41 lines
1.3 KiB
JavaScript
41 lines
1.3 KiB
JavaScript
// lib/parsers/jaccard.mjs
|
||
// Jaccard similarity for SC4 determinism floor.
|
||
//
|
||
// jaccard(A, B) = |A ∩ B| / |A ∪ B|
|
||
// Inputs are arrays of strings; deduplicated internally.
|
||
// Both empty → 1.0 (vacuously identical). One empty → 0.0.
|
||
|
||
/**
|
||
* Compute Jaccard similarity between two string sets.
|
||
* @param {string[]} setA
|
||
* @param {string[]} setB
|
||
* @returns {number} similarity in [0, 1]
|
||
*/
|
||
export function jaccardSimilarity(setA, setB) {
|
||
if (!Array.isArray(setA) || !Array.isArray(setB)) {
|
||
throw new TypeError('jaccardSimilarity: both inputs must be arrays');
|
||
}
|
||
const a = new Set(setA);
|
||
const b = new Set(setB);
|
||
if (a.size === 0 && b.size === 0) return 1.0;
|
||
if (a.size === 0 || b.size === 0) return 0.0;
|
||
|
||
let intersection = 0;
|
||
for (const x of a) {
|
||
if (b.has(x)) intersection += 1;
|
||
}
|
||
const union = a.size + b.size - intersection;
|
||
return intersection / union;
|
||
}
|
||
|
||
/**
|
||
* Check whether a similarity meets a threshold.
|
||
* @param {number} similarity
|
||
* @param {number} threshold
|
||
* @returns {boolean}
|
||
*/
|
||
export function meetsThreshold(similarity, threshold) {
|
||
if (typeof similarity !== 'number' || typeof threshold !== 'number') return false;
|
||
if (!Number.isFinite(similarity) || !Number.isFinite(threshold)) return false;
|
||
return similarity >= threshold;
|
||
}
|