cmpstr/dist/cjs/metric/JaroWinkler.cjs

CmpStr is a lightweight, fast and well performing package for calculating string similarity

// CmpStr v3.2.2 build-bb61120-260311 by Paul Köhler @komed3 / MIT License
'use strict';

var Pool = require('../utils/Pool.cjs');
var Metric = require('./Metric.cjs');

class JaroWinklerDistance extends Metric.Metric {
  constructor(a, b, opt = {}) {
    super('jaroWinkler', a, b, opt, true);
  }
  compute(a, b, m, n) {
    const [matchA, matchB] = Pool.Pool.acquireMany('int32', [m, n]);
    try {
      for (let i = 0; i < m; i++) matchA[i] = 0;
      for (let i = 0; i < n; i++) matchB[i] = 0;
      const matchWindow = Math.max(0, Math.floor(n / 2) - 1);
      let matches = 0;
      for (let i = 0; i < m; i++) {
        const start = Math.max(0, i - matchWindow);
        const end = Math.min(i + matchWindow + 1, n);
        for (let j = start; j < end; j++) {
          if (!matchB[j] && a[i] === b[j]) {
            matchA[i] = 1;
            matchB[j] = 1;
            matches++;
            break;
          }
        }
      }
      let transpos = 0,
        jaro = 0,
        prefix = 0,
        res = 0;
      if (matches > 0) {
        let k = 0;
        for (let i = 0; i < m; i++) {
          if (matchA[i]) {
            while (!matchB[k]) k++;
            if (a[i] !== b[k]) transpos++;
            k++;
          }
        }
        transpos /= 2;
        jaro = (matches / m + matches / n + (matches - transpos) / matches) / 3;
        for (let i = 0; i < Math.min(4, m, n); i++) {
          if (a[i] === b[i]) prefix++;
          else break;
        }
        res = jaro + prefix * 0.1 * (1 - jaro);
      }
      return {
        res: Metric.Metric.clamp(res),
        raw: { matchWindow, matches, transpos, jaro, prefix }
      };
    } finally {
      Pool.Pool.release('int32', matchA, m);
      Pool.Pool.release('int32', matchB, n);
    }
  }
}
Metric.MetricRegistry.add('jaroWinkler', JaroWinklerDistance);

exports.JaroWinklerDistance = JaroWinklerDistance;