Convert and detect character encoding in JavaScript.

README (日本語)

• Features
  • How to Use Character Encoding in Strings?
• Installation
  • npm
    • TypeScript
  • Browser (standalone)
  • CDN
• Supported encodings
  • About UNICODE
• Example usage
• Demo
• API
  • detect : Detects character encoding
  • convert : Converts character encoding
    • Specify conversion options to the argument to as an object
    • Specify the return type by the type option
    • Specify handling for unrepresentable characters
    • Replacing characters with HTML entities when they cannot be represented
    • Ignoring characters when they cannot be represented
    • Throwing an Error when they cannot be represented
    • Specify BOM in UTF-16
  • urlEncode : Encodes to percent-encoded string
  • urlDecode : Decodes from percent-encoded string
  • base64Encode : Encodes to Base64 formatted string
  • base64Decode : Decodes from Base64 formatted string
  • codeToString : Converts character code array to string
  • stringToCode : Converts string to character code array
  • Japanese Zenkaku/Hankaku conversion
• Other examples
  • Example using the fetch API and Typed Arrays (Uint8Array)
  • Convert encoding for file using the File APIs
• Contributing
• License

encoding.js is a JavaScript library for converting and detecting character encodings, supporting both Japanese character encodings (Shift_JIS, EUC-JP, ISO-2022-JP) and Unicode formats (UTF-8, UTF-16).

Since JavaScript string values are internally encoded as UTF-16 code units (ref: ECMAScript® 2019 Language Specification - 6.1.4 The String Type), they cannot directly handle other character encodings as strings. However, encoding.js overcomes this limitation by treating these encodings as arrays instead of strings, enabling the conversion between different character sets.

Each character encoding is represented as an array of numbers corresponding to character code values, for example, [130, 160] represents "あ" in Shift_JIS.

The array of character codes used in its methods can also be utilized with TypedArray objects, such as Uint8Array, or with Buffer in Node.js.

Numeric arrays of character codes can be converted to strings using methods such as Encoding.codeToString. However, due to the JavaScript specifications mentioned above, some character encodings may not be handled properly when converted directly to strings.

If you prefer to use strings instead of numeric arrays, you can convert them to percent-encoded strings, such as '%82%A0', using Encoding.urlEncode and Encoding.urlDecode for passing to other resources. Similarly, Encoding.base64Encode and Encoding.base64Decode allow for encoding and decoding to and from base64, which can then be passed as strings.

encoding.js is published under the package name encoding-japanese on npm.

npm install --save encoding-japanese

import Encoding from 'encoding-japanese';

const Encoding = require('encoding-japanese');

TypeScript type definitions for encoding.js are available at @types/encoding-japanese (thanks to @rhysd).

npm install --save-dev @types/encoding-japanese

To use encoding.js in a browser environment, you can either install it via npm or download it directly from the release list. The package includes both encoding.js and encoding.min.js.

Note: Cloning the repository via git clone might give you access to the master (or main) branch, which could still be in a development state.

<!-- To include the full version -->
<script src="encoding.js"></script>

<!-- Or, to include the minified version for production -->
<script src="encoding.min.js"></script>

When the script is loaded, the object Encoding is defined in the global scope (i.e., window.Encoding).

You can use encoding.js (package name: encoding-japanese) directly from a CDN via a script tag:

<script src="https://unpkg.com/encoding-japanese@2.2.0/encoding.min.js"></script>

In this example we use unpkg, but you can use any CDN that provides npm packages, for example cdnjs or jsDelivr.

In encoding.js, UNICODE is defined as the internal character encoding that JavaScript strings (JavaScript string objects) can handle directly.

As mentioned in the Features section, JavaScript strings are internally encoded using UTF-16 code units. This means that other character encodings cannot be directly handled without conversion. Therefore, when converting to a character encoding that is properly representable in JavaScript, you should specify UNICODE.

(Note: Even if the HTML file's encoding is UTF-8, you should specify UNICODE instead of UTF8 when processing the encoding in JavaScript.)

When using Encoding.convert, if you specify a character encoding other than UNICODE (such as UTF8 or SJIS), the values in the returned character code array will range from 0-255. However, if you specify UNICODE, the values will range from 0-65535, which corresponds to the range of values returned by String.prototype.charCodeAt() (Code Units).

Convert character encoding from JavaScript string (UNICODE) to SJIS.

const unicodeArray = Encoding.stringToCode('こんにちは'); // Convert string to code array
const sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE'
});
console.log(sjisArray);
// [130, 177, 130, 241, 130, 201, 130, 191, 130, 205] ('こんにちは' array in SJIS)

Convert character encoding from SJIS to UNICODE.

const sjisArray = [
  130, 177, 130, 241, 130, 201, 130, 191, 130, 205
]; // 'こんにちは' array in SJIS

const unicodeArray = Encoding.convert(sjisArray, {
  to: 'UNICODE',
  from: 'SJIS'
});
const str = Encoding.codeToString(unicodeArray); // Convert code array to string
console.log(str); // 'こんにちは'

Detect character encoding.

const data = [
  227, 129, 147, 227, 130, 147, 227, 129, 171, 227, 129, 161, 227, 129, 175
]; // 'こんにちは' array in UTF-8

const detectedEncoding = Encoding.detect(data);
console.log(`Character encoding is ${detectedEncoding}`); // 'Character encoding is UTF8'

(Node.js) Example of reading a text file written in SJIS.

const fs = require('fs');
const Encoding = require('encoding-japanese');

const sjisBuffer = fs.readFileSync('./sjis.txt');
const unicodeArray = Encoding.convert(sjisBuffer, {
  to: 'UNICODE',
  from: 'SJIS'
});
console.log(Encoding.codeToString(unicodeArray));

• Playground for testing character encoding conversion and detection
• Test run for reading sample files and converting character encodings
• Demo for converting and detecting character encoding by specifying a file

• detect
• convert
• urlEncode
• urlDecode
• base64Encode
• base64Decode
• codeToString
• stringToCode
• Japanese Zenkaku/Hankaku conversion

Detects the character encoding of the given data.

• data (Array<number>|TypedArray|Buffer|string) : The code array or string to detect character encoding.
• [encodings] (string|Array<string>|Object) : (Optional) Specifies a specific character encoding, or an array of encodings to limit the detection. Detects automatically if this argument is omitted or AUTO is specified. Supported encoding values can be found in the "Supported encodings" section.

(string|boolean): Returns a string representing the detected encoding (e.g., SJIS, UTF8) listed in the "Supported encodings" section, or false if the encoding cannot be detected. If the encodings argument is provided, it returns the name of the detected encoding if the data matches any of the specified encodings, or false otherwise.

Example of detecting character encoding.

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const detectedEncoding = Encoding.detect(sjisArray);
console.log(`Encoding is ${detectedEncoding}`); // 'Encoding is SJIS'

Example of using the encodings argument to specify the character encoding to be detected. This returns a string detected encoding if the specified encoding matches, or false otherwise:

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const detectedEncoding = Encoding.detect(sjisArray, 'SJIS');
if (detectedEncoding) {
  console.log('Encoding is SJIS');
} else {
  console.log('Encoding does not match SJIS');
}

Example of specifying multiple encodings:

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const detectedEncoding = Encoding.detect(sjisArray, ['UTF8', 'SJIS']);
if (detectedEncoding) {
  console.log(`Encoding is ${detectedEncoding}`); // 'Encoding is SJIS'
} else {
  console.log('Encoding does not match UTF8 and SJIS');
}

Converts the character encoding of the given data.

