public BinaryTree(Node i_node, ref BinaryTree i_left, ref BinaryTree i_right)
{
IsInTheCurrentLevel = true;
Node = i_node;
Left = i_left;
Right = i_right;
}
// create the Huffman tree based on the counted frequency
private void CreateHuffmanTree()
{
uint positiveCount = GetNumberOfPositiveValuesInFrequencyTable();
m_HuffmanTree = new BinaryTree[2 * positiveCount - 1];
for (uint i = 0; i < positiveCount; i++)
{
byte existingByte = GetExistingByteAtIndex(i);
m_HuffmanTree[i] = new BinaryTree(new Node(existingByte, GetFrequencyOfByte(existingByte)));
}
for (uint i = positiveCount; i < 2 * positiveCount - 1; i++)
{
SortHuffmanTree();
m_HuffmanTree[i - 1].IsInTheCurrentLevel = false;
m_HuffmanTree[i - 2].IsInTheCurrentLevel = false;
m_HuffmanTree[i] = new BinaryTree(new Node(m_HuffmanTree[i - 1].Node.Frequency + m_HuffmanTree[i - 2].Node.Frequency), ref m_HuffmanTree[i - 1], ref m_HuffmanTree[i - 2]);
}
SortHuffmanTree();
BinaryTree finalHuffmanTree = m_HuffmanTree[2 * positiveCount - 2];
Utility.PrintDebug(finalHuffmanTree.ToString());
Utility.PrintDebug(finalHuffmanTree.GetHeight().ToString());
Utility.PrintDebug();
m_HuffmanDictionary = new Dictionary<byte, uint>();
CreateHuffmanMap(ref finalHuffmanTree, 1);
foreach (byte key in m_HuffmanDictionary.Keys)
{
Utility.PrintDebug("HuffmanDict[ " + key + " ] = