//
// This code parses binary format of WebM file.
//  recognizes only some important TAGs
//
// Limitation:
//  This programs reads all binary at once in memory (100MB).
//  It is very bad imprementation, but it is still enough for some small WebM file. 
//
// Usage:
//  node parse_webm.js filename
//
// Refer:
//  http://www.matroska.org/technical/specs/index.html
//
 
var path = require('path');
var fs   = require('fs');
var util = require('util');
 

// check args
var args = process.argv.length;
if (args != 3) {
 console.log('usage: node parse_webm.js filename');
 return -1;
}

var filename = process.argv[2];
console.log('START parsing: ' + filename);
var fd = fs.openSync(filename, 'r');

var buf_size = 100*1024*1024; // 100MB
var buffer = new Buffer(buf_size);
var position = 0;
//var tagBytes[4];
var tagSize = 0;
var dataSizeSize = 0;
var dataSize = 0;

var tagDictionary = setupTagDictionary();

var readBytes = fs.readSync(fd, buffer, 0, buf_size, 0);
console.log('readBytes=' + readBytes + "  " + addrHex(readBytes));
console.log(' ');

// --- parse webm ----
parseWebm(0, buffer, position, readBytes);
console.log('---- END ----- ');
return 0;

// ============

function parseWebm(level, buffer, position, maxPosition) {
 while (position < maxPosition) {
  var spc = spacer(level);
  
  // -- ADDRESS --
  console.log(spc + 'ADDR 0x' + addrHex(position) + ' -- Level:' + level + ' BEGIN' );
 
  // -- TAG --
  var result = scanWebmTag(buffer, position);
  if (! result) {
   console.log(spc + 'TAG scan end');
   break; 
  }
  var tagName = tagDictionary[result.str];
  console.log(spc + 'Tag size=' + result.size + ' Tag=' + result.str + ' <' + tagName + '> TagVal=' + result.value); 
  position += result.size;
 
  // --- DATA SIZE ---
  result = scanDataSize(buffer, position);
  if (! result) {
   console.log(spc + 'DATA SIZE scan end');
   break; 
  }
  console.log(spc + 'DataSize size=' + result.size + ' DataSize str=' + result.str + ' DataSize Val=' + result.value);
  position += result.size;
 
  // ---- DATA ----
  if (tagName === 'EBML') {
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'Tracks') {
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'TrackEntry') {
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'CodecName') {
   var codec = scanDataUTF8(buffer, position, result.value);
   console.log(spc + 'codecName=' + codec);
  }
  else if (tagName === 'Video') {
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'PixelWidth') {
   var width = scanDataValueU(buffer, position, result.value);
   console.log(spc + 'pixelWidth=' + width);
  }
  else if (tagName === 'PixelHeight') {
   var height = scanDataValueU(buffer, position, result.value);
   console.log(spc + 'pixelHeight=' + height);
  }
  else if (tagName === 'FrameRate') {
   var rate = scanDataFloat(buffer, position, result.value);
   console.log(spc + 'frameRate=' + rate);
  }
  else if (tagName === 'Segment') {
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'Cluster') {   
   parseWebm(level+1, buffer, position, (position + result.value));
  }
  else if (tagName === 'Timecode') {
   var timecode = scanDataValueU(buffer, position, result.value);
   console.log(spc + 'timecode=' + timecode);
   
   return position;
  }
 
  if (result.value >= 0) {
   position += result.value;
  }
  else {
   console.log(spc + 'DATA SIZE ffffffff.. cont.')
  }
  console.log(' ');
 