• data (Array<number>|TypedArray|Buffer|string) : The code array or string to convert character encoding.
• to (string|Object) : The character encoding name of the conversion destination as a string, or conversion options as an object.
• [from] (string|Array<string>) : (Optional) The character encoding name of the conversion source as a string, or an array of encoding names. Detects automatically if this argument is omitted or AUTO is specified. Supported encoding values can be found in the "Supported encodings" section.

(Array<number>|TypedArray|string) : Returns a numeric character code array of the converted character encoding if data is an array or a buffer, or returns the converted string if data is a string.

Example of converting a character code array to Shift_JIS from UTF-8:

const utf8Array = [227, 129, 130]; // 'あ' in UTF-8
const sjisArray = Encoding.convert(utf8Array, 'SJIS', 'UTF8');
console.log(sjisArray); // [130, 160] ('あ' in SJIS)

TypedArray such as Uint8Array, and Buffer of Node.js can be converted in the same usage:

const utf8Array = new Uint8Array([227, 129, 130]);
const sjisArray = Encoding.convert(utf8Array, 'SJIS', 'UTF8');

Converts character encoding by auto-detecting the encoding name of the source:

// The character encoding is automatically detected when the argument `from` is omitted
const utf8Array = [227, 129, 130];
let sjisArray = Encoding.convert(utf8Array, 'SJIS');
// Or explicitly specify 'AUTO' to auto-detecting
sjisArray = Encoding.convert(utf8Array, 'SJIS', 'AUTO');

You can pass the second argument to as an object for improving readability. Also, the following options such as type, fallback, and bom must be specified with an object.

const utf8Array = [227, 129, 130];
const sjisArray = Encoding.convert(utf8Array, {
  to: 'SJIS',
  from: 'UTF8'
});

convert returns an array by default, but you can change the return type by specifying the type option. Also, if the argument data is passed as a string and the type option is not specified, then type ='string' is assumed (returns as a string).

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const unicodeString = Encoding.convert(sjisArray, {
  to: 'UNICODE',
  from: 'SJIS',
  type: 'string' // Specify 'string' to return as string
});
console.log(unicodeString); // 'おはよ'

The following type options are supported.

• string : Return as a string.
• arraybuffer : Return as an ArrayBuffer (Actually returns a Uint16Array due to historical reasons).
• array : Return as an Array. (default)

type: 'string' can be used as a shorthand for converting a code array to a string, as performed by Encoding.codeToString.
Note: Specifying type: 'string' may not handle conversions properly, except when converting to UNICODE.

With the fallback option, you can specify how to handle characters that cannot be represented in the target encoding. The fallback option supports the following values:

• html-entity: Replace characters with HTML entities (decimal HTML numeric character references).
• html-entity-hex: Replace characters with HTML entities (hexadecimal HTML numeric character references).
• ignore: Ignore characters that cannot be represented.
• error: Throw an error if any character cannot be represented.

Characters that cannot be represented in the target character set are replaced with '?' (U+003F) by default, but by specifying html-entity as the fallback option, you can replace them with HTML entities (Numeric character references), such as 🍣.

Example of specifying { fallback: 'html-entity' } option:

const unicodeArray = Encoding.stringToCode('寿司🍣ビール🍺');
// No fallback specified
let sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE'
});
console.log(sjisArray); // Converted to a code array of '寿司?ビール?'

// Specify `fallback: html-entity`
sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE',
  fallback: 'html-entity'
});
console.log(sjisArray); // Converted to a code array of '寿司🍣ビール🍺'

Example of specifying { fallback: 'html-entity-hex' } option:

const unicodeArray = Encoding.stringToCode('ホッケの漢字は𩸽');
const sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE',
  fallback: 'html-entity-hex'
});
console.log(sjisArray); // Converted to a code array of 'ホッケの漢字は𩸽'

By specifying ignore as a fallback option, characters that cannot be represented in the target encoding format can be ignored.

Example of specifying { fallback: 'ignore' } option:

const unicodeArray = Encoding.stringToCode('寿司🍣ビール🍺');
// No fallback specified
let sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE'
});
console.log(sjisArray); // Converted to a code array of '寿司?ビール?'

