From a93a9cd55791a78432c2dc81e07c884b1e8e0a0c Mon Sep 17 00:00:00 2001 From: Dmitry Shirokov Date: Mon, 4 Mar 2013 00:08:00 +0400 Subject: [PATCH] imports --- .gitignore | 2 + encoding/iso2022.js | 148 +++++++++++++ encoding/mbcs.js | 488 ++++++++++++++++++++++++++++++++++++++++++ encoding/sbcs.js | 508 ++++++++++++++++++++++++++++++++++++++++++++ encoding/unicode.js | 106 +++++++++ encoding/utf8.js | 84 ++++++++ match.js | 6 + 7 files changed, 1342 insertions(+) create mode 100644 .gitignore create mode 100644 encoding/iso2022.js create mode 100644 encoding/mbcs.js create mode 100644 encoding/sbcs.js create mode 100644 encoding/unicode.js create mode 100644 encoding/utf8.js create mode 100644 match.js diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..e1f4792 --- /dev/null +++ b/.gitignore @@ -0,0 +1,2 @@ +testing.js +java \ No newline at end of file diff --git a/encoding/iso2022.js b/encoding/iso2022.js new file mode 100644 index 0000000..610bad2 --- /dev/null +++ b/encoding/iso2022.js @@ -0,0 +1,148 @@ +var util = require('util'), + Match = require ('../match'); + + +/** + * This is a superclass for the individual detectors for + * each of the detectable members of the ISO 2022 family + * of encodings. + */ + +function ISO_2022() {}; + +ISO_2022.prototype.match = function(det) { + + /** + * Matching function shared among the 2022 detectors JP, CN and KR + * Counts up the number of legal an unrecognized escape sequences in + * the sample of text, and computes a score based on the total number & + * the proportion that fit the encoding. + * + * + * @param text the byte buffer containing text to analyse + * @param textLen the size of the text in the byte. + * @param escapeSequences the byte escape sequences to test for. + * @return match quality, in the range of 0-100. + */ + + var i, j; + var escN; + var hits = 0; + var misses = 0; + var shifts = 0; + var quality; + + // TODO: refactor me + // int match(byte [] text, int textLen, byte [][] escapeSequences) { + // det.fInputBytes, det.fInputLen + var text = det.fInputBytes; + var textLen = det.fInputLen; + + scanInput: + for (i=0; i>> 1); + if (searchValue > arr[mid]) { + return find(arr, searchValue, mid + 1, right); + } else if (searchValue < arr[mid]) { + return find(arr, searchValue, left, mid - 1); + } else { + 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(1) + // 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 = "Shift_JIS"; + this.language = "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); + // console.log('fb', firstByte) + 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 = "Big5"; + this.language = "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; + firstByte = iter.charValue = iter.nextByte(det); + if (firstByte < 0) { + return false; + } + if (firstByte <= 0x7f || firstByte == 0xff) { + // single byte character. + 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. + * + */ +module.exports.euc = function() { + + /* + * 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) { + it.index = it.