Grammalecte  Diff

import metagraphe
import metaphone2


# Dictionnaire des caractères pour le tri naturel.
# Ordre souhaitable, mais pose problème pour la recherche, car engendre des égalités de lemmes différents.
# Il faut donc travailler sur un dictionnaire trié *numériquement* et le sauvegarder selon le tri *naturel*             
# Il faut donc travailler sur un dictionnaire trié *numériquement* et le sauvegarder selon le tri *naturel*
CHARMAP = str.maketrans({ 'à': 'a',  'À': 'A',  'â': 'a',  'Â': 'A',  'ä': 'a',  'Ä': 'A',  'å': 'a',  'Å': 'A',  'ā': 'a',  'Ā': 'A',
                          'ç': 'c',  'Ç': 'C',
                          'é': 'e',  'É': 'E',  'è': 'e',  'È': 'E',  'ê': 'e',  'Ê': 'E',  'ë': 'e',  'Ë': 'E',  'ē': 'e',  'Ē': 'E',
                          'î': 'i',  'Î': 'I',  'ï': 'i',  'Ï': 'I',  'ī': 'i',  'Ī': 'I',
                          'ñ': 'n',
                          'ô': 'o',  'Ô': 'O',  'ö': 'o',  'Ö': 'O',  'ō': 'o',  'Ō': 'O',
                          'ù': 'u',  'Ù': 'U',  'û': 'u',  'Û': 'U',  'ü': 'u',  'Ü': 'U',  'ū': 'u',  'Ū': 'U',

        # Affixes
        self.sSettings = '' # enregistre tout avant la ligne # END
        self.dFlags = collections.OrderedDict()
        self.bShortenTags = False
        self.dAM = collections.OrderedDict() # étiquettes morphologiques
        self.dAF = collections.OrderedDict() # étiquettes drapeaux
        # Flexions
        self.lFlexions = []           # liste des flexions avec lemme, morphologie et occurrences 
        self.lFlexions = []           # liste des flexions avec lemme, morphologie et occurrences
        self.lStatsLex = []
        self.nTotOccurRecognizedWords = 0
        self.aFlexions = None
    

    def readDictionary (self, spf):
        "Lecture du dictionnaire"
        echo('Dictionnaire << [ {} ]'.format(spf), end=' ')
        for sLine in readfile(spf):
            sLine = sLine.strip()
            if not sLine.isdigit() and not sLine.startswith("#"):
                self.lEntry.append(Entree(sLine))

                dAF[oEntry.flags] = dAF.get(oEntry.flags, 0) + 1
            sMorph = oEntry.getMorph(nMode).strip()
            if sMorph:
                dAM[sMorph] = dAM.get(sMorph, 0) + 1

        lAF = sorted(dAF.items(), key = lambda x: (x[1], x[0]), reverse=True)
        lAM = sorted(dAM.items(), key = lambda x: (x[1], x[0]), reverse=True)
        

        with open(spDst, 'a', encoding='utf-8', newline="\n") as hDst:
            hDst.write("\n\nDrapeaux :\n")
            for nAF, elem in enumerate(lAF, 1):
                self.dAF[elem[0]] = str(nAF)
                hDst.write("  > {0[1]:>8} : {0[0]}\n".format(elem))
            hDst.write("\n\nMorphologies :\n")
            for nAM, elem in enumerate(lAM, 1):
                self.dAM[elem[0]] = str(nAM)
                hDst.write("  > {0[1]:>8} : {0[0]}\n".format(elem))

    def writeDictionary (self, spDst, dTplVars, nMode, bSimplified):
        "Écrire le fichier dictionnaire (.dic)"
        echo(' * Dictionnaire >> [ {}.dic ] ({})'.format(dTplVars['asciiName'], dTplVars['subDicts']))
        nEntry = 0
        for oEntry in self.lEntry:
            if oEntry.di in dTplVars['subDicts']:
            if oEntry.di in dTplVars['subDicts'] and " " not in oEntry.lemma:
                nEntry += 1
        with open(spDst+'/'+dTplVars['asciiName']+'.dic', 'w', encoding='utf-8', newline="\n") as hDst:
            hDst.write(str(nEntry)+"\n")
            for oEntry in self.lEntry:
                if oEntry.di in dTplVars['subDicts']:
                    hDst.write(oEntry.getEntryLine(self, nMode, bSimplified))
    
                if oEntry.di in dTplVars['subDicts'] and " " not in oEntry.lemma:
                    hDst.write(oEntry.getHunspellLine(self, nMode, bSimplified))

    def writeAffixes (self, spDst, dTplVars, nMode, bSimplified):
        "Écrire le fichier des affixes (.aff)"
        echo(' * Dictionnaire >> [ {}.aff ]'.format(dTplVars['asciiName']))
        info = "# This Source Code Form is subject to the terms of the Mozilla Public\n" + \
               "# License, v. 2.0. If a copy of the MPL was not distributed with this\n" + \
               "# file, You can obtain one at http://mozilla.org/MPL/2.0/.\n\n" + \
               "# AFFIXES DU {} v{}\n".format(dTplVars['name'], self.sVersion) + \
               "# par Olivier R. -- licence MPL 2.0\n" + \
               "# Généré le " + time.strftime("%d-%m-%Y à %H:%M") + "\n" \
               "# Pour améliorer le dictionnaire, allez sur http://www.dicollecte.org/\n\n"
               