// Specify `fallback: ignore`
sjisArray = Encoding.convert(unicodeArray, {
  to: 'SJIS',
  from: 'UNICODE',
  fallback: 'ignore'
});
console.log(sjisArray); // Converted to a code array of '寿司ビール'

If you need to throw an error when a character cannot be represented in the target character encoding, specify error as a fallback option. This will cause an exception to be thrown.

Example of specifying { fallback: 'error' } option:

const unicodeArray = Encoding.stringToCode('おにぎり🍙ラーメン🍜');
try {
  const sjisArray = Encoding.convert(unicodeArray, {
    to: 'SJIS',
    from: 'UNICODE',
    fallback: 'error' // Specify 'error' to throw an exception
  });
} catch (e) {
  console.error(e); // Error: Character cannot be represented: [240, 159, 141, 153]
}

You can add a BOM (byte order mark) by specifying the bom option when converting to UTF16. The default is no BOM.

const utf16Array = Encoding.convert(utf8Array, {
  to: 'UTF16',
  from: 'UTF8',
  bom: true // Specify to add the BOM
});

UTF16 byte order is big-endian by default. If you want to convert as little-endian, specify the { bom: 'LE' } option.

const utf16leArray = Encoding.convert(utf8Array, {
  to: 'UTF16',
  from: 'UTF8',
  bom: 'LE' // Specify to add the BOM as little-endian
});

If you do not need BOM, use UTF16BE or UTF16LE. UTF16BE is big-endian, and UTF16LE is little-endian, and both have no BOM.

const utf16beArray = Encoding.convert(utf8Array, {
  to: 'UTF16BE',
  from: 'UTF8'
});

Encodes a numeric character code array into a percent-encoded string formatted as a URI component in %xx format.

urlEncode escapes all characters except the following, just like encodeURIComponent().

A-Z a-z 0-9 - _ . ! ~ * ' ( )

• data (Array<number>|TypedArray|Buffer|string) : The numeric character code array or string that will be encoded into a percent-encoded URI component.

(string) : Returns a percent-encoded string formatted as a URI component in %xx format.

Example of URL encoding a Shift_JIS array:

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const encoded = Encoding.urlEncode(sjisArray);
console.log(encoded); // '%82%A8%82%CD%82%E6'

Decodes a percent-encoded string formatted as a URI component in %xx format to a numeric character code array.

• string (string) : The string to decode.

(Array<number>) : Returns a numeric character code array.

Example of decoding a percent-encoded Shift_JIS string:

const encoded = '%82%A8%82%CD%82%E6'; // 'おはよ' encoded as percent-encoded SJIS string
const sjisArray = Encoding.urlDecode(encoded);
console.log(sjisArray); // [130, 168, 130, 205, 130, 230]

Encodes a numeric character code array into a Base64 encoded string.

• data (Array<number>|TypedArray|Buffer|string) : The numeric character code array or string to encode.

(string) : Returns a Base64 encoded string.

Example of Base64 encoding a Shift_JIS array:

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const encodedStr = Encoding.base64Encode(sjisArray);
console.log(encodedStr); // 'gqiCzYLm'

Decodes a Base64 encoded string to a numeric character code array.

• string (string) : The Base64 encoded string to decode.

(Array<number>) : Returns a Base64 decoded numeric character code array.

Example of base64Encode and base64Decode:

const sjisArray = [130, 177, 130, 241, 130, 201, 130, 191, 130, 205]; // 'こんにちは' array in SJIS
const encodedStr = Encoding.base64Encode(sjisArray);
console.log(encodedStr); // 'grGC8YLJgr+CzQ=='

const decodedArray = Encoding.base64Decode(encodedStr);
console.log(decodedArray); // [130, 177, 130, 241, 130, 201, 130, 191, 130, 205]

Converts a numeric character code array to string.

