Cosine | `1 − distance` | | InnerProduct | `−distance` (scalar_distance returns neg-IP) | | All others | `−distance` |

(d: f32, metric: DistanceMetric)

/// | InnerProduct   | `−distance` (scalar_distance returns neg-IP) |
/// | All others     | `−distance`         |
fn dist_to_sim(d: f32, metric: DistanceMetric) -> f32 {
    match metric {
        DistanceMetric::Cosine => 1.0 - d,
        // All other metrics (and unknown future metrics) use negated distance.
        _ => -d,
    }
}

// ---------------------------------------------------------------------------
// Public API