Character set detection is the process of determining the character set, or encoding, of character data in an unknown format.

A simple binding of ICU character set detection (http://userguide.icu-project.org/conversion/detection) for Node.js.

At first, install libicu into your system (See this instruction for details).

After that, install node-icu-charset-detector from npm.

npm install node-icu-charset-detector

• Debian (Ubuntu)

  apt-get install libicu-dev

• Gentoo

  emerge icu

• Fedora/CentOS

  yum install libicu-devel

• MacPorts

  port install icu +devel

• Homebrew

brew install icu4c
brew link icu4c --force

If experiencing issues with 'homebrew' installing version 50.1 of icu4c, try the following:

brew search icu4c
brew tap homebrew/versions
brew versions icu4c
cd $(brew --prefix) && git pull --rebase
git checkout c25fd2f $(brew --prefix)/Library/Formula/icu4c.rb
brew install icu4c

• From source

curl -O http://download.icu-project.org/files/icu4c/52.1/icu4c-52_1-src.tgz
tar xzvf icu4c-4_4_2-src.tgz
cd icu/source
chmod +x runConfigureICU configure install-sh
./runConfigureICU MacOSX
make
sudo make install
xcode-select --install

node-icu-charset-detector provides a function detectCharset(buffer), where buffer is an instance of Buffer whose charset should be detected.

var charsetDetector = require("node-icu-charset-detector");

var buffer = fs.readFileSync("/path/to/the/file");
var charset = charsetDetector.detectCharset(buffer);

console.log("charset name: " + charset.toString());
console.log("language: " + charset.language);
console.log("detection confidence: " + charset.confidence);

detectCharset(buffer) returns the detected charset name for buffer, and the returned charset name has two extra properties language and confidence:

• charset.language
  • language name for the detected character set.
• charset.confidence
  • confidence of the charset detection for charset.

Since ICU itself does not have a feature to convert character sets, you may need to use node-iconv (https://github.com/bnoordhuis/node-iconv), which has a powerful character sets converting feature.

Here is a simple example to leverage node-iconv to convert character sets not supported by Node itself.

function bufferToString(buffer) {
  var charsetDetector = require("node-icu-charset-detector");
  var charset = charsetDetector.detectCharset(buffer).toString();

  try {
    return buffer.toString(charset);
  } catch (x) {
    var Iconv = require("iconv").Iconv;
    var charsetConverter = new Iconv(charset, "utf8");
    return charsetConverter.convert(buffer).toString();
  }
}

var buffer = fs.readFileSync("/path/to/the/file");
var bufferString = bufferToString(buffer);