• code (Array<number>|TypedArray|Buffer) : The numeric character code array to convert.

(string) : Returns a converted string.

Example of converting a character code array to a string:

const sjisArray = [130, 168, 130, 205, 130, 230]; // 'おはよ' array in SJIS
const unicodeArray = Encoding.convert(sjisArray, {
  to: 'UNICODE',
  from: 'SJIS'
});
const unicodeStr = Encoding.codeToString(unicodeArray);
console.log(unicodeStr); // 'おはよ'

Converts a string to a numeric character code array.

• string (string) : The string to convert.

(Array<number>) : Returns a numeric character code array converted from the string.

Example of converting a string to a character code array:

const unicodeArray = Encoding.stringToCode('おはよ');
console.log(unicodeArray); // [12362, 12399, 12424]

The following methods convert Japanese full-width (zenkaku) and half-width (hankaku) characters, suitable for use with UNICODE strings or numeric character code arrays of UNICODE.

Returns a converted string if the argument data is a string. Returns a numeric character code array if the argument data is a code array.

• Encoding.toHankakuCase (data) : Converts full-width (zenkaku) symbols and alphanumeric characters to their half-width (hankaku) equivalents.
• Encoding.toZenkakuCase (data) : Converts half-width (hankaku) symbols and alphanumeric characters to their full-width (zenkaku) equivalents.
• Encoding.toHiraganaCase (data) : Converts full-width katakana to full-width hiragana.
• Encoding.toKatakanaCase (data) : Converts full-width hiragana to full-width katakana.
• Encoding.toHankanaCase (data) : Converts full-width katakana to half-width katakana.
• Encoding.toZenkanaCase (data) : Converts half-width katakana to full-width katakana.
• Encoding.toHankakuSpace (data) : Converts the em space (U+3000) to the single space (U+0020).
• Encoding.toZenkakuSpace (data) : Converts the single space (U+0020) to the em space (U+3000).

• data (Array<number>|TypedArray|Buffer|string) : The string or numeric character code array to convert.

(Array<number>|string) : Returns a converted string or numeric character code array.

Example of converting zenkaku and hankaku strings:

console.log(Encoding.toHankakuCase('ａｂｃＤＥＦ１２３＠！＃＊＝')); // 'abcDEF123@!#*='
console.log(Encoding.toZenkakuCase('abcDEF123@!#*=')); // 'ａｂｃＤＥＦ１２３＠！＃＊＝'
console.log(Encoding.toHiraganaCase('アイウエオァィゥェォヴボポ')); // 'あいうえおぁぃぅぇぉゔぼぽ'
console.log(Encoding.toKatakanaCase('あいうえおぁぃぅぇぉゔぼぽ')); // 'アイウエオァィゥェォヴボポ'
console.log(Encoding.toHankanaCase('アイウエオァィゥェォヴボポ')); // 'ｱｲｳｴｵｧｨｩｪｫｳﾞﾎﾞﾎﾟ'
console.log(Encoding.toZenkanaCase('ｱｲｳｴｵｧｨｩｪｫｳﾞﾎﾞﾎﾟ')); // 'アイウエオァィゥェォヴボポ'
console.log(Encoding.toHankakuSpace('あいうえお　abc　123')); // 'あいうえお abc 123'
console.log(Encoding.toZenkakuSpace('あいうえお abc 123')); // 'あいうえお　abc　123'