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); + }; +}; +util.inherits(module.exports.euc, mbcs); + + + +/** + * 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 = "EUC-JP"; + this.language = "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 + ]; +}; +util.inherits(module.exports.euc_jp, module.exports.euc); + + + +/** + * 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 = "EUC-KR"; + this.language = "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 + ]; +}; +util.inherits(module.exports.euc_kr, module.exports.euc); + + + +/** + * GB-18030 recognizer. Uses simplified Chinese statistics. + */ +module.exports.gb_18030 = function() { + this.name = "GB18030"; + this.language = "zh"; + /* + * (non-Javadoc) + * 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); diff --git a/encoding/sbcs.js b/encoding/sbcs.js new file mode 100644 index 0000000..d136c48 --- /dev/null +++ b/encoding/sbcs.js @@ -0,0 +1,508 @@ +var util = require('util'), + Match = require ('../match'); + +/** + * This class recognizes single-byte encodings. Because the encoding scheme is so + * simple, language statistics are used to do the matching. + */ + +function NGramParser(theNgramList, theByteMap) { + var N_GRAM_MASK = 0xFFFFFF; + + this.byteIndex = 0; + this.ngram = 0; + + this.ngramList = theNgramList; + this.byteMap = theByteMap; + + this.ngramCount = 0; + this.hitCount = 0; + + this.spaceChar; + + /* + * Binary search for value in table, which must have exactly 64 entries. + */ + this.search = function(table, value) { + var index = 0; + + if (table[index + 32] <= value) { + index += 32; + } + + if (table[index + 16] <= value) { + index += 16; + } + + if (table[index + 8] <= value) { + index += 8; + } + + if (table[index + 4] <= value) { + index += 4; + } + + if (table[index + 2] <= value) { + index += 2; + } + + if (table[index + 1] <= value) { + index += 1; + } + + if (table[index] > value) { + index -= 1; + } + + if (index < 0 || table[index] != value) { + return -1; + } + + return index; + }; + + this.lookup = function(thisNgram) { + this.ngramCount += 1; + if (this.search(this.ngramList, thisNgram) >= 0) { + this.hitCount += 1; + } + }; + + this.addByte = function(b) { + this.ngram = ((this.ngram << 8) + (b & 0xFF)) & N_GRAM_MASK; + this.lookup(this.ngram); + } + + this.nextByte = function(det) + { + if (this.byteIndex >= det.fInputLen) { + return -1; + } + + return det.fInputBytes[this.byteIndex++] & 0xFF; + } + + // public int parse(CharsetDetector det) + // { + // return parse (det, (byte)0x20); + // } + + this.parse = function(det, spaceCh) { + var b; + var ignoreSpace = false; + this.spaceChar = spaceCh; + + while ((b = this.nextByte(det)) >= 0) { + var mb = this.byteMap[b]; + + // TODO: 0x20 might not be a space in all character sets... + if (mb != 0) { + if (!(mb == this.spaceChar && ignoreSpace)) { + this.addByte(mb); + } + + ignoreSpace = (mb == this.spaceChar); + } + } + + // TODO: Is this OK? The buffer could have ended in the middle of a word... + this.addByte(this.spaceChar); + + var rawPercent = this.hitCount / this.ngramCount; + +// if (rawPercent <= 2.0) { +// return 0; +// } + + // TODO - This is a bit of a hack to take care of a case + // were we were getting a confidence of 135... + if (rawPercent > 0.33) { + return 98; + } + + return Math.floor(rawPercent * 300.0); + }; +}; + +function NGramsPlusLang(la, ng) { + this.fLang = la; + this.fNGrams = ng; +}; + +function sbcs() {}; +sbcs.prototype.spaceChar = 0x20; +sbcs.prototype.ngrams = []; +sbcs.prototype.byteMap = []; +sbcs.prototype.match = function(det) { + + var multiple = (this.ngrams instanceof Array && this.ngrams[0] instanceof Array); + if (multiple) { + var bestConfidenceSoFar = -1; + var lang = null; + + for (var i = this.ngrams.length - 1; i >= 0; i--) { + var ngl = this.ngrams[i]; + + var parser = new NGramParser(ngl.fNGrams, this.byteMap); + var confidence = parser.parse(det, this.spaceChar); + if (confidence > bestConfidenceSoFar) { + bestConfidenceSoFar = confidence; + lang = ngl.fLang; + } + } + + var name = this.getName(det); + return bestConfidenceSoFar <= 0 ? null : new Match(det, this, bestConfidenceSoFar, name, lang); + } + + var parser = new NGramParser(this.ngrams, this.byteMap); + var confidence = parser.parse(det, this.spaceChar); + return confidence <= 0 ? null : new Match(det, this, confidence); +}; + + +module.exports.ISO_8859_1 = function() { + this.byteMap = [ + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0xAA, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0xB5, 0x20, 0x20, + 0x20, 0x20, 0xBA, 0x20, 0x20, 0x20, 0x20, 0x20, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF + ]; + + this.ngrams = [ + new NGramsPlusLang("da", [ + 0x206166, 0x206174, 0x206465, 0x20656E, 0x206572, 0x20666F, 0x206861, 0x206920, 0x206D65, 0x206F67, 0x2070E5, 0x207369, 0x207374, 0x207469, 0x207669, 0x616620, + 0x616E20, 0x616E64, 0x617220, 0x617420, 0x646520, 0x64656E, 0x646572, 0x646574, 0x652073, 0x656420, 0x656465, 0x656E20, 0x656E64, 0x657220, 0x657265, 0x657320, + 0x657420, 0x666F72, 0x676520, 0x67656E, 0x676572, 0x696765, 0x696C20, 0x696E67, 0x6B6520, 0x6B6B65, 0x6C6572, 0x6C6967, 0x6C6C65, 0x6D6564, 0x6E6465, 0x6E6520, + 0x6E6720, 0x6E6765, 0x6F6720, 0x6F6D20, 0x6F7220, 0x70E520, 0x722064, 0x722065, 0x722073, 0x726520, 0x737465, 0x742073, 0x746520, 0x746572, 0x74696C, 0x766572 + ]), + new NGramsPlusLang("de", [ + 0x20616E, 0x206175, 0x206265, 0x206461, 0x206465, 0x206469, 0x206569, 0x206765, 0x206861, 0x20696E, 0x206D69, 0x207363, 0x207365, 0x20756E, 0x207665, 0x20766F, + 0x207765, 0x207A75, 0x626572, 0x636820, 0x636865, 0x636874, 0x646173, 0x64656E, 0x646572, 0x646965, 0x652064, 0x652073, 0x65696E, 0x656974, 0x656E20, 0x657220, + 0x657320, 0x67656E, 0x68656E, 0x687420, 0x696368, 0x696520, 0x696E20, 0x696E65, 0x697420, 0x6C6963, 0x6C6C65, 0x6E2061, 0x6E2064, 0x6E2073, 0x6E6420, 0x6E6465, + 0x6E6520, 0x6E6720, 0x6E6765, 0x6E7465, 0x722064, 0x726465, 0x726569, 0x736368, 0x737465, 0x742064, 0x746520, 0x74656E, 0x746572, 0x756E64, 0x756E67, 0x766572 + ]), + new NGramsPlusLang("en", [ + 0x206120, 0x20616E, 0x206265, 0x20636F, 0x20666F, 0x206861, 0x206865, 0x20696E, 0x206D61, 0x206F66, 0x207072, 0x207265, 0x207361, 0x207374, 0x207468, 0x20746F, + 0x207768, 0x616964, 0x616C20, 0x616E20, 0x616E64, 0x617320, 0x617420, 0x617465, 0x617469, 0x642061, 0x642074, 0x652061, 0x652073, 0x652074, 0x656420, 0x656E74, + 0x657220, 0x657320, 0x666F72, 0x686174, 0x686520, 0x686572, 0x696420, 0x696E20, 0x696E67, 0x696F6E, 0x697320, 0x6E2061, 0x6E2074, 0x6E6420, 0x6E6720, 0x6E7420, + 0x6F6620, 0x6F6E20, 0x6F7220, 0x726520, 0x727320, 0x732061, 0x732074, 0x736169, 0x737420, 0x742074, 0x746572, 0x746861, 0x746865, 0x74696F, 0x746F20, 0x747320 + ]), + new NGramsPlusLang("es", [ + 0x206120, 0x206361, 0x20636F, 0x206465, 0x20656C, 0x20656E, 0x206573, 0x20696E, 0x206C61, 0x206C6F, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207265, 0x207365, + 0x20756E, 0x207920, 0x612063, 0x612064, 0x612065, 0x61206C, 0x612070, 0x616369, 0x61646F, 0x616C20, 0x617220, 0x617320, 0x6369F3, 0x636F6E, 0x646520, 0x64656C, + 0x646F20, 0x652064, 0x652065, 0x65206C, 0x656C20, 0x656E20, 0x656E74, 0x657320, 0x657374, 0x69656E, 0x69F36E, 0x6C6120, 0x6C6F73, 0x6E2065, 0x6E7465, 0x6F2064, + 0x6F2065, 0x6F6E20, 0x6F7220, 0x6F7320, 0x706172, 0x717565, 0x726120, 0x726573, 0x732064, 0x732065, 0x732070, 0x736520, 0x746520, 0x746F20, 0x756520, 0xF36E20 + ]), + new NGramsPlusLang("fr", [ + 0x206175, 0x20636F, 0x206461, 0x206465, 0x206475, 0x20656E, 0x206574, 0x206C61, 0x206C65, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207365, 0x20736F, 0x20756E, + 0x20E020, 0x616E74, 0x617469, 0x636520, 0x636F6E, 0x646520, 0x646573, 0x647520, 0x652061, 0x652063, 0x652064, 0x652065, 0x65206C, 0x652070, 0x652073, 0x656E20, + 0x656E74, 0x657220, 0x657320, 0x657420, 0x657572, 0x696F6E, 0x697320, 0x697420, 0x6C6120, 0x6C6520, 0x6C6573, 0x6D656E, 0x6E2064, 0x6E6520, 0x6E7320, 0x6E7420, + 0x6F6E20, 0x6F6E74, 0x6F7572, 0x717565, 0x72206C, 0x726520, 0x732061, 0x732064, 0x732065, 0x73206C, 0x732070, 0x742064, 0x746520, 0x74696F, 0x756520, 0x757220 + ]), + new NGramsPlusLang("it", [ + 0x20616C, 0x206368, 0x20636F, 0x206465, 0x206469, 0x206520, 0x20696C, 0x20696E, 0x206C61, 0x207065, 0x207072, 0x20756E, 0x612063, 0x612064, 0x612070, 0x612073, + 0x61746F, 0x636865, 0x636F6E, 0x64656C, 0x646920, 0x652061, 0x652063, 0x652064, 0x652069, 0x65206C, 0x652070, 0x652073, 0x656C20, 0x656C6C, 0x656E74, 0x657220, + 0x686520, 0x692061, 0x692063, 0x692064, 0x692073, 0x696120, 0x696C20, 0x696E20, 0x696F6E, 0x6C6120, 0x6C6520, 0x6C6920, 0x6C6C61, 0x6E6520, 0x6E6920, 0x6E6F20, + 0x6E7465, 0x6F2061, 0x6F2064, 0x6F2069, 0x6F2073, 0x6F6E20, 0x6F6E65, 0x706572, 0x726120, 0x726520, 0x736920, 0x746120, 0x746520, 0x746920, 0x746F20, 0x7A696F + ]), + new NGramsPlusLang("nl", [ + 0x20616C, 0x206265, 0x206461, 0x206465, 0x206469, 0x206565, 0x20656E, 0x206765, 0x206865, 0x20696E, 0x206D61, 0x206D65, 0x206F70, 0x207465, 0x207661, 0x207665, + 0x20766F, 0x207765, 0x207A69, 0x61616E, 0x616172, 0x616E20, 0x616E64, 0x617220, 0x617420, 0x636874, 0x646520, 0x64656E, 0x646572, 0x652062, 0x652076, 0x65656E, + 0x656572, 0x656E20, 0x657220, 0x657273, 0x657420, 0x67656E, 0x686574, 0x696520, 0x696E20, 0x696E67, 0x697320, 0x6E2062, 0x6E2064, 0x6E2065, 0x6E2068, 0x6E206F, + 0x6E2076, 0x6E6465, 0x6E6720, 0x6F6E64, 0x6F6F72, 0x6F7020, 0x6F7220, 0x736368, 0x737465, 0x742064, 0x746520, 0x74656E, 0x746572, 0x76616E, 0x766572, 0x766F6F + ]), + new NGramsPlusLang("no", [ + 0x206174, 0x206176, 0x206465, 0x20656E, 0x206572, 0x20666F, 0x206861, 0x206920, 0x206D65, 0x206F67, 0x2070E5, 0x207365, 0x20736B, 0x20736F, 0x207374, 0x207469, + 0x207669, 0x20E520, 0x616E64, 0x617220, 0x617420, 0x646520, 0x64656E, 0x646574, 0x652073, 0x656420, 0x656E20, 0x656E65, 0x657220, 0x657265, 0x657420, 0x657474, + 0x666F72, 0x67656E, 0x696B6B, 0x696C20, 0x696E67, 0x6B6520, 0x6B6B65, 0x6C6520, 0x6C6C65, 0x6D6564, 0x6D656E, 0x6E2073, 0x6E6520, 0x6E6720, 0x6E6765, 0x6E6E65, + 0x6F6720, 0x6F6D20, 0x6F7220, 0x70E520, 0x722073, 0x726520, 0x736F6D, 0x737465, 0x742073, 0x746520, 0x74656E, 0x746572, 0x74696C, 0x747420, 0x747465, 0x766572 + ]), + new NGramsPlusLang("pt", [ + 0x206120, 0x20636F, 0x206461, 0x206465, 0x20646F, 0x206520, 0x206573, 0x206D61, 0x206E6F, 0x206F20, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207265, 0x207365, + 0x20756D, 0x612061, 0x612063, 0x612064, 0x612070, 0x616465, 0x61646F, 0x616C20, 0x617220, 0x617261, 0x617320, 0x636F6D, 0x636F6E, 0x646120, 0x646520, 0x646F20, + 0x646F73, 0x652061, 0x652064, 0x656D20, 0x656E74, 0x657320, 0x657374, 0x696120, 0x696361, 0x6D656E, 0x6E7465, 0x6E746F, 0x6F2061, 0x6F2063, 0x6F2064, 0x6F2065, + 0x6F2070, 0x6F7320, 0x706172, 0x717565, 0x726120, 0x726573, 0x732061, 0x732064, 0x732065, 0x732070, 0x737461, 0x746520, 0x746F20, 0x756520, 0xE36F20, 0xE7E36F + ]), + new NGramsPlusLang("sv", [ + 0x206174, 0x206176, 0x206465, 0x20656E, 0x2066F6, 0x206861, 0x206920, 0x20696E, 0x206B6F, 0x206D65, 0x206F63, 0x2070E5, 0x20736B, 0x20736F, 0x207374, 0x207469, + 0x207661, 0x207669, 0x20E472, 0x616465, 0x616E20, 0x616E64, 0x617220, 0x617474, 0x636820, 0x646520, 0x64656E, 0x646572, 0x646574, 0x656420, 0x656E20, 0x657220, + 0x657420, 0x66F672, 0x67656E, 0x696C6C, 0x696E67, 0x6B6120, 0x6C6C20, 0x6D6564, 0x6E2073, 0x6E6120, 0x6E6465, 0x6E6720, 0x6E6765, 0x6E696E, 0x6F6368, 0x6F6D20, + 0x6F6E20, 0x70E520, 0x722061, 0x722073, 0x726120, 0x736B61, 0x736F6D, 0x742073, 0x746120, 0x746520, 0x746572, 0x74696C, 0x747420, 0x766172, 0xE47220, 0xF67220, + ]) + ]; + + this.getName = function(det) { + if (typeof det == 'undefined') + return "ISO-8859-1"; + return det.fC1Bytes ? "windows-1252" : "ISO-8859-1"; + }; +}; +util.inherits(module.exports.ISO_8859_1, sbcs); + + +module.exports.ISO_8859_2 = function() { + this.byteMap = [ + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0xB1, 0x20, 0xB3, 0x20, 0xB5, 0xB6, 0x20, + 0x20, 0xB9, 0xBA, 0xBB, 0xBC, 0x20, 0xBE, 0xBF, + 0x20, 0xB1, 0x20, 0xB3, 0x20, 0xB5, 0xB6, 0xB7, + 0x20, 0xB9, 0xBA, 0xBB, 0xBC, 0x20, 0xBE, 0xBF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0x20 + ]; + + this.ngrams = [ + new NGramsPlusLang("cs", [ + 0x206120, 0x206279, 0x20646F, 0x206A65, 0x206E61, 0x206E65, 0x206F20, 0x206F64, 0x20706F, 0x207072, 0x2070F8, 0x20726F, 0x207365, 0x20736F, 0x207374, 0x20746F, + 0x207620, 0x207679, 0x207A61, 0x612070, 0x636520, 0x636820, 0x652070, 0x652073, 0x652076, 0x656D20, 0x656EED, 0x686F20, 0x686F64, 0x697374, 0x6A6520, 0x6B7465, + 0x6C6520, 0x6C6920, 0x6E6120, 0x6EE920, 0x6EEC20, 0x6EED20, 0x6F2070, 0x6F646E, 0x6F6A69, 0x6F7374, 0x6F7520, 0x6F7661, 0x706F64, 0x706F6A, 0x70726F, 0x70F865, + 0x736520, 0x736F75, 0x737461, 0x737469, 0x73746E, 0x746572, 0x746EED, 0x746F20, 0x752070, 0xBE6520, 0xE16EED, 0xE9686F, 0xED2070, 0xED2073, 0xED6D20, 0xF86564, + ]), + new NGramsPlusLang("hu", [ + 0x206120, 0x20617A, 0x206265, 0x206567, 0x20656C, 0x206665, 0x206861, 0x20686F, 0x206973, 0x206B65, 0x206B69, 0x206BF6, 0x206C65, 0x206D61, 0x206D65, 0x206D69, + 0x206E65, 0x20737A, 0x207465, 0x20E973, 0x612061, 0x61206B, 0x61206D, 0x612073, 0x616B20, 0x616E20, 0x617A20, 0x62616E, 0x62656E, 0x656779, 0x656B20, 0x656C20, + 0x656C65, 0x656D20, 0x656E20, 0x657265, 0x657420, 0x657465, 0x657474, 0x677920, 0x686F67, 0x696E74, 0x697320, 0x6B2061, 0x6BF67A, 0x6D6567, 0x6D696E, 0x6E2061, + 0x6E616B, 0x6E656B, 0x6E656D, 0x6E7420, 0x6F6779, 0x732061, 0x737A65, 0x737A74, 0x737AE1, 0x73E967, 0x742061, 0x747420, 0x74E173, 0x7A6572, 0xE16E20, 0xE97320, + ]), + new NGramsPlusLang("pl", [ + 0x20637A, 0x20646F, 0x206920, 0x206A65, 0x206B6F, 0x206D61, 0x206D69, 0x206E61, 0x206E69, 0x206F64, 0x20706F, 0x207072, 0x207369, 0x207720, 0x207769, 0x207779, + 0x207A20, 0x207A61, 0x612070, 0x612077, 0x616E69, 0x636820, 0x637A65, 0x637A79, 0x646F20, 0x647A69, 0x652070, 0x652073, 0x652077, 0x65207A, 0x65676F, 0x656A20, + 0x656D20, 0x656E69, 0x676F20, 0x696120, 0x696520, 0x69656A, 0x6B6120, 0x6B6920, 0x6B6965, 0x6D6965, 0x6E6120, 0x6E6961, 0x6E6965, 0x6F2070, 0x6F7761, 0x6F7769, + 0x706F6C, 0x707261, 0x70726F, 0x70727A, 0x727A65, 0x727A79, 0x7369EA, 0x736B69, 0x737461, 0x776965, 0x796368, 0x796D20, 0x7A6520, 0x7A6965, 0x7A7920, 0xF37720, + ]), + new NGramsPlusLang("ro", [ + 0x206120, 0x206163, 0x206361, 0x206365, 0x20636F, 0x206375, 0x206465, 0x206469, 0x206C61, 0x206D61, 0x207065, 0x207072, 0x207365, 0x2073E3, 0x20756E, 0x20BA69, + 0x20EE6E, 0x612063, 0x612064, 0x617265, 0x617420, 0x617465, 0x617520, 0x636172, 0x636F6E, 0x637520, 0x63E320, 0x646520, 0x652061, 0x652063, 0x652064, 0x652070, + 0x652073, 0x656120, 0x656920, 0x656C65, 0x656E74, 0x657374, 0x692061, 0x692063, 0x692064, 0x692070, 0x696520, 0x696920, 0x696E20, 0x6C6120, 0x6C6520, 0x6C6F72, + 0x6C7569, 0x6E6520, 0x6E7472, 0x6F7220, 0x70656E, 0x726520, 0x726561, 0x727520, 0x73E320, 0x746520, 0x747275, 0x74E320, 0x756920, 0x756C20, 0xBA6920, 0xEE6E20, + ]) + ]; + + this.getName = function(det) { + if (typeof det == 'undefined') + return "ISO-8859-2"; + return det.fC1Bytes ? "windows-1250" : "ISO-8859-2"; + }; +}; +util.inherits(module.exports.ISO_8859_2, sbcs); + + +module.exports.ISO_8859_5 = function() { + this.byteMap = [ + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0x20, 0xFE, 0xFF, + 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, + 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, + 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0x20, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0x20, 0xFE, 0xFF + ]; + + this.ngrams = [ + 0x20D220, 0x20D2DE, 0x20D4DE, 0x20D7D0, 0x20D820, 0x20DAD0, 0x20DADE, 0x20DDD0, 0x20DDD5, 0x20DED1, 0x20DFDE, 0x20DFE0, 0x20E0D0, 0x20E1DE, 0x20E1E2, 0x20E2DE, + 0x20E7E2, 0x20EDE2, 0xD0DDD8, 0xD0E2EC, 0xD3DE20, 0xD5DBEC, 0xD5DDD8, 0xD5E1E2, 0xD5E220, 0xD820DF, 0xD8D520, 0xD8D820, 0xD8EF20, 0xDBD5DD, 0xDBD820, 0xDBECDD, + 0xDDD020, 0xDDD520, 0xDDD8D5, 0xDDD8EF, 0xDDDE20, 0xDDDED2, 0xDE20D2, 0xDE20DF, 0xDE20E1, 0xDED220, 0xDED2D0, 0xDED3DE, 0xDED920, 0xDEDBEC, 0xDEDC20, 0xDEE1E2, + 0xDFDEDB, 0xDFE0D5, 0xDFE0D8, 0xDFE0DE, 0xE0D0D2, 0xE0D5D4, 0xE1E2D0, 0xE1E2D2, 0xE1E2D8, 0xE1EF20, 0xE2D5DB, 0xE2DE20, 0xE2DEE0, 0xE2EC20, 0xE7E2DE, 0xEBE520 + ]; + + this.getName = function(det) { + return "ISO-8859-5"; + }; + + this.getLanguage = function() { + return "ru"; + }; +}; +util.inherits(module.exports.ISO_8859_5, sbcs); + + +module.exports.ISO_8859_6 = function() { + this.byteMap = [ + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, + 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, + 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, + 0xD8, 0xD9, 0xDA, 0x20, 0x20, 0x20, 0x20, 0x20, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20 + ]; + + this.ngrams = [ + 0x20C7E4, 0x20C7E6, 0x20C8C7, 0x20D9E4, 0x20E1EA, 0x20E4E4, 0x20E5E6, 0x20E8C7, 0xC720C7, 0xC7C120, 0xC7CA20, 0xC7D120, 0xC7E420, 0xC7E4C3, 0xC7E4C7, 0xC7E4C8, + 0xC7E4CA, 0xC7E4CC, 0xC7E4CD, 0xC7E4CF, 0xC7E4D3, 0xC7E4D9, 0xC7E4E2, 0xC7E4E5, 0xC7E4E8, 0xC7E4EA, 0xC7E520, 0xC7E620, 0xC7E6CA, 0xC820C7, 0xC920C7, 0xC920E1, + 0xC920E4, 0xC920E5, 0xC920E8, 0xCA20C7, 0xCF20C7, 0xCFC920, 0xD120C7, 0xD1C920, 0xD320C7, 0xD920C7, 0xD9E4E9, 0xE1EA20, 0xE420C7, 0xE4C920, 0xE4E920, 0xE4EA20, + 0xE520C7, 0xE5C720, 0xE5C920, 0xE5E620, 0xE620C7, 0xE720C7, 0xE7C720, 0xE8C7E4, 0xE8E620, 0xE920C7, 0xEA20C7, 0xEA20E5, 0xEA20E8, 0xEAC920, 0xEAD120, 0xEAE620 + ]; + + this.getName = function(det) { + return "ISO-8859-6"; + }; + + this.getLanguage = function() { + return "ar"; + }; +}; +util.inherits(module.exports.ISO_8859_6, sbcs); + + +module.exports.ISO_8859_7 = function() { + this.byteMap = [ + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, + 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, + 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0xA1, 0xA2, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xDC, 0x20, + 0xDD, 0xDE, 0xDF, 0x20, 0xFC, 0x20, 0xFD, 0xFE, + 0xC0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0x20, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xFB, 0xDC, 0xDD, 0xDE, 0xDF, + 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, + 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, + 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, + 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0x20 + ]; + + this.ngrams = [ + 0x20E1ED, 0x20E1F0, 0x20E3E9, 0x20E4E9, 0x20E5F0, 0x20E720, 0x20EAE1, 0x20ECE5, 0x20EDE1, 0x20EF20, 0x20F0E1, 0x20F0EF, 0x20F0F1, 0x20F3F4, 0x20F3F5, 0x20F4E7, + 0x20F4EF, 0xDFE120, 0xE120E1, 0xE120F4, 0xE1E920, 0xE1ED20, 0xE1F0FC, 0xE1F220, 0xE3E9E1, 0xE5E920, 0xE5F220, 0xE720F4, 0xE7ED20, 0xE7F220, 0xE920F4, 0xE9E120, + 0xE9EADE, 0xE9F220, 0xEAE1E9, 0xEAE1F4, 0xECE520, 0xED20E1, 0xED20E5, 0xED20F0, 0xEDE120, 0xEFF220, 0xEFF520, 0xF0EFF5, 0xF0F1EF, 0xF0FC20, 0xF220E1, 0xF220E5, + 0xF220EA, 0xF220F0, 0xF220F4, 0xF3E520, 0xF3E720, 0xF3F4EF, 0xF4E120, 0xF4E1E9, 0xF4E7ED, 0xF4E7F2, 0xF4E9EA, 0xF4EF20, 0xF4EFF5, 0xF4F9ED, 0xF9ED20, 0xFEED20 + ]; + + this.getName = function(det) { + if (typeof det == 'undefined') + return "ISO-8859-7"; + return det.fC1Bytes ? "windows-1253" : "ISO-8859-7"; + }; + + this.getLanguage = function() { + return "el"; + }; +}; +util.inherits(module.exports.ISO_8859_7, sbcs); + + + diff --git a/encoding/unicode.js b/encoding/unicode.js new file mode 100644 index 0000000..9298352 --- /dev/null +++ b/encoding/unicode.js @@ -0,0 +1,106 @@ +var util = require('util'), + Match = require ('../match'); + +/** + * This class matches UTF-16 and UTF-32, both big- and little-endian. The + * BOM will be used if it is present. + */ + +module.exports.UTF_16BE = function() { + this.name = "UTF-16BE"; + this.match = function(det) { + var input = det.fRawInput; + + if (input.length>=2 && ((input[0] & 0xFF) == 0xFE && (input[1] & 0xFF) == 0xFF)) { + var confidence = 100; + return new Match(det, this, confidence); + } + + // TODO: Do some statistics to check for unsigned UTF-16BE + return null; + } +}; + +module.exports.UTF_16LE = function() { + this.name = "UTF-16LE"; + this.match = function(det) { + var input = det.fRawInput; + + if (input.length >= 2 && ((input[0] & 0xFF) == 0xFF && (input[1] & 0xFF) == 0xFE)) { + // An LE BOM is present. + if (input.length>=4 && input[2] == 0x00 && input[3] == 0x00) { + // It is probably UTF-32 LE, not UTF-16 + return null; + } + var confidence = 100; + return new Match(det, this, confidence); + } + + // TODO: Do some statistics to check for unsigned UTF-16LE + return null; + } +}; + +function UTF_32() {}; +UTF_32.prototype.match = function(det) { + var input = det.fRawInput, + limit = (det.fRawLength / 4) * 4, + numValid = 0, + numInvalid = 0, + hasBOM = false, + confidence = 0; + + if (limit==0) { + return null; + } + + if (this.getChar(input, 0) == 0x0000FEFF) { + hasBOM = true; + } + + for(var i = 0; i < limit; i += 4) { + var ch = this.getChar(input, i); + + if (ch < 0 || ch >= 0x10FFFF || (ch >= 0xD800 && ch <= 0xDFFF)) { + numInvalid += 1; + } else { + numValid += 1; + } + } + + // Cook up some sort of confidence score, based on presence of a BOM + // and the existence of valid and/or invalid multi-byte sequences. + if (hasBOM && numInvalid==0) { + confidence = 100; + } else if (hasBOM && numValid > numInvalid*10) { + confidence = 80; + } else if (numValid > 3 && numInvalid == 0) { + confidence = 100; + } else if (numValid > 0 && numInvalid == 0) { + confidence = 80; + } else if (numValid > numInvalid*10) { + // Probably corrupt UTF-32BE data. Valid sequences aren't likely by chance. + confidence = 25; + } + + // return confidence == 0 ? null : new CharsetMatch(det, this, confidence); + return confidence == 0 ? null : new Match(det, this, confidence); +}; + +module.exports.UTF_32BE = function() { + this.name = "UTF_32BE"; + this.getChar = function(input, index) { + return (input[index + 0] & 0xFF) << 24 | (input[index + 1] & 0xFF) << 16 | + (input[index + 2] & 0xFF) << 8 | (input[index + 3] & 0xFF); + }; +}; +util.inherits(module.exports.UTF_32BE, UTF_32); + +module.exports.UTF_32LE = function() { + this.name = "UTF_32LE"; + this.getChar = function(input, index) { + return (input[index + 3] & 0xFF) << 24 | (input[index + 2] & 0xFF) << 16 | + (input[index + 1] & 0xFF) << 8 | (input[index + 0] & 0xFF); + }; +}; +util.inherits(module.exports.UTF_32LE, UTF_32); diff --git a/encoding/utf8.js b/encoding/utf8.js new file mode 100644 index 0000000..c0cd442 --- /dev/null +++ b/encoding/utf8.js @@ -0,0 +1,84 @@ + +var Match = require ('../match'); + +/** + * Charset recognizer for UTF-8 + */ +module.exports = function() { + this.name = "UTF-8"; + this.match = function(det) { + + var hasBOM = false, + numValid = 0, + numInvalid = 0, + input = det.fRawInput, + i, + trailBytes = 0, + confidence; + + if (det.fRawLength >= 3 && + (input[0] & 0xFF) == 0xef && (input[1] & 0xFF) == 0xbb && (input[2] & 0xFF) == 0xbf) { + hasBOM = true; + } + + // Scan for multi-byte sequences + for (i=0; i 5) { + break; + } + trailBytes = 0; + } + + // Verify that we've got the right number of trail bytes in the sequence + for (;;) { + i++; + if (i>=det.fRawLength) { + break; + } + b = input[i]; + if ((b & 0xc0) != 0x080) { + numInvalid++; + break; + } + if (--trailBytes == 0) { + numValid++; + break; + } + } + } + + // Cook up some sort of confidence score, based on presense of a BOM + // and the existence of valid and/or invalid multi-byte sequences. + confidence = 0; + if (hasBOM && numInvalid==0) { + confidence = 100; + } else if (hasBOM && numValid > numInvalid*10) { + confidence = 80; + } else if (numValid > 3 && numInvalid == 0) { + confidence = 100; + } else if (numValid > 0 && numInvalid == 0) { + confidence = 80; + } else if (numValid == 0 && numInvalid == 0) { + // Plain ASCII. + confidence = 10; + } else if (numValid > numInvalid*10) { + // Probably corruput utf-8 data. Valid sequences aren't likely by chance. + confidence = 25; + } + return confidence == 0 ? null : new Match(det, this, confidence); + }; +}; diff --git a/match.js b/match.js new file mode 100644 index 0000000..b8ec5a6 --- /dev/null +++ b/match.js @@ -0,0 +1,6 @@ + +module.exports = function(det, rec, confidence) { + // console.log(det, rec, confidence); + // this.res = 1; + this.confidence = confidence; +} \ No newline at end of file