        with open(spDst+'/'+dTplVars['asciiName']+'.aff', 'w', encoding='utf-8', newline="\n") as hDst:
            hDst.write(info)
            hDst.write(self.sSettings + "\n")
            if self.bShortenTags:
                hDst.write("AM {}\n".format(len(self.dAM)))
                for item in self.dAM.items():
                    hDst.write("AM {}\n".format(item[0]))

    def sortEntriesNatural (self):
        echo(' * Dictionnaire - Tri naturel des entrées...')
        self.lEntry = sorted(self.lEntry, key=Entree.keyTriNat)

    def sortEntriesNumerical (self):
        echo(' * Dictionnaire - Tri numérique des entrées...')
        self.lEntry = sorted(self.lEntry, key=Entree.keyTriNum)        
        self.lEntry = sorted(self.lEntry, key=Entree.keyTriNum)

    def sortLexiconByFlexion (self):
        echo(' * Dictionnaire - tri du lexique (par flexion)...')
        self.lFlexions = sorted(self.lFlexions, key=Flexion.keyFlexion)

    def sortLexiconByFreq (self):
        echo(' * Dictionnaire - tri du lexique (par fréquence)...')

                d[oFlex.sFlexion] = [oFlex.oEntry]
        for oFlex in self.lFlexions:
            oFlex.lMulti = list(d[oFlex.sFlexion])
            oFlex.nMulti = len(oFlex.lMulti)
        for oFlex in self.lFlexions:
            oFlex.lMulti.remove(oFlex.oEntry)
            oFlex.nMulti -= 1
        

    def setTagsFrom (self, other):
        echo(' * Dictionnaire - copie des tags...')
        for i in range(self.nEntry):
            for oEntry in other.lEntry:
                if self.lEntry[i].lemma == oEntry.lemma and self.lEntry[i].flags == oEntry.flags:
                    self.lEntry[i].setTagsFrom(oEntry)

    def calculateStats (self, oStatsLex, spfDst):
        echo(" * Dictionnaire - calculs...")
        with open(spfDst, 'w', encoding='utf-8', newline="\n") as hDst:
            # Occurrences brutes des formes fléchies
            echo("   comptage des occurrences...")
            hDst.write(oStatsLex.getInfo())
            for oFlex in self.lFlexions:
                oFlex.setOccur(oStatsLex.getFlexionOccur(oFlex.sFlexion))
            self.nTotOccurRecognizedWords = 0
            for oFlex in self.lFlexions:
                oFlex.calcOccur()
                self.nTotOccurRecognizedWords += oFlex.nOccur
            

            # Report des occurrences
            echo("   report des occurrences des formes fléchies multiples...")
            hDst.write("Report des occurrences des formes fléchies multiples :\n")
            hDst.write("  Légende :\n")
            hDst.write("    >>   le nombre d’occurrences de la flexion est ramené à la moyenne.\n")
            hDst.write("    +>   le nombre d’occurrences de la flexion est augmenté avec le surplus d’occurrences des flexions ramenées à la moyenne.\n")
            hDst.write("    %>   le nombre d’occurrences de la flexion est pondéré avec le poids de la moyenne de l’entrée.\n\n")

            for oEntry in self.lEntry:
                oEntry.calcOccurFromFlexions()
                oEntry.calcAverageKnownOccurrence()
                oEntry.solveOccurMultipleFlexions(hDst, oStatsLex)
                oEntry.calcOccurFromFlexions()
            

            # Fréquences
            echo("   calcul des fréquences et indices de fréquence...")
            for oFlex in self.lFlexions:
                oFlex.calcFreq(self.nTotOccurRecognizedWords)
            for oEntry in self.lEntry:
                oEntry.calcFreq(self.nTotOccurRecognizedWords)
            

            # Entrées, statistiques
            echo("   statistiques...")
            hDst.write("\n\nNatures grammaticales :\n")
            d = {}
            for oEntry in self.lEntry:
                po = re.sub("(?<=v[0-3])[itnpqrmaezx_]+", "", oEntry.po)
                d[po] = d.get(po, 0) + 1
            for e in sorted(d.items(), key = lambda x: (x[1], x[0]), reverse=True):
                hDst.write(" * {0[1]:<15} : {0[0]}\n".format(e))
            

            hDst.write("\n\nVentilation des entrées par indice de fréquence :\n")
            d1 = {}
            d2 = {}
            for oEntry in self.lEntry:
                d1[oEntry.fq] = d1.get(oEntry.fq, 0) + 1
                d2[oEntry.fq] = d2.get(oEntry.fq, 0) + oEntry.fFreq
            for k in sorted(d1.keys()):
                hDst.write(" * {} : {} entrées ({:.2f} %)  → {:.9f} %\n".format(k, d1[k], (d1[k]*100)/self.nEntry, d2[k]))
                    

            hDst.write("\n\nRépartition des entrées par sous-dictionnaire :\n")
            d = {}
            for oEntry in self.lEntry:
                d[oEntry.di] = d.get(oEntry.di, 0) + 1
            for sKey, nVal in d.items():
                hDst.write(" * {0:<15} : {1} entrées ({2:.2f} %)\n".format(dSUBDIC[sKey], nVal, (nVal*100)/self.nEntry))
            