Example of converting zenkaku and hankaku code arrays:

const unicodeArray = Encoding.stringToCode('ａｂｃ１２３！＃　あいうアイウ ABCｱｲｳ');
console.log(Encoding.codeToString(Encoding.toHankakuCase(unicodeArray)));
// 'abc123!#　あいうアイウ ABCｱｲｳ'
console.log(Encoding.codeToString(Encoding.toZenkakuCase(unicodeArray)));
// 'ａｂｃ１２３！＃　あいうアイウ ＡＢＣｱｲｳ'
console.log(Encoding.codeToString(Encoding.toHiraganaCase(unicodeArray)));
// 'ａｂｃ１２３！＃　あいうあいう ABCｱｲｳ'
console.log(Encoding.codeToString(Encoding.toKatakanaCase(unicodeArray)));
// 'ａｂｃ１２３！＃　アイウアイウ ABCｱｲｳ'
console.log(Encoding.codeToString(Encoding.toHankanaCase(unicodeArray)));
// 'ａｂｃ１２３！＃　あいうｱｲｳ ABCｱｲｳ'
console.log(Encoding.codeToString(Encoding.toZenkanaCase(unicodeArray)));
// 'ａｂｃ１２３！＃　あいうアイウ ABCアイウ'
console.log(Encoding.codeToString(Encoding.toHankakuSpace(unicodeArray)));
// 'ａｂｃ１２３！＃ あいうアイウ ABCｱｲｳ'
console.log(Encoding.codeToString(Encoding.toZenkakuSpace(unicodeArray)));
// 'ａｂｃ１２３！＃　あいうアイウ　ABCｱｲｳ'

This example reads a text file encoded in Shift_JIS as binary data, and displays it as a string after converting it to Unicode using Encoding.convert.

(async () => {
  try {
    const response = await fetch('shift_jis.txt');
    const buffer = await response.arrayBuffer();

    // Code array with Shift_JIS file contents
    const sjisArray = new Uint8Array(buffer);

    // Convert encoding to UNICODE (JavaScript Code Units) from Shift_JIS
    const unicodeArray = Encoding.convert(sjisArray, {
      to: 'UNICODE',
      from: 'SJIS'
    });

    // Convert to string from code array for display
    const unicodeString = Encoding.codeToString(unicodeArray);
    console.log(unicodeString);
  } catch (error) {
    console.error('Error loading the file:', error);
  }
})();

const req = new XMLHttpRequest();
req.open('GET', 'shift_jis.txt', true);
req.responseType = 'arraybuffer';

req.onload = (event) => {
  const buffer = req.response;
  if (buffer) {
    // Code array with Shift_JIS file contents
    const sjisArray = new Uint8Array(buffer);

    // Convert encoding to UNICODE (JavaScript Code Units) from Shift_JIS
    const unicodeArray = Encoding.convert(sjisArray, {
      to: 'UNICODE',
      from: 'SJIS'
    });

    // Convert to string from code array for display
    const unicodeString = Encoding.codeToString(unicodeArray);
    console.log(unicodeString);
  }
};

req.send(null);

This example uses the File API to read the content of a selected file, detects its character encoding, and converts the file content to UNICODE from any character encoding such as Shift_JIS or EUC-JP. The converted content is then displayed in a textarea.

<input type="file" id="file">
<div id="encoding"></div>
<textarea id="content" rows="5" cols="80"></textarea>

<script>
function onFileSelect(event) {
  const file = event.target.files[0];

  const reader = new FileReader();
  reader.onload = function(e) {
    const codes = new Uint8Array(e.target.result);

    const detectedEncoding = Encoding.detect(codes);
    const encoding = document.getElementById('encoding');
    encoding.textContent = `Detected encoding: ${detectedEncoding}`;

    // Convert encoding to UNICODE
    const unicodeString = Encoding.convert(codes, {
      to: 'UNICODE',
      from: detectedEncoding,
      type: 'string'
    });
    document.getElementById('content').value = unicodeString;
  };

  reader.readAsArrayBuffer(file);
}

document.getElementById('file').addEventListener('change', onFileSelect);
</script>

Demo

We welcome contributions from everyone. For bug reports and feature requests, please create an issue on GitHub.

Before submitting a pull request, please run npm run test to ensure there are no errors. We only accept pull requests that pass all tests.

This project is licensed under the terms of the MIT license. See the LICENSE file for details.