Chinese-Cedict/Programmes/cedict2csv.js

var fs = require("fs");
var readline = require("readline");
var Stream = require("stream");

// librairie de traitement du pinyin
const pinyinizer = require("pinyinizer");

// utilitaires
var tab = "\t";
var endLine = "\n";

var ligne = {
  hanzi: "",
  traditional: "",
  pinyin: "",
  translations: []
};

function resetLigne() {
  ligne.hanzi = "";
  ligne.traditional = "";
  ligne.pinyin = "";
  ligne.translations = [];
}

// la ligne CSV qu'on va écrire dans le fichier CSV
var ligneCSV = "";

function readFileLineByLine(inputFile, outputFile) {
  var instream = fs.createReadStream(inputFile);
  var outstream = new Stream();
  outstream.readable = true;
  outstream.writable = true;

  var rl = readline.createInterface({
    input: instream,
    output: outstream,
    terminal: false
  });

  rl.on("line", function(line) {
    if (line[0] != "#") {
      chinois = line.split(" ", 2);
      // console.log(chinois[0], " ## ", chinois[1]);
      ligne.traditional = chinois[0];
      ligne.hanzi = chinois[1];
      i = line.search(/\[/);
      j = line.search(/\]/);
      phonetic = line.substr(i + 1, j - i - 1);
      // traitement des champs pinyin avec des chiffres
      phonetic = phonetic.toLowerCase();
      // on va tester ça ...
      try {
        var sauve = phonetic;
        phonetic = pinyinizer.pinyinize(phonetic);
        phonetic = phonetic.replace(/5/g, "");
      } catch (err) {
        console.log("Erreur: " + pinyin);
        phonetic = sauve;
      }
      ligne.pinyin = phonetic;
      // console.log(phonetic);

      // mettre en pinyin accentué les parties entre crochets

      ligne.translations = line.split(/\//g);
      var l = ligne.translations.length;
      ligne.translations.pop();
      ligne.translations.shift();
      var temp = ligne.translations.join(" / ");

      var tableau = temp.match(/\[.*?\]/g);
      if (tableau != null) {
        tableau.forEach(function(t) {
          t = t.replace("u:","v");
          var t1 = pinyinizer.pinyinize(t);
          //console.log(t, "-", t1);
          temp = temp.replace(t, t1);
          temp = temp.replace("5","");
          //console.log(temp);
        });
      }

      // console.log(l);
      // console.log(ligne.translations);
      // console.log(temp);
      ligneCSV =
        ligne.hanzi +
        tab +
        ligne.traditional +
        tab +
        ligne.pinyin +
        tab +
        temp +
        endLine;

      console.log(ligneCSV);
      fs.appendFileSync(outputFile, ligneCSV);
      resetLigne();
    }
  });
}

readFileLineByLine("cedict.txt", "cedict.csv");