  // -- check EOF ---
  if (position == maxPosition) {
   console.log(spc + '--level:' + level + ' reached END---');
   break;
  }
  else if (position > maxPosition) {
   console.log(spc + '--level:' + level + ' --OVER END---' + ' pos=' + position + ' max=' + maxPosition );
   break;
  }
 }
 
 return position;
}


function addrHex(pos) {
 var str = '00000000' + pos.toString(16);
 var len = str.length;
 return str.substring(len - 8).toUpperCase();
}

function byteToHex(b) {
 var str = '0' + b.toString(16);
 var len = str.length;
 return str.substring(len - 2).toUpperCase();
}

function spacer(level) {
 var str = '          ';
 str = str.substring(0, level);
 return str;
}

function setupTagDictionary() {
 // T - Element Type - The form of data the element contains. m: Master, u: unsigned int, i: signed integer, s: string, 8: UTF-8 string, b: binary, f: float, d: date
 
 var tagDict = new Array();
 tagDict['[1A][45][DF][A3]'] = 'EBML'; // EBML 0	[1A][45][DF][A3] m
 tagDict['[42][86]'] = 'EBMLVersion'; //EBMLVersion	1	[42][86] u
 tagDict['[42][F7]'] =  'EBMLReadVersion'; // EBMLReadVersion	1	[42][F7] u
 tagDict['[42][F2]'] =  'EBMLMaxIDLength'; // EBMLMaxIDLength	1	[42][F2] u
 tagDict['[42][F3]'] =  'EBMLMaxSizeLength'; // EBMLMaxSizeLength	1	[42][F3] u
 tagDict['[42][82]'] =  'DocType'; // DocType	1	[42][82] s
 tagDict['[42][87]'] =  'DocTypeVersion'; // DocTypeVersion	1	[42][87] u
 tagDict['[42][85]'] =  'DocTypeReadVersion'; // DocTypeReadVersion	1	[42][85] u
 
 tagDict['[EC]'] =  'Void'; // Void	g	[EC] b
 tagDict['[BF]'] =  'CRC-32'; // CRC-32	g	[BF] b
 tagDict['[1C][53][BB][6B]'] =  'Cues'; // Cues	1	[1C][53][BB][6B] m
 
 tagDict['[18][53][80][67]'] = 'Segment';  // Segment	0	[18][53][80][67] m
 tagDict['[11][4D][9B][74]'] = 'SeekHead'; // SeekHead	1	[11][4D][9B][74] m
 tagDict['[4D][BB]'] = 'Seek'; // Seek	2	[4D][BB] m
 tagDict['[53][AB]'] = 'SeekID'; // SeekID	3	[53][AB] b
 tagDict['[53][AC]'] = 'SeekPosition'; // SeekPosition	3	[53][AC] u
  
 tagDict['[15][49][A9][66]'] = 'Info'; // Info	1	[15][49][A9][66] m 

 tagDict['[16][54][AE][6B]'] = 'Tracks'; // Tracks	1	[16][54][AE][6B] m
 tagDict['[AE]'] = 'TrackEntry'; // TrackEntry	2	[AE] m
 tagDict['[D7]'] = 'TrackNumber'; // TrackNumber	3	[D7] u
 tagDict['[73][C5]'] = 'TrackUID'; // TrackUID	3	[73][C5] u
 tagDict['[83]'] = 'TrackType'; // TrackType	3	[83] u
 tagDict['[23][E3][83]'] = 'DefaultDuration'; // DefaultDuration	3	[23][E3][83] u
 tagDict['[23][31][4F]'] = 'TrackTimecodeScale'; // TrackTimecodeScale	3	[23][31][4F] f
 tagDict['[86]'] = 'CodecID'; // CodecID	3	[86] s
 tagDict['[63][A2]'] = 'CodecPrivate'; // CodecPrivate	3	[63][A2] b
 tagDict['[25][86][88]'] = 'CodecName'; // CodecName	3	[25][86][88] 8
 tagDict['[E0]'] = 'Video'; // Video	3	[E0] m
 tagDict['[B0]'] = 'PixelWidth'; // PixelWidth	4	[B0] u
 tagDict['[BA]'] = 'PixelHeight'; // PixelHeight	4	[BA] u
 tagDict['[23][83][E3]'] = 'FrameRate'; // FrameRate	4	[23][83][E3] f
 tagDict['[E1]'] = 'Audio'; // Audio	3	[E1] m
 tagDict['[B5]'] = 'SamplingFrequency'; // SamplingFrequency	4	[B5] f
 tagDict['[9F]'] = 'Channels'; // Channels	4	[9F] u
 