            # Occurrences des lettres
            echo("   occurrences des lettres...")
            d = {}
            for oFlex in self.lFlexions:
                for c in oFlex.sFlexion:
                    d[c] = d.get(c, 0) + oFlex.nOccur
            nTot = 0

            hDst.write("\n\nNombre de formes fléchies : {}\n".format(len(self.lFlexions)))
            hDst.write("\n\nNombre de graphies : {}\n".format(len(self.aFlexions)))

    def calcMetagraphe (self):
        echo(" * Lexique - Metagraphe")
        for oFlex in self.lFlexions:
            oFlex.calcMetagraphe()
    

    def calcMetaphone2 (self):
        echo(" * Lexique - Metaphone 2")
        for oFlex in self.lFlexions:
            oFlex.calcMetaphone2()
    

    def createNgrams (self, spDest, n):
        echo(" * Lexique - Ngrams " + str(n))
        if n < 2:
            echo("erreur: n = " + str(n))
            return
        dOccur = {} # ngram:n
        dRefW = {} # ngram:set(idx)

        file_util.copy_file('_templates/ooo/french_flag.png', spExt)
        file_util.copy_file('_templates/ooo/french_flag_16.bmp', spExt+'/ui')
        copyTemplate('_templates/ooo', spExt, 'description.xml', dTplVars)
        copyTemplate('_templates/ooo', spExt, 'dictionaries.xcu', dTplVars)
        #file_util.copy_file('_templates/ooo/dictionaries.xcu.tpl.xml', spExt)
        copyTemplate('_templates/ooo', spExt, 'package-description.txt', dTplVars)
        for dVars in lDictVars:
            dicPath = spBuild + '/' + PREFIX_DICT_PATH + self.sVersion 
            dicPath = spBuild + '/' + PREFIX_DICT_PATH + self.sVersion
            file_util.copy_file(dicPath+'/'+dVars['asciiName']+'.dic', spExt+'/dictionaries/'+dVars['asciiName']+'.dic')
            file_util.copy_file(dicPath+'/'+dVars['asciiName']+'.aff', spExt+'/dictionaries/'+dVars['asciiName']+'.aff')
        copyTemplate('orthographe', spExt+'/dictionaries', 'README_dict_fr.txt', dTplVars)
        # thesaurus
        file_util.copy_file('thesaurus/thes_fr.dat', spExt+'/dictionaries')
        file_util.copy_file('thesaurus/thes_fr.idx', spExt+'/dictionaries')
        file_util.copy_file('thesaurus/README_thes_fr.txt', spExt+'/dictionaries')
        # hyphenation
        file_util.copy_file('césures/hyph_fr.dic', spExt+'/dictionaries')
        file_util.copy_file('césures/hyph_fr.iso8859-1.dic', spExt+'/dictionaries')
        file_util.copy_file('césures/frhyph.tex', spExt+'/dictionaries')
        file_util.copy_file('césures/hyph-fr.tex', spExt+'/dictionaries')
        file_util.copy_file('césures/README_hyph_fr-3.0.txt', spExt+'/dictionaries')
        file_util.copy_file('césures/README_hyph_fr-2.9.txt', spExt+'/dictionaries')
        # zip
        createZipFiles(spExt, spBuild, sExtensionName + '.oxt')
        # copy to Grammalecte Project
        if spDestGL:
            echo("   extension copiée dans Grammalecte...")
            dir_util.copy_tree(spExt+'/dictionaries', spDestGL)
    

    def createMozillaExtensions (self, spBuild, dTplVars, lDictVars, spDestGL=""):
        # Mozilla extension 1
        echo(" * Dictionnaire >> extension pour Mozilla")
        dTplVars['version'] = self.sVersion
        sExtensionName = EXT_PREFIX_MOZ + self.sVersion
        spExt = spBuild + '/' + sExtensionName
        dir_util.mkpath(spExt+'/dictionaries')
        copyTemplate('_templates/moz', spExt, 'install.rdf', dTplVars)
        spDict = spBuild + '/' + PREFIX_DICT_PATH + self.sVersion
        file_util.copy_file(spDict+'/fr-classique.dic', spExt+'/dictionaries/fr-classic.dic')
        file_util.copy_file(spDict+'/fr-classique.aff', spExt+'/dictionaries/fr-classic.aff')
        copyTemplate('orthographe', spExt, 'README_dict_fr.txt', dTplVars)
        createZipFiles(spExt, spBuild, sExtensionName + '.xpi')
        # Grammalecte
        if spDestGL:
            echo(" * Dictionnaire >> copie des dicos dans Grammalecte")
            for dVars in lDictVars:
                file_util.copy_file(spDict+'/'+dVars['asciiName']+'.dic', spDestGL+'/'+dVars['mozAsciiName']+"/"+dVars['mozAsciiName']+'.dic')
                file_util.copy_file(spDict+'/'+dVars['asciiName']+'.aff', spDestGL+'/'+dVars['mozAsciiName']+"/"+dVars['mozAsciiName']+'.aff')
    

    def createFileIfqForDB (self, spBuild):
        echo(" * Dictionnaire >> indices de fréquence pour la DB...")
        with open(spBuild+'/dictIdxIfq-'+self.sVersion+'.diff.txt', 'w', encoding='utf-8', newline="\n") as hDiff, \
             open(spBuild+'/dictIdxIfq-'+self.sVersion+'.notes.txt', 'w', encoding='utf-8', newline="\n") as hNotes:
            for oEntry in self.lEntry:
                if oEntry.fq != oEntry.oldFq:
                    hDiff.write("{0.iD}\t{0.fq}\n".format(oEntry))
                    hNotes.write("{0.lemma}/{0.flags}\t{0.oldFq} > {0.fq}\n".format(oEntry))
        

    def createLexiconPackages (self, spBuild, version, oStatsLex, spDestGL=""):
        sLexName = LEX_PREFIX + version
        spLex = spBuild + '/' + sLexName
        dir_util.mkpath(spLex)
        # write Dicollecte lexicon
        self.sortLexiconByFreq()
        self.writeLexicon(spLex + '/' + sLexName + '.txt', version, oStatsLex)

        self.iD = '0'

        # autres
        self.comment = ''
        self.err = ''
        self.nFlexions = 0
        self.lFlexions = []
        self.sRadical = ''
        self.sStem = ''
        self.nOccur = 0
        self.nAKO = -1   # Average known occurrences
        self.fFreq = 0
        self.oldFq = ''
        

        sLine = sLine.rstrip(" \n")
        # commentaire
        if '#' in sLine:
            sLine, comment = sLine.split('#', 1)
            self.comment = comment.strip()
        # éléments de la ligne
        elems = sLine.split()
        elems = sLine.split("\t")
        nElems = len(elems)
        # lemme et drapeaux
        firstElems = elems[0].split('/')
        self.lemma = firstElems[0]
        self.flags = firstElems[1]  if len(firstElems) > 1  else ''
        # morph
        for i in range(1, nElems):

                else:
                    echo('  ## Champ inconnu: {}  dans  {}/{}'.format(fields[0], self.lemma, self.flags))
            else:
                self.err = self.err + elems[i]
        if self.err:
            echo("\n## Erreur dans le dictionnaire : {}".format(self.err))
            echo("   dans : " + self.lemma)
                

    def __str__ (self):
        return "{0.lemma}/{0.flags} {1}".format(self, self.getMorph(2))

    def check (self):
        sErr = ''
        if self.lemma == '':
            sErr += 'lemme vide'
        if not re.match(r"[a-zA-ZéÉôÔàâáÂîÎïèÈêÊÜœŒæÆçÇ0-9µåÅΩ&αβγδεζηθικλμνξοπρστυφχψωΔℓΩ_]", self.lemma):
            sErr += 'premier caractère inconnu: ' + self.lemma[0]
        if re.search(r"\s$", self.lemma):
            sErr += 'espace en fin de lemme'
        if re.match(r"v[0123]", self.po) and not re.match(r"[eas_][ix_][tx_][nx_][pqreuvx_][mx_][ex_z][ax_z]\b", self.po[2:]):
            sErr += 'verbe inconnu: ' + self.po
        if (re.match(r"S[*.]", self.flags) and re.search("[sxz]$", self.lemma)) or (re.match(r"X[*.]", self.flags) and not re.search("[ul]$", self.lemma)):
            sErr += 'drapeau inutile'
        if self.iz == '' and re.match(r"[SXAI](?!=)", self.flags) and self.po:
            sErr += '[is]'
            sErr += '[is] vide'
        if re.match(r"pl|sg|inv", self.iz):
            sErr += '[is]'
            sErr += '[is] incomplet'
        if re.match(r"[FW]", self.flags) and re.search(r"epi|mas|fem|inv|sg|pl", self.iz):
            sErr += '[is]'
            sErr += '[is] incohérent'
        if re.match(r"[FW]", self.flags) and re.search(r"[^eë]$", self.lemma):
            sErr += "fin de lemme inapproprié"
        if re.match(r".\*", self.flags) and re.match(r"[bcdfgjklmnpqrstvwxz]", self.lemma):
            sErr += 'drapeau pour lemme commençant par une voyelle'
        if re.search(r"pl|sg|inv", self.iz) and re.match(r"[SXAIFW](?!=)", self.flags):
            sErr += '[is]'
            sErr += '[is] incohérent'
        if re.search(r"nom|adj", self.po) and re.match(r"(?i)[aâàäáeéèêëiîïíìoôöóòuûüúù]", self.lemma) and re.match("[SFWXAI][.]", self.flags) \
           and "pel" not in self.lx:
            sErr += 'le drapeau derait finir avec *'
        if not self.flags and self.iz.endswith(("mas", "fem", "epi")):
            sErr += '[is] incomplet'
        if self.flags.startswith(("a", "b", "c", "d")) and not self.lemma.endswith("er"):
            sErr += "drapeau pour verbe du 1ᵉʳ groupe sur un lemme non conforme"

    def keyTriNat (self):
        return (self.lemma.translate(CHARMAP), self.flags, self.po)

    def keyTriNum (self):
        return (self.lemma, self.flags, self.po)

    def getEntryLine (self, oDict, nMode, bSimplified=False):    
        sLine = self.lemma
    def getHunspellLine (self, oDict, nMode, bSimplified=False):
        sLine = self.lemma.replace("’", "'")
        if self.flags:
            sLine += '/'
            sLine += self.flags  if not oDict.bShortenTags or bSimplified  else oDict.dAF[self.flags]
        if bSimplified:
            return sLine.replace("()", "") + "\n"
        if nMode > 0:
            sMorph = self.getMorph(nMode)

                if not sMorph.endswith((" mas", " fem", " epi")):
                    self.lFlexions.append( Flexion(self, sFlex, sMorph, sDic) )
                    self.nFlexions += 1
                else:
                    #echo(sFlex + " " + sMorph + ", ")
                    pass
        # Drapeaux dont le lemme féminin doit être remplacé par le masculin dans la gestion des formes fléchies
        if self.st:
            self.sStem = self.st
        else:
        if self.flags.startswith(("F.", "F*", "W.", "W*")):
            # recherche de la forme masculine
            for t in lTuples:
                sMorph = self.clean(t[1])
                if sMorph.endswith('mas') or sMorph.endswith('mas sg') or sMorph.endswith('mas inv'): 
                    self.sRadical = t[0]
        else:
            self.sRadical = self.lemma
            if self.flags.startswith(("F.", "F*", "W.", "W*")):
                # recherche de la forme masculine
                for t in lTuples:
                    sMorph = self.clean(t[1])
                    if sMorph.endswith(('mas', 'mas sg', 'mas inv')):
                        self.sStem = t[0]
            else:
                self.sStem = self.lemma
        # Tag duplicates
        d = {}
        for oFlex in self.lFlexions:
            d[oFlex.sFlexion] = d.get(oFlex.sFlexion, 0) + 1
        for oFlex in self.lFlexions:
            oFlex.nDup = d[oFlex.sFlexion]

                                        lFlexions.append( (oRule.add+flex[0], flex[1]+ruleMorph) )
                                else:
                                    lFlexions.append(flexion)
                            else:
                                flexion = (self.lemma.replace(oRule.cut, oRule.add, 1), ruleMorph+morph, oRule.di)
                                if oFlag.bMix:
                                    lFlexPrefix.append(flexion)
                                    for flex in lFlexSuffix: 
                                    for flex in lFlexSuffix:
                                        lFlexions.append( (flex[0].replace(oRule.cut, oRule.add, 1), flex[1]+ruleMorph) )
                                else:
                                    lFlexions.append(flexion)
                            if oRule.flags != '' and oRule.flags != '**':
                                lFlexions.extend(Entree(flexion[0]+'/'+oRule.flags)._flechir(dFlags, flexion[1], iPR+1))
                else:
                    # cas des suffixes

    def calcOccurFromFlexions (self):
        self.nOccur = 0
        for o in self.lFlexions:
            self.nOccur += o.nOccur

    def calcAverageKnownOccurrence (self):
        # nous calculons la moyenne des occurrences des formes fléchies
        # qui n’ont pas d’équivalent dans les autres entrées (nMulti = 0) 
        # qui n’ont pas d’équivalent dans les autres entrées (nMulti = 0)
        nOccur = 0
        nFlex = 0
        for oFlex in self.lFlexions:
            if oFlex.nMulti == 0:
                nOccur += oFlex.nOccur
                nFlex += 1
        # moyenne des formes fléchies sans équivalent ou -1
        self.nAKO = math.ceil(nOccur / nFlex)  if nFlex > 0  else -1
    

    def solveOccurMultipleFlexions (self, hDst, oStatsLex):
        sBlank = "           "
        if self.nAKO >= 0:
            for oFlex in self.lFlexions:
                if oFlex.nMulti > 0 and not oFlex.bBlocked:
                    # on trie les entrées avec AKO et sans AKO
                    lEntWithAKO = []
                    lEntNoAKO = []
                    for oEntry in oFlex.lMulti:
                        if oEntry.nAKO >= 0:
                            lEntWithAKO.append(oEntry)
                        else:
                            lEntNoAKO.append(oEntry)
                    

                    if lEntNoAKO:
                        # on calcule la différence totale occasionnée par du passage des flexions appartenant à des entrées avec AKO au niveau AKO
                        nDiff = (oFlex.nOccur - self.nAKO) * oFlex.nDup
                        for oEntry in lEntWithAKO:
                            for oFlexM in oEntry.lFlexions:
                                if oFlex.sFlexion == oFlexM.sFlexion:
                                    nDiff += oFlexM.nOccur - oEntry.nAKO

                                        oFlexM.setOccurAndBlock(nNewOccur)
                    else:
                        # Toutes les entrées sont avec AKO : on pondère
                        nFlexOccur = oStatsLex.getFlexionOccur(oFlex.sFlexion)
                        nTotAKO = self.nAKO
                        for oEnt in oFlex.lMulti:
                            nTotAKO += oEnt.nAKO
                        

                        hDst.write(" = {0.sFlexion}\n".format(oFlex))
                        hDst.write("       moyennes connues\n")
                        for oFlexD in self.lFlexions:
                            if oFlex.sFlexion == oFlexD.sFlexion:
                                nNewOccur = math.ceil((nFlexOccur * (self.nAKO / nTotAKO)) / oFlexD.nDup)  if nTotAKO  else 0
                                hDst.write(sBlank + "{2:<30} {0.sMorph:<30}  {0.nOccur:>10}  %> {1:>10}\n".format(oFlexD, nNewOccur, self.getShortDescr()))
                                oFlexD.setOccurAndBlock(nNewOccur)
                        for oEntry in oFlex.lMulti:
                            for oFlexM in oEntry.lFlexions:
                                if oFlex.sFlexion == oFlexM.sFlexion:
                                    nNewOccur = math.ceil((nFlexOccur * (oEntry.nAKO / nTotAKO)) / oFlexM.nDup)  if nTotAKO  else 0
                                    hDst.write(sBlank + "{2:<30} {0.sMorph:<30}  {0.nOccur:>10}  %> {1:>10}\n".format(oFlexM, nNewOccur, oEntry.getShortDescr()))
                                    oFlexM.setOccurAndBlock(nNewOccur)
        

    def calcFreq (self, nTot):
        self.fFreq = (self.nOccur * 100) / nTot
        self.oldFq = self.fq
        self.fq = getIfq(self.fFreq)

        self.nDup    = 0    # duplicates in the same entry
        self.nMulti  = 0    # duplicates with other entries
        self.lMulti  = []   # list of similar flexions
        self.fFreq   = 0
        self.cFq     = ''
        self.metagfx = ''   # métagraphe
        self.metaph2 = ''   # métaphone 2
    

    def setOccur (self, n):
        self.nOccur = n

    def setOccurAndBlock (self, n):
        self.nOccur = n
        self.bBlocked = True

    def calcOccur (self):
        self.nOccur = math.ceil((self.nOccur / (self.nMulti+1)) / self.nDup)
    

    def calcFreq (self, nTot):
        self.fFreq = (self.nOccur * 100) / nTot
        self.cFq = getIfq(self.fFreq)
    

    def calcMetagraphe (self):
        t = metagraphe.getMetagraphe(self.sFlexion, self.sMorph)
        self.metagfx = t[0]  if not t[1]  else t[0]+"/"+t[1]

    def calcMetaphone2 (self):
        t = metaphone2.dm(self.sFlexion)
        self.metaph2 = t[0]  if not t[1]  else t[0]+"/"+t[1]

    @classmethod
    def header (cls, oStatsLex):
        sOccurs = ''
        for t in oStatsLex.lLex:
            sOccurs += t[1] + "\t"
        return "id\tFlexion\tLemme\tÉtiquettes\tMétagraphe (β)\tMetaphone2\tNotes\tSémantique\tÉtymologie\tSous-dictionnaire\t" + sOccurs + "Total occurrences\tDoublons\tMultiples\tFréquence\tIndice de fréquence\n"

    def __str__ (self, oStatsLex):
        sOccurs = ''
        for v in oStatsLex.dFlexions[self.sFlexion]:
            sOccurs += str(v) + "\t"
        return "{0.oEntry.iD}\t{0.sFlexion}\t{0.oEntry.sRadical}\t{0.sMorph}\t{0.metagfx}\t{0.metaph2}\t{0.oEntry.lx}\t{0.oEntry.se}\t{0.oEntry.et}\t{0.oEntry.di}{2}\t{1}{0.nOccur}\t{0.nDup}\t{0.nMulti}\t{0.fFreq:.15f}\t{0.cFq}\n".format(self, sOccurs, "/"+self.cDic if self.cDic != "*" else "")
        return "{0.oEntry.iD}\t{0.sFlexion}\t{0.oEntry.sStem}\t{0.sMorph}\t{0.metagfx}\t{0.metaph2}\t{0.oEntry.lx}\t{0.oEntry.se}\t{0.oEntry.et}\t{0.oEntry.di}{2}\t{1}{0.nOccur}\t{0.nDup}\t{0.nMulti}\t{0.fFreq:.15f}\t{0.cFq}\n".format(self, sOccurs, "/"+self.cDic if self.cDic != "*" else "")

