mbcs.js

var util = require('util'),
  Match = require ('../match');

/**
 * Binary search implementation (recursive)
 */
function binarySearch(arr, searchValue) {
  function find(arr, searchValue, left, right) {
    if (right < left)
      return -1;

    /*
    int mid = mid = (left + right) / 2;
    There is a bug in the above line;
    Joshua Bloch suggests the following replacement:
    */
    var mid = Math.floor((left + right) >>> 1);
    if (searchValue > arr[mid])
      return find(arr, searchValue, mid + 1, right);

    if (searchValue < arr[mid])
      return find(arr, searchValue, left, mid - 1);

    return mid;
  };

  return find(arr, searchValue, 0, arr.length - 1);
};

// 'Character'  iterated character class.
//    Recognizers for specific mbcs encodings make their 'characters' available
//    by providing a nextChar() function that fills in an instance of iteratedChar
//    with the next char from the input.
//    The returned characters are not converted to Unicode, but remain as the raw
//    bytes (concatenated into an int) from the codepage data.
//
//  For Asian charsets, use the raw input rather than the input that has been
//   stripped of markup.  Detection only considers multi-byte chars, effectively
//   stripping markup anyway, and double byte chars do occur in markup too.
//
function IteratedChar() {

  this.charValue = 0; // 1-4 bytes from the raw input data
  this.index     = 0;
  this.nextIndex = 0;
  this.error     = false;
  this.done      = false;

  this.reset = function() {
    this.charValue = 0;
    this.index     = -1;
    this.nextIndex = 0;
    this.error     = false;
    this.done      = false;
  };

  this.nextByte = function(det) {
    if (this.nextIndex >= det.fRawLength) {
      this.done = true;
      return -1;
    }
    var byteValue = det.fRawInput[this.nextIndex++] & 0x00ff;
    return byteValue;
  };
};



/**
 * Asian double or multi-byte - charsets.
 * Match is determined mostly by the input data adhering to the
 * encoding scheme for the charset, and, optionally,
 * frequency-of-occurence of characters.
 */

function mbcs() {};

/**
 * Test the match of this charset with the input text data
 *      which is obtained via the CharsetDetector object.
 *
 * @param det  The CharsetDetector, which contains the input text
 *             to be checked for being in this charset.
 * @return     Two values packed into one int  (Damn java, anyhow)
 *             bits 0-7:  the match confidence, ranging from 0-100
 *             bits 8-15: The match reason, an enum-like value.
 */
mbcs.prototype.match = function(det) {

  var singleByteCharCount = 0,  //TODO Do we really need this?
    doubleByteCharCount = 0,
    commonCharCount     = 0,
    badCharCount        = 0,
    totalCharCount      = 0,
    confidence          = 0;

  var iter = new IteratedChar();

  detectBlock: {
    for (iter.reset(); this.nextChar(iter, det);) {
      totalCharCount++;
      if (iter.error) {
        badCharCount++;
      } else {
        var cv = iter.charValue & 0xFFFFFFFF;

        if (cv <= 0xff) {
          singleByteCharCount++;
        } else {
          doubleByteCharCount++;
          if (this.commonChars != null) {
            // NOTE: This assumes that there are no 4-byte common chars.
            if (binarySearch(this.commonChars, cv) >= 0) {
              commonCharCount++;
            }
          }
        }
      }
      if (badCharCount >= 2 && badCharCount * 5 >= doubleByteCharCount) {
        // console.log('its here!')
        // Bail out early if the byte data is not matching the encoding scheme.
        break detectBlock;
      }
    }

    if (doubleByteCharCount <= 10 && badCharCount== 0) {
      // Not many multi-byte chars.
      if (doubleByteCharCount == 0 && totalCharCount < 10) {
        // There weren't any multibyte sequences, and there was a low density of non-ASCII single bytes.
        // We don't have enough data to have any confidence.
        // Statistical analysis of single byte non-ASCII charcters would probably help here.
        confidence = 0;
      }
      else {
        //   ASCII or ISO file?  It's probably not our encoding,
        //   but is not incompatible with our encoding, so don't give it a zero.
        confidence = 10;
      }
      break detectBlock;
    }

    //
    //  No match if there are too many characters that don't fit the encoding scheme.
    //    (should we have zero tolerance for these?)
    //
    if (doubleByteCharCount < 20 * badCharCount) {
      confidence = 0;
      break detectBlock;
    }

    if (this.commonChars == null) {
      // We have no statistics on frequently occuring characters.
      //  Assess confidence purely on having a reasonable number of
      //  multi-byte characters (the more the better
      confidence = 30 + doubleByteCharCount - 20 * badCharCount;
      if (confidence > 100) {
        confidence = 100;
      }
    } else {
      //
      // Frequency of occurence statistics exist.
      //
      var maxVal = Math.log(parseFloat(doubleByteCharCount) / 4);
      var scaleFactor = 90.0 / maxVal;
      confidence = Math.floor(Math.log(commonCharCount + 1) * scaleFactor + 10);
      confidence = Math.min(confidence, 100);
    }
  }   // end of detectBlock:

  return confidence == 0 ? null : new Match(det, this, confidence);
};

/**
 * Get the next character (however many bytes it is) from the input data
 *    Subclasses for specific charset encodings must implement this function
 *    to get characters according to the rules of their encoding scheme.
 *
 *  This function is not a method of class iteratedChar only because
 *   that would require a lot of extra derived classes, which is awkward.
 * @param it  The iteratedChar 'struct' into which the returned char is placed.
 * @param det The charset detector, which is needed to get at the input byte data
 *            being iterated over.
 * @return    True if a character was returned, false at end of input.
 */

mbcs.prototype.nextChar = function(iter, det) {};



/**
 * Shift-JIS charset recognizer.
 */
module.exports.sjis = function() {
  this.name = function() {
    return 'Shift-JIS';
  };
  this.language = function() {
    return 'ja';
  };

  // TODO:  This set of data comes from the character frequency-
  //        of-occurence analysis tool.  The data needs to be moved
  //        into a resource and loaded from there.
  this.commonChars = [
    0x8140, 0x8141, 0x8142, 0x8145, 0x815b, 0x8169, 0x816a, 0x8175, 0x8176, 0x82a0,
    0x82a2, 0x82a4, 0x82a9, 0x82aa, 0x82ab, 0x82ad, 0x82af, 0x82b1, 0x82b3, 0x82b5,
    0x82b7, 0x82bd, 0x82be, 0x82c1, 0x82c4, 0x82c5, 0x82c6, 0x82c8, 0x82c9, 0x82cc,
    0x82cd, 0x82dc, 0x82e0, 0x82e7, 0x82e8, 0x82e9, 0x82ea, 0x82f0, 0x82f1, 0x8341,
    0x8343, 0x834e, 0x834f, 0x8358, 0x835e, 0x8362, 0x8367, 0x8375, 0x8376, 0x8389,
    0x838a, 0x838b, 0x838d, 0x8393, 0x8e96, 0x93fa, 0x95aa
  ];

  this.nextChar = function(iter, det) {
    iter.index = iter.nextIndex;
    iter.error = false;

    var firstByte;
    firstByte = iter.charValue = iter.nextByte(det);
    if (firstByte < 0)
      return false;

    if (firstByte <= 0x7f || (firstByte > 0xa0 && firstByte <= 0xdf))
      return true;

    var secondByte = iter.nextByte(det);
    if (secondByte < 0)
      return false;

    iter.charValue = (firstByte << 8) | secondByte;
    if (! ((secondByte >= 0x40 && secondByte <= 0x7f) || (secondByte >= 0x80 && secondByte <= 0xff))) {
      // Illegal second byte value.
      iter.error = true;
    }
    return true;
  };
};
util.inherits(module.exports.sjis, mbcs);



/**
 *   Big5 charset recognizer.
 */
module.exports.big5 = function() {
  this.name = function() {
    return 'Big5';
  };
  this.language = function() {
    return 'zh';
  };
  // TODO:  This set of data comes from the character frequency-
  //        of-occurence analysis tool.  The data needs to be moved
  //        into a resource and loaded from there.
  this.commonChars = [
    0xa140, 0xa141, 0xa142, 0xa143, 0xa147, 0xa149, 0xa175, 0xa176, 0xa440, 0xa446,
    0xa447, 0xa448, 0xa451, 0xa454, 0xa457, 0xa464, 0xa46a, 0xa46c, 0xa477, 0xa4a3,
    0xa4a4, 0xa4a7, 0xa4c1, 0xa4ce, 0xa4d1, 0xa4df, 0xa4e8, 0xa4fd, 0xa540, 0xa548,
    0xa558, 0xa569, 0xa5cd, 0xa5e7, 0xa657, 0xa661, 0xa662, 0xa668, 0xa670, 0xa6a8,
    0xa6b3, 0xa6b9, 0xa6d3, 0xa6db, 0xa6e6, 0xa6f2, 0xa740, 0xa751, 0xa759, 0xa7da,
    0xa8a3, 0xa8a5, 0xa8ad, 0xa8d1, 0xa8d3, 0xa8e4, 0xa8fc, 0xa9c0, 0xa9d2, 0xa9f3,
    0xaa6b, 0xaaba, 0xaabe, 0xaacc, 0xaafc, 0xac47, 0xac4f, 0xacb0, 0xacd2, 0xad59,
    0xaec9, 0xafe0, 0xb0ea, 0xb16f, 0xb2b3, 0xb2c4, 0xb36f, 0xb44c, 0xb44e, 0xb54c,
    0xb5a5, 0xb5bd, 0xb5d0, 0xb5d8, 0xb671, 0xb7ed, 0xb867, 0xb944, 0xbad8, 0xbb44,
    0xbba1, 0xbdd1, 0xc2c4, 0xc3b9, 0xc440, 0xc45f
  ];
  this.nextChar = function(iter, det) {
    iter.index = iter.nextIndex;
    iter.error = false;

    var firstByte = iter.charValue = iter.nextByte(det);

    if (firstByte < 0)
      return false;

    // single byte character.
    if (firstByte <= 0x7f || firstByte == 0xff)
      return true;

    var secondByte = iter.nextByte(det);

    if (secondByte < 0)
      return false;

    iter.charValue = (iter.charValue << 8) | secondByte;

    if (secondByte < 0x40 || secondByte == 0x7f || secondByte == 0xff)
      iter.error = true;

    return true;
  };
};
util.inherits(module.exports.big5, mbcs);



/**
 *  EUC charset recognizers.  One abstract class that provides the common function
 *  for getting the next character according to the EUC encoding scheme,
 *  and nested derived classes for EUC_KR, EUC_JP, EUC_CN.
 *
 *  Get the next character value for EUC based encodings.
 *  Character 'value' is simply the raw bytes that make up the character
 *     packed into an int.
 */
function eucNextChar(iter, det) {
  iter.index = iter.nextIndex;
  iter.error = false;
  var firstByte  = 0;
  var secondByte = 0;
  var thirdByte  = 0;
  //int fourthByte = 0;
  buildChar: {
    firstByte = iter.charValue = iter.nextByte(det);
    if (firstByte < 0) {
      // Ran off the end of the input data
      iter.done = true;
      break buildChar;
    }
    if (firstByte <= 0x8d) {
      // single byte char
      break buildChar;
    }
    secondByte = iter.nextByte(det);
    iter.charValue = (iter.charValue << 8) | secondByte;
    if (firstByte >= 0xA1 && firstByte <= 0xfe) {
      // Two byte Char
      if (secondByte < 0xa1) {
        iter.error = true;
      }
      break buildChar;
    }
    if (firstByte == 0x8e) {
      // Code Set 2.
      //   In EUC-JP, total char size is 2 bytes, only one byte of actual char value.
      //   In EUC-TW, total char size is 4 bytes, three bytes contribute to char value.
      // We don't know which we've got.
      // Treat it like EUC-JP.  If the data really was EUC-TW, the following two
      //   bytes will look like a well formed 2 byte char.
      if (secondByte < 0xa1) {
        iter.error = true;
      }
      break buildChar;
    }
    if (firstByte == 0x8f) {
      // Code set 3.
      // Three byte total char size, two bytes of actual char value.
      thirdByte = iter.nextByte(det);
      iter.charValue = (iter.charValue << 8) | thirdByte;
      if (thirdByte < 0xa1) {
        iter.error = true;
      }
    }
  }
  return iter.done == false;
};



/**
 * The charset recognize for EUC-JP.  A singleton instance of this class
 *    is created and kept by the public CharsetDetector class
 */
module.exports.euc_jp = function() {
  this.name = function() {
    return 'EUC-JP';
  };
  this.language = function() {
    return 'ja';
  };

  // TODO:  This set of data comes from the character frequency-
  //        of-occurence analysis tool.  The data needs to be moved
  //        into a resource and loaded from there.
  this.commonChars = [
    0xa1a1, 0xa1a2, 0xa1a3, 0xa1a6, 0xa1bc, 0xa1ca, 0xa1cb, 0xa1d6, 0xa1d7, 0xa4a2,
    0xa4a4, 0xa4a6, 0xa4a8, 0xa4aa, 0xa4ab, 0xa4ac, 0xa4ad, 0xa4af, 0xa4b1, 0xa4b3,
    0xa4b5, 0xa4b7, 0xa4b9, 0xa4bb, 0xa4bd, 0xa4bf, 0xa4c0, 0xa4c1, 0xa4c3, 0xa4c4,
    0xa4c6, 0xa4c7, 0xa4c8, 0xa4c9, 0xa4ca, 0xa4cb, 0xa4ce, 0xa4cf, 0xa4d0, 0xa4de,
    0xa4df, 0xa4e1, 0xa4e2, 0xa4e4, 0xa4e8, 0xa4e9, 0xa4ea, 0xa4eb, 0xa4ec, 0xa4ef,
    0xa4f2, 0xa4f3, 0xa5a2, 0xa5a3, 0xa5a4, 0xa5a6, 0xa5a7, 0xa5aa, 0xa5ad, 0xa5af,
    0xa5b0, 0xa5b3, 0xa5b5, 0xa5b7, 0xa5b8, 0xa5b9, 0xa5bf, 0xa5c3, 0xa5c6, 0xa5c7,
    0xa5c8, 0xa5c9, 0xa5cb, 0xa5d0, 0xa5d5, 0xa5d6, 0xa5d7, 0xa5de, 0xa5e0, 0xa5e1,
    0xa5e5, 0xa5e9, 0xa5ea, 0xa5eb, 0xa5ec, 0xa5ed, 0xa5f3, 0xb8a9, 0xb9d4, 0xbaee,
    0xbbc8, 0xbef0, 0xbfb7, 0xc4ea, 0xc6fc, 0xc7bd, 0xcab8, 0xcaf3, 0xcbdc, 0xcdd1
  ];

  this.nextChar = eucNextChar;
};
util.inherits(module.exports.euc_jp, mbcs);



/**
 * The charset recognize for EUC-KR.  A singleton instance of this class
 *    is created and kept by the public CharsetDetector class
 */
module.exports.euc_kr = function() {
  this.name = function() {
    return 'EUC-KR';
  };
  this.language = function() {
    return 'ko';
  };

  // TODO:  This set of data comes from the character frequency-
  //        of-occurence analysis tool.  The data needs to be moved
  //        into a resource and loaded from there.
  this.commonChars = [
    0xb0a1, 0xb0b3, 0xb0c5, 0xb0cd, 0xb0d4, 0xb0e6, 0xb0ed, 0xb0f8, 0xb0fa, 0xb0fc,
    0xb1b8, 0xb1b9, 0xb1c7, 0xb1d7, 0xb1e2, 0xb3aa, 0xb3bb, 0xb4c2, 0xb4cf, 0xb4d9,
    0xb4eb, 0xb5a5, 0xb5b5, 0xb5bf, 0xb5c7, 0xb5e9, 0xb6f3, 0xb7af, 0xb7c2, 0xb7ce,
    0xb8a6, 0xb8ae, 0xb8b6, 0xb8b8, 0xb8bb, 0xb8e9, 0xb9ab, 0xb9ae, 0xb9cc, 0xb9ce,
    0xb9fd, 0xbab8, 0xbace, 0xbad0, 0xbaf1, 0xbbe7, 0xbbf3, 0xbbfd, 0xbcad, 0xbcba,
    0xbcd2, 0xbcf6, 0xbdba, 0xbdc0, 0xbdc3, 0xbdc5, 0xbec6, 0xbec8, 0xbedf, 0xbeee,
    0xbef8, 0xbefa, 0xbfa1, 0xbfa9, 0xbfc0, 0xbfe4, 0xbfeb, 0xbfec, 0xbff8, 0xc0a7,
    0xc0af, 0xc0b8, 0xc0ba, 0xc0bb, 0xc0bd, 0xc0c7, 0xc0cc, 0xc0ce, 0xc0cf, 0xc0d6,
    0xc0da, 0xc0e5, 0xc0fb, 0xc0fc, 0xc1a4, 0xc1a6, 0xc1b6, 0xc1d6, 0xc1df, 0xc1f6,
    0xc1f8, 0xc4a1, 0xc5cd, 0xc6ae, 0xc7cf, 0xc7d1, 0xc7d2, 0xc7d8, 0xc7e5, 0xc8ad
  ];

  this.nextChar = eucNextChar;
};
util.inherits(module.exports.euc_kr, mbcs);



