rdf-dataset-fragmenter/lib/quadmatcher/QuadMatcherTermValue.js

Fragments an RDF dataset into multiple parts

"use strict";
var __importDefault = (this && this.__importDefault) || function (mod) {
    return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
exports.QuadMatcherTermValue = void 0;
const imurmurhash_1 = __importDefault(require("imurmurhash"));
// MurmurHash3 produces 32-bit (unsigned, according to the docs) integer values
const MURMURHASH3_MAX_VALUE = Number.MAX_SAFE_INTEGER >>> 0;
/**
 * Matches a quad by the given component regex, with optional probability.
 */
class QuadMatcherTermValue {
    term;
    regex;
    probability;
    constructor(options) {
        this.term = options.term;
        this.regex = new RegExp(options.regex, 'u');
        this.probability = options.probability;
    }
    matches(quad) {
        const termValue = quad[this.term].value;
        const termMatch = this.regex.exec(termValue);
        if (termMatch) {
            const hashValue = (0, imurmurhash_1.default)(termMatch.at(1) ?? termValue).result();
            const hashValueRelativeToMax = hashValue / MURMURHASH3_MAX_VALUE;
            return hashValueRelativeToMax <= this.probability;
        }
        return false;
    }
}
exports.QuadMatcherTermValue = QuadMatcherTermValue;
//# sourceMappingURL=QuadMatcherTermValue.js.map