 
 
 
 
 tagDict['[1F][43][B6][75]'] = 'Cluster'; // Cluster	1	[1F][43][B6][75] m
 tagDict['[E7]'] = 'Timecode'; // Timecode	2	[E7] u
 tagDict['[A3]'] = 'SimpleBlock'; // SimpleBlock	2	[A3] b
 
 return tagDict;
}

function scanWebmTag(buff, pos) {
 var tagSize = 0;
 var followByte;
 var firstByte = buff.readUInt8(pos);
 var firstMask = 0xff;
 
 if (firstByte & 0x80) {
  tagSize = 1;
 }
 else if (firstByte & 0x40) {
  tagSize = 2;
 }
 else if (firstByte & 0x20) {
  tagSize = 3;
 }
 else if (firstByte & 0x10) {
  tagSize = 4;
 }
 else {
  console.log('ERROR: bad TAG byte');
  return null;
 }

 var decodeRes = decodeBytes(buff, pos, tagSize, firstByte, firstMask); 
 return decodeRes;
}


function scanDataSize(buff, pos) {
 var dataSizeSize = 0;
 var followByte;
 var firstByte = buff.readUInt8(pos);
 var firstMask;
 
 if (firstByte & 0x80) {
  dataSizeSize = 1;
  firstMask = 0x7f;
 }
 else if (firstByte & 0x40) {
  dataSizeSize = 2;
  firstMask = 0x3f;
 }
 else if (firstByte & 0x20) {
  dataSizeSize = 3;
  firstMask = 0x1f;
 }
 else if (firstByte & 0x10) {
  dataSizeSize = 4;
  firstMask = 0x0f;
 }
 else if (firstByte & 0x08) {
  dataSizeSize = 5;
  firstMask = 0x07;
 }
 else if (firstByte & 0x04) {
  dataSizeSize = 6;
  firstMask = 0x03;
 }
 else if (firstByte & 0x02) {
  dataSizeSize = 7;
  firstMask = 0x01;
 }
 else if (firstByte & 0x01) {
  dataSizeSize = 8;
  firstMask = 0x00;
 }
 else {
  console.log('ERROR: bad DATA byte');
  return null;
 }
  
 var decodeRes = decodeBytes(buff, pos, dataSizeSize, firstByte, firstMask); 
 return decodeRes;
}

function scanDataValueU(buff, pos, size) {
 var uVal = 0;
 var byteVal;
 for (var i = 0; i < size; i++) {
  byteVal = buff.readUInt8(pos + i);
  //console.log('scanDataValueU pos=' + pos + ' i=' + i + ' byte=' + byteToHex(byteVal));
  uVal = (uVal << 8) + byteVal;
 }
 
 return uVal;
}

function scanDataUTF8(buff, pos, size) {
 var sVal = buff.toString('utf8', pos, pos+size);
 return sVal;
}

function scanDataFloat(buff, pos, size) {
 if (size === 4) {
  var f = buff.readFloatBE(pos);
  return f;
 }
 else if (size === 8) {
  var df = buff.readDoubleBE(pos);
  return df;
 }
 else {
  console.error('ERROR. Bad Float size=' + size);
  return null;
 }
}



function decodeBytes(buff, pos, size, firstByte, firstMask) {
 var value = firstByte & firstMask;
 var str = ('[' + byteToHex(firstByte) + ']');
 var followByte;
 for (var i = 1; i < size; i++) {
   followByte = buff.readUInt8(pos + i);
   str += '[';
   str += byteToHex(followByte);
   str += ']';
   value = (value << 8) + followByte;
 }
 
 var res = {};
 res.str = str;
 res.size = size;
 res.value = value;
 
 return res;
}