    @classmethod
    def simpleHeader (cls):
        return "# :POS ;LEX ~SEM =FQ /DIC\n"

    def getGrammarCheckerRepr (self):
        return "{0.sFlexion}\t{0.oEntry.lemma}\t{1}\n".format(self, self._getSimpleTags())

        "ipre": ":Ip", "iimp": ":Iq", "ipsi": ":Is", "ifut": ":If",
        "spre": ":Sp", "simp": ":Sq", "cond": ":K", "impe": ":E",
        "1sg": ":1s", "1isg": ":1ś", "1jsg": ":1ŝ", "2sg": ":2s", "3sg": ":3s", "1pl": ":1p", "2pl": ":2p", "3pl": ":3p", "3pl!": ":3p!",
        "prepv": ":Rv", "prep": ":R", "loc.prep": ":Ŕ",
        "detpos": ":Dp", "detdem": ":Dd", "detind": ":Di", "detneg": ":Dn", "detex": ":De", "det": ":D",
        "advint": ":U",
        "prodem": ":Od", "proind": ":Oi", "proint": ":Ot", "proneg": ":On", "prorel": ":Or", "proadv": ":Ow",
        "properobj": ":Oo", "propersuj": ":Os", "1pe": ":O1", "2pe": ":O2", "3pe": ":O3",
        "properobj": ":Oo", "propersuj": ":Os", "1pe": ":O1", "2pe": ":O2", "3pe": ":O3", "preverb": "Ov",
        "cjco": ":Cc", "cjsub": ":Cs", "cj": ":C", "loc.cj": ":Ĉ", "loc.cjsub": ":Ĉs",
        "prn": ":M1", "patr": ":M2", "loc.patr": ":Ḿ2", "npr": ":MP", "nompr": ":NM",
        "pfx": ":Zp", "sfx": ":Zs",
        "div": ":H",
        "err": ":#",
        # LEX
        "symb": ";S"

            s += "/" + self.oEntry.di
        return s

    def keyTriNat (self):
        return (self.sFlexion.translate(CHARMAP), self.sMorph)

    def keyFreq (self):
        return (100-self.fFreq, self.oEntry.sRadical, self.sFlexion)
        return (100-self.fFreq, self.oEntry.sStem, self.sFlexion)

    def keyOcc (self):
        return (self.nOccur, self.oEntry.sRadical, self.sFlexion)
        
        return (self.nOccur, self.oEntry.sStem, self.sFlexion)

    def keyIdx (self):
        return self.oEntry.iD

    def keyFlexion (self):
        return self.sFlexion



class Flag:
    def __init__ (self, sFlagType, sFlagName, sMix):
        self.sFlagName = sFlagName
        self.bSfx = True  if sFlagType == 'SFX'  else False
        self.bMix = True  if sMix == 'Y'  else False
        self.lRules = []
        self.nRules = 0
        self.nOccur = 0
        

    def addAffixRule (self, line):
        "ajoute une règle au drapeau"
        oRule = AffixRule(line)
        self.lRules.append(oRule)
        self.nRules += 1

    def getFlag (self, subDicts, oDict, nMode, bSimplified):

        # champs de Dicollecte
        self.lx = ''
        self.di = '*'
        # erreurs
        self.err = ''
        # autres champs
        self.nOccur = 0
        

        sLine = sLine.rstrip(" \n")
        # commentaire
        if '#' in sLine:
            sLine, comment = sLine.split('#', 1)
            self.comment = comment.strip()
        # éléments de la ligne
        elems = sLine.split()

                    self.lx = fields[1]  if self.lx == ''  else self.lx + ' ' + fields[1]
                elif fields[0] == 'di':
                    self.di = fields[1]
                else:
                    echo('Champ inconnu: {}  dans  {}'.format(fields[0], self.sFlagName))
            else:
                echo("  # Erreur affixe : {}".format(line))
    

    def isReplicationRule (self):
        "is this rule used for replication of a virtual lemma"
        return self.flags == "" and ((self.cut == "0" and self.add == "") or self.cut == self.add)

    def getRuleLine (self, oDict, nMode, bSimplified=False):
        sLine = 'SFX'  if self.bSfx  else 'PFX'
        sLine += ' ' + self.sFlagName + ' ' + self.cut + ' '
        sLine += self.add  if self.add  else '0'
        if self.flags != '':
            sLine += '/'
            sLine += self.flags  if not oDict.bShortenTags or bSimplified  else oDict.dAF[self.flags]
            if bSimplified:
                sLine = sLine.replace("()", "")
        sLine += ' ' + self.cond
        if not bSimplified and nMode > 0:
            sMorph = self.getMorph(nMode)
            if sMorph:
                sLine += sMorph  if not oDict.bShortenTags or bSimplified  else ' ' + oDict.dAM[sMorph.strip()]
        return sLine + "\n"
    

    def getMorph (self, nMode):
        # morphology for Hunspell
        txt = ''
        if self.po: txt += fieldToHunspell('po', self.po)
        if self.iz: txt += fieldToHunspell('is', self.iz)
        if self.ds: txt += fieldToHunspell('ds', self.ds)
        if self.ts: txt += fieldToHunspell('ts', self.ts)

class StatsLex:
    def __init__ (self, oDict):
        echo("Lexique statistique")
        self.dFlexions = { oFlex.sFlexion: []  for oFlex in oDict.lFlexions }
        self.lLex = []
        

    def addLexFromFile (self, sPathFile, cLexID, sLexName):
        if not os.path.isfile(sPathFile):
            echo(' * Lexique statistique - fichier {} introuvable'.format(sPathFile))
            return None
        if len(cLexID) != 1:
            echo(' * Lexique statistique - fichier {} - identifiant incorrect, 1 caractère requis'.format(sPathFile))
            return None

                hDst.write(str(t)+"\n")
            for e in self.dFlexions.items():
                hDst.write("{} - {}\n".format(e[0], e[1]))



def main ():

    xParser = argparse.ArgumentParser()
    xParser.add_argument("-v", "--verdic", help="set dictionary version, i.e. 5.4", type=str, default="X.Y.z")
    xParser.add_argument("-m", "--mode", help="0: no tags,  1: Hunspell tags (default),  2: All tags", type=int, choices=[0, 1, 2], default=1)
    xParser.add_argument("-u", "--uncompress", help="do not use Hunspell compression", action="store_true")
    xParser.add_argument("-s", "--simplify", help="no virtual lemmas", action="store_true")
    xParser.add_argument("-sv", "--spellvariants", help="generate spell variants", action="store_true")
    xParser.add_argument("-gl", "--grammalecte", help="copy generated files to Grammalecte folders", action="store_true")
    xArgs = xParser.parse_args()

    if xArgs.simplify:
        xArgs.mode = 0
        xArgs.uncompress = True

    echo("Python: " + sys.version)
    echo("Version: " + xArgs.verdic)
    echo("Simplify: " + str(xArgs.simplify))
    echo("Mode: " + str(xArgs.mode))
    echo("Compression: " + str(not(xArgs.uncompress)))
    

    ### création du répertoire
    spBuild = BUILD_PATH + '/' + xArgs.verdic
    dir_util.mkpath(spBuild)
    

    ### Lecture des fichiers et création du dictionnaire
    oFrenchDict = Dictionnaire(xArgs.verdic, "French dictionary")
    for sFile in ['orthographe/FRANCAIS.dic']:
        oFrenchDict.readDictionary(sFile)
    oFrenchDict.readAffixes('orthographe/FRANCAIS_5.aff')
    

    ### Contrôle
    oFrenchDict.sortEntriesNatural()
    oFrenchDict.checkEntries()
    

    ### Lexique
    oFrenchDict.generateFlexions()
    oFrenchDict.calcMetagraphe()
    oFrenchDict.calcMetaphone2()

    #oFrenchDict.createNgrams(spBuild, 3)
    if xArgs.spellvariants:
        oFrenchDict.generateSpellVariants(1, spBuild)

    ### Statistiques
    spfStats = spBuild+'/'+STATS_NAME+xArgs.verdic+'.txt'
    oStatsLex = StatsLex(oFrenchDict)
    oStatsLex.addLexFromFile('lexique/corpus_data/stats_google_ngram_1.txt', 'G', 'Google 1-grams')
    oStatsLex.addLexFromFile('lexique/corpus_data/stats_frwiki.txt', 'W', 'Wikipédia')
    oStatsLex.addLexFromFile('lexique/corpus_data/stats_frwikisource.txt', 'S', 'Wikisource')
    oStatsLex.addLexFromFile('lexique/corpus_data/stats_litterature.txt', 'L', 'Littérature')
    oStatsLex.write(spBuild+'/test_lex.txt')
    oFrenchDict.calculateStats(oStatsLex, spfStats)
    

    ### écriture des paquets
    echo("Création des paquets...")

    spLexiconDestGL = "../../../lexicons"  if xArgs.grammalecte  else ""
    spLibreOfficeExtDestGL = "../oxt/Dictionnaires/dictionaries"  if xArgs.grammalecte  else ""
    spMozillaExtDestGL = "../xpi/data/dictionaries"  if xArgs.grammalecte  else ""
    spMozillaExtDestGL = ""  if xArgs.grammalecte  else "" # no more Hunspell dictionaries in Mozilla extensions for now
    spDataDestGL = "../data"  if xArgs.grammalecte  else ""

    if not xArgs.uncompress:
        oFrenchDict.defineAbreviatedTags(xArgs.mode, spfStats)
    oFrenchDict.createFiles(spBuild, [dMODERNE, dTOUTESVAR, dCLASSIQUE, dREFORME1990], xArgs.mode, xArgs.simplify)
    oFrenchDict.createLexiconPackages(spBuild, xArgs.verdic, oStatsLex, spLexiconDestGL)
    oFrenchDict.createFileIfqForDB(spBuild)