/**
 *   GB-18030 recognizer. Uses simplified Chinese statistics.
 */
module.exports.gb_18030 = function() {
  this.name = function() {
    return 'GB18030';
  };
  this.language = function() {
    return 'zh';
  };

  /*
   *  Get the next character value for EUC based encodings.
   *  Character 'value' is simply the raw bytes that make up the character
   *     packed into an int.
   */
  this.nextChar = function(iter, det) {
    iter.index = iter.nextIndex;
    iter.error = false;
    var firstByte  = 0;
    var secondByte = 0;
    var thirdByte  = 0;
    var fourthByte = 0;
    buildChar: {
      firstByte = iter.charValue = iter.nextByte(det);
      if (firstByte < 0) {
        // Ran off the end of the input data
        iter.done = true;
        break buildChar;
      }
      if (firstByte <= 0x80) {
        // single byte char
        break buildChar;
      }
      secondByte = iter.nextByte(det);
      iter.charValue = (iter.charValue << 8) | secondByte;
      if (firstByte >= 0x81 && firstByte <= 0xFE) {
        // Two byte Char
        if ((secondByte >= 0x40 && secondByte <= 0x7E) || (secondByte >=80 && secondByte <= 0xFE)) {
          break buildChar;
        }
        // Four byte char
        if (secondByte >= 0x30 && secondByte <= 0x39) {
          thirdByte = iter.nextByte(det);
          if (thirdByte >= 0x81 && thirdByte <= 0xFE) {
            fourthByte = iter.nextByte(det);
            if (fourthByte >= 0x30 && fourthByte <= 0x39) {
              iter.charValue = (iter.charValue << 16) | (thirdByte << 8) | fourthByte;
              break buildChar;
            }
          }
        }
        iter.error = true;
        break buildChar;
      }
    }
    return iter.done == false;
  };

  // TODO:  This set of data comes from the character frequency-
  //        of-occurence analysis tool.  The data needs to be moved
  //        into a resource and loaded from there.
  this.commonChars = [
    0xa1a1, 0xa1a2, 0xa1a3, 0xa1a4, 0xa1b0, 0xa1b1, 0xa1f1, 0xa1f3, 0xa3a1, 0xa3ac,
    0xa3ba, 0xb1a8, 0xb1b8, 0xb1be, 0xb2bb, 0xb3c9, 0xb3f6, 0xb4f3, 0xb5bd, 0xb5c4,
    0xb5e3, 0xb6af, 0xb6d4, 0xb6e0, 0xb7a2, 0xb7a8, 0xb7bd, 0xb7d6, 0xb7dd, 0xb8b4,
    0xb8df, 0xb8f6, 0xb9ab, 0xb9c9, 0xb9d8, 0xb9fa, 0xb9fd, 0xbacd, 0xbba7, 0xbbd6,
    0xbbe1, 0xbbfa, 0xbcbc, 0xbcdb, 0xbcfe, 0xbdcc, 0xbecd, 0xbedd, 0xbfb4, 0xbfc6,
    0xbfc9, 0xc0b4, 0xc0ed, 0xc1cb, 0xc2db, 0xc3c7, 0xc4dc, 0xc4ea, 0xc5cc, 0xc6f7,
    0xc7f8, 0xc8ab, 0xc8cb, 0xc8d5, 0xc8e7, 0xc9cf, 0xc9fa, 0xcab1, 0xcab5, 0xcac7,
    0xcad0, 0xcad6, 0xcaf5, 0xcafd, 0xccec, 0xcdf8, 0xceaa, 0xcec4, 0xced2, 0xcee5,
    0xcfb5, 0xcfc2, 0xcfd6, 0xd0c2, 0xd0c5, 0xd0d0, 0xd0d4, 0xd1a7, 0xd2aa, 0xd2b2,
    0xd2b5, 0xd2bb, 0xd2d4, 0xd3c3, 0xd3d0, 0xd3fd, 0xd4c2, 0xd4da, 0xd5e2, 0xd6d0
  ];
};
util.inherits(module.exports.gb_18030, mbcs);