wordmap/dist/algorithms/CharacterLength.js

Multi-Lingual Word Alignment Prediction

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
const Algorithm_1 = require("../Algorithm");
/**
 * Determines the likely hood that an n-gram is a phrase.
 */
class CharacterLength extends Algorithm_1.default {
    constructor() {
        super(...arguments);
        this.name = "character length";
    }
    execute(prediction) {
        let weight = 0;
        // TRICKY: do not score null alignments
        if (!prediction.target.isNull()) {
            // sentence lengths
            const sourceSentenceLength = prediction.source.sentenceCharacterLength;
            const targetSentenceLength = prediction.target.sentenceCharacterLength;
            // n-gram lengths
            const sourceLength = prediction.source.characterLength;
            const targetLength = prediction.target.characterLength;
            const primaryLengthRatio = sourceLength / sourceSentenceLength;
            const secondaryLengthRatio = targetLength / targetSentenceLength;
            // length affinity
            const delta = Math.abs(primaryLengthRatio - secondaryLengthRatio);
            // TRICKY: the power of 5 improves the curve
            weight = Math.pow(1 - delta, 5);
        }
        prediction.setScore("characterLength", weight);
        return prediction;
    }
}
exports.default = CharacterLength;