Grammalecte  Diff

// JavaScript

// FSA DICTIONARY BUILDER
//
// by Olivier R.
// License: MPL 2
//
// This tool encodes lexicon into an indexable binary dictionary
// Input files MUST be encoded in UTF-8.

/* jshint esversion:6, -W097 */
/* jslint esversion:6 */
/* global require, exports, console, helpers */

"use strict";

if(typeof process !== 'undefined') {
    var str_transform = require('./str_transform.js');
} else if (typeof require !== 'undefined') {
    var str_transform = require('resource://grammalecte/graphspell/str_transform.js');
}


${map}


class DAWG {
    /*  DIRECT ACYCLIC WORD GRAPH
        This code is inspired from Steve Hanov’s DAWG, 2011. (http://stevehanov.ca/blog/index.php?id=115)

        let lWord = [];
        for (let [sFlex, iAff, iTag] of lEntry) {
            let lTemp = [];
            for (let c of sFlex) {
                lTemp.push(dChar.get(c));
            }
            lTemp.push(iAff+nChar);
            lTemp.push(iTag+nChar+nAff);
            lWord.push(lTemp);
        }
        lEntry.length = 0; // clear the array

        // Dictionary of arc values occurrency, to sort arcs of each node
        let lKeyVal = [];
        for (let c of dChar.keys()) { lKeyVal.push([dChar.get(c), dCharOccur.get(c)]); }

        this.nNextId += 1;
        return this.nNextId-1;
    },

    reset: function () {
        this.nNextId = 0;
    }
};


class DawgNode {

    constructor () {
        this.i = oNodeCounter.getId();
        this.final = false;

    convValueToHexString (nVal, nByte) {
        // nVal: value to convert, nByte: number of bytes
        let sHexVal = nVal.toString(16); // conversion to hexadecimal string
        //console.log(`value: ${nVal} in ${nByte} bytes`);
        if (sHexVal.length < (nByte*2)) {
            return "0".repeat((nByte*2) - sHexVal.length) + sHexVal;
        } else if (sHexVal.length == (nByte*2)) {
            return sHexVal;
        } else {
            throw "Conversion to byte string: value bigger than allowed.";
        }
    }
}