Grammalecte  Check-in [9c995c9f7f]

        This code is inspired from Steve Hanov’s DAWG, 2011. (http://stevehanov.ca/blog/index.php?id=115)
        We store suffix/affix codes and tags within the graph after the “real” word.
        A word is a list of numbers [ c1, c2, c3 . . . cN, iAffix, iTags]
        Each arc is an index in this.lArcVal, where are stored characters, suffix/affix codes for stemming and tags.
        Important: As usual, the last node (after ‘iTags’) is tagged final, AND the node after ‘cN’ is ALSO tagged final.
    */

    constructor (lEntrySrc, sLangCode, sLangName, sDicName, cStemming, xProgressBarNode=null) {
        console.log("===== Direct Acyclic Word Graph - Minimal Acyclic Finite State Automaton =====");
        let funcStemmingGen = null;
        switch (cStemming.toUpperCase()) {
            case "A":
                funcStemmingGen = str_transform.defineAffixCode; break;
            case "S":
                funcStemmingGen = str_transform.defineSuffixCode; break;

        let lKeyVal = [];
        for (let c of dChar.keys()) { lKeyVal.push([dChar.get(c), dCharOccur.get(c)]); }
        for (let sAff of dAff.keys()) { lKeyVal.push([dAff.get(sAff)+nChar, dAffOccur.get(sAff)]); }
        for (let sTag of dTag.keys()) { lKeyVal.push([dTag.get(sTag)+nChar+nAff, dTagOccur.get(sTag)]); }
        let dValOccur = new Map(lKeyVal);
        lKeyVal.length = 0; // clear the array

        this.sLangCode = sLangCode;
        this.sLangName = sLangName;
        this.sDicName = sDicName;
        this.nEntry = lWord.length;
        this.aPreviousEntry = [];
        oNodeCounter.reset();
        this.oRoot = new DawgNode();
        this.lUncheckedNodes = [];          // list of nodes that have not been checked for duplication.
        this.dMinimizedNodes = new Map();   // list of unique nodes that have been checked for duplication.
        this.nNode = 0;

        if (nCompressionMethod == 1) {
            sByDic = this.oRoot.convToBytes1(this.nBytesArc, this.nBytesNodeAddress);
            for (let oNode of this.dMinimizedNodes.values()) {
                sByDic += oNode.convToBytes1(this.nBytesArc, this.nBytesNodeAddress);
            }
        }
        let oJSON = {
            "sHeader": "/pyfsa/",
            "sLangCode": this.sLangCode,
            "sLangName": this.sLangName,
            "sDicName": this.sDicName,

            "sFileName": "[none]",
            "sDate": this._getDate(),




            "nEntries": this.nEntry,
            "nChar": this.nChar,
            "nAff": this.nAff,
            "nTag": this.nTag,
            "cStemming": this.cStemming,
            "dChar": helpers.mapToObject(this.dChar),
            "nNode": this.nNode,
            "nArc": this.nArc,
            "lArcVal": this.lArcVal,
            "nArcVal": this.nArcVal,
            "nCompressionMethod": nCompressionMethod,
            "nBytesArc": this.nBytesArc,
            "nBytesNodeAddress": this.nBytesNodeAddress,
            "nBytesOffset": this.nBytesOffset,
            "sByDic": sByDic    // binary word graph
        };
        return oJSON;
    },

    _getDate () {
        let oDate = new Date();
        let sMonth = (oDate.getMonth() + 1).toString().padStart(2, "0"); // Month+1: Because JS always sucks somehow.

            Object.assign(this, oData);
        }
        catch (e) {
            throw Error("# Error. File not found or not loadable.\n" + e.message + "\n");
        }
        /*
            Properties:
            sName, nCompressionMethod, sHeader, lArcVal, nArcVal, sByDic, sLang, nChar, nBytesArc, nBytesNodeAddress,
            nEntries, nNode, nArc, nAff, cStemming, nTag, dChar, nBytesOffset,
        */

        /*
            Bug workaround.
            Mozilla’s JS parser sucks. Can’t read file bigger than 4 Mb!
            So we convert huge hexadecimal string to list of numbers…
            https://github.com/mozilla/addons-linter/issues/1361
        */
        let lTemp = [];
        for (let i = 0;  i < this.sByDic.length;  i+=2) {
            lTemp.push(parseInt(this.sByDic.slice(i, i+2), 16));
        }
        this.sByDic = "";
        this.byDic = lTemp;
        //this.byDic = new Uint8Array(lTemp);  // not quicker, even slower
        /* end of bug workaround */

        if (!this.sHeader.startsWith("/pyfsa/")) {
            throw TypeError("# Error. Not a pyfsa binary dictionary. Header: " + this.sHeader);
        }
        if (!(this.nCompressionMethod == "1" || this.nCompressionMethod == "2" || this.nCompressionMethod == "3")) {
            throw RangeError("# Error. Unknown dictionary compression method: " + this.nCompressionMethod);
        }
        // <dChar> to get the value of an arc, <dCharVal> to get the char of an arc with its value
        this.dChar = helpers.objectToMap(this.dChar);
        this.dCharVal = this.dChar.gl_reverse();


        if (this.cStemming == "S") {
            this.funcStemming = str_transform.changeWordWithSuffixCode;
        } else if (this.cStemming == "A") {
            this.funcStemming = str_transform.changeWordWithAffixCode;
        } else {
            this.funcStemming = str_transform.noStemming;

        }
        //console.log(this.getInfo());
        this.bOptNumSigle = true;
        this.bOptNumAtLast = false;
    }

    getInfo () {
        return  `  Language: ${this.sLangName}   Lang code: ${this.sLangCode}   Dictionary name: ${this.sDicName}\n` +
                `  Compression method: ${this.nCompressionMethod}   Date: ${this.sDate}   Stemming: ${this.cStemming}FX\n` +
                `  Arcs values:  ${this.nArcVal} = ${this.nChar} characters,  ${this.nAff} affixes,  ${this.nTag} tags\n` +
                `  Dictionary: ${this.nEntries} entries,    ${this.nNode} nodes,   ${this.nArc} arcs\n` +
                `  Address size: ${this.nBytesNodeAddress} bytes,  Arc size: ${this.nBytesArc} bytes\n`;
    }

    isValidToken (sToken) {
        // checks if sToken is valid (if there is hyphens in sToken, sToken is split, each part is checked)

    """DIRECT ACYCLIC WORD GRAPH"""
    # This code is inspired from Steve Hanov’s DAWG, 2011. (http://stevehanov.ca/blog/index.php?id=115)
    # We store suffix/affix codes and tags within the graph after the “real” word.
    # A word is a list of numbers [ c1, c2, c3 . . . cN, iAffix, iTags]
    # Each arc is an index in self.lArcVal, where are stored characters, suffix/affix codes for stemming and tags.
    # Important: As usual, the last node (after ‘iTags’) is tagged final, AND the node after ‘cN’ is ALSO tagged final.

    def __init__ (self, spfSrc, sLangCode, sLangName, sDicName, cStemming):
        print("===== Direct Acyclic Word Graph - Minimal Acyclic Finite State Automaton =====")
        cStemming = cStemming.upper()
        if cStemming == "A":
            funcStemmingGen = st.defineAffixCode
        elif cStemming == "S":
            funcStemmingGen = st.defineSuffixCode
        elif cStemming == "N":

                        + [ (dAff[aff]+nChar, dAffOccur[aff]) for aff in dAff ] \
                        + [ (dTag[tag]+nChar+nAff, dTagOccur[tag]) for tag in dTag ] )
        #with open(spfSrc[:-8]+".valuesfreq.txt", 'w', encoding='utf-8') as hFreqDst:  # DEBUG
        #    for iKey, nOcc in sorted(dValOccur.items(), key=lambda t: t[1], reverse=True):
        #        hFreqDst.write("{}: {}\n".format(lVal[iKey], nOcc))
        #    hFreqDst.close()
        
        self.sFileName = spfSrc
        self.sLangCode = sLangCode
        self.sLangName = sLangName
        self.sDicName = sDicName
        self.nEntry = len(lWord)
        self.aPreviousEntry = []
        DawgNode.resetNextId()
        self.oRoot = DawgNode()
        self.lUncheckedNodes = []  # list of nodes that have not been checked for duplication.
        self.lMinimizedNodes = {}  # list of unique nodes that have been checked for duplication.
        self.lSortedNodes = []     # version 2 and 3

            for oNode in self.lSortedNodes:
                byDic += oNode.convToBytes3(self.nBytesArc, self.nBytesNodeAddress, self.nBytesOffset)

        with open(spfDst, "w", encoding="utf-8", newline="\n") as hDst:
            if bInJSModule:
                hDst.write('// JavaScript\n// Generated data (do not edit)\n\n"use strict";\n\nconst dictionary = ')
            hDst.write(json.dumps({
                            "sHeader": "/pyfsa/",
                            "sLangCode": self.sLangCode,
                            "sLangName": self.sLangName,
                            "sDicName": self.sDicName,

                            "sFileName": self.sFileName,
                            "sDate": str(datetime.datetime.now())[:-7],
                            "nEntries": self.nEntry,
                            "nChar": self.nChar,
                            "nAff": self.nAff,
                            "nTag": self.nTag,
                            "cStemming": self.cStemming,
                            "dChar": self.dChar,
                            "nNode": self.nNode,
                            "nArc": self.nArc,
                            "nArcVal": self.nArcVal,
                            "lArcVal": self.lArcVal,
                            "nCompressionMethod": nCompressionMethod,
                            "nBytesArc": self.nBytesArc,
                            "nBytesNodeAddress": self.nBytesNodeAddress,
                            "nBytesOffset": self.nBytesOffset
                            # JavaScript is a pile of shit, so Mozilla’s JS parser don’t like file bigger than 4 Mb!
                            # So, if necessary, we use an hexadecimal string, that we will convert later in Firefox’s extension.
                            # https://github.com/mozilla/addons-linter/issues/1361
                            "sByDic": byDic.hex()  if bBinaryDictAsHexString  else [ e  for e in byDic ],












                        }, ensure_ascii=False))
            if bInJSModule:
                hDst.write(";\n\nexports.dictionary = dictionary;\n")

    def _writeBinary (self, sPathFile, nCompressionMethod):
        """
        Format of the binary indexable dictionary:

        if not sPathFile.endswith(".bdic"):
            sPathFile += "."+str(nCompressionMethod)+".bdic"
        with open(sPathFile, 'wb') as hDst:
            # header
            hDst.write("/pyfsa/{}/".format(nCompressionMethod).encode("utf-8"))
            hDst.write(b"\0\0\0\0")
            # infos
            hDst.write("{}/{}/{}/{}/{}/{}/{}/{}/{}".format(self.sLangName, self.nChar, self.nBytesArc, self.nBytesNodeAddress, \
                                                           self.nEntry, self.nNode, self.nArc, self.nAff, self.cStemming).encode("utf-8"))
            hDst.write(b"\0\0\0\0")
            # lArcVal
            hDst.write("\t".join(self.lArcVal).encode("utf-8"))
            hDst.write(b"\0\0\0\0")
            # DAWG: nodes / arcs
            if nCompressionMethod == 1:

    def writeResults (self, sPathFile):
        bFileExits = os.path.isfile("_lexicons.res.txt")
        with open("_lexicons.res.txt", "a", encoding='utf-8', newline="\n") as hDst:
            sFormat1 = "{:<12} {:>12} {:>5} {:>8} {:>8} {:>6} {:>8} {:>9} {:>9} {:>15} {:>12} {:>12}\n"
            sFormat2 = "{:<12} {:>12,} {:>5,} {:>8,} {:>8} {:>6,} {:>8,} {:>9,} {:>9,} {:>15,} {:>12,} {:>12,}\n"
            if not bFileExits:
                hDst.write(sFormat1.format("Lexicon", "Entries", "Chars", "Affixes", "Stemming", "Tags", "Values", "Nodes", "Arcs", "Lexicon (Kb)", "Dict (Kb)", "LT Dict (Kb)"))
            hDst.write(sFormat2.format(self.sLangName, self.nEntry, self.nChar, self.nAff, self.cStemming + "FX", self.nTag, self.nArcVal, \
                                       self.nNode, self.nArc, os.path.getsize(self.sFileName), os.path.getsize(sPathFile), \
                                       os.path.getsize("cfsa/dict/{}.dict".format(self.sLangName)) if os.path.isfile("cfsa/dict/{}.dict".format(self.sLangName)) else 0))
            hDst.close()



class DawgNode:
    NextId = 0
    NextPos = 1 # (version 2)

        else:
            raise ValueError("  # Error: unknown code: {}".format(self.nCompressionMethod))

        self.bOptNumSigle = False
        self.bOptNumAtLast = False

    def getInfo (self):
        return  "  Language: {0.sLangName}   Lang code: {0.sLangCode}   Dictionary name: {0.sDicName}" \
                "  Compression method: {0.nCompressionMethod:>2}   Date: {0.sDate}   Stemming: {0.cStemming}FX\n" \
                "  Arcs values:  {0.nArcVal:>10,} = {0.nChar:>5,} characters,  {0.nAff:>6,} affixes,  {0.nTag:>6,} tags\n" \
                "  Dictionary: {0.nEntries:>12,} entries,    {0.nNode:>11,} nodes,   {0.nArc:>11,} arcs\n" \
                "  Address size: {0.nBytesNodeAddress:>1} bytes,  Arc size: {0.nBytesArc:>1} bytes\n".format(self)

    def writeAsJSObject (self, spfDest, bInJSModule=False, bBinaryDictAsHexString=False):
        "write IBDAWG as a JavaScript object in a JavaScript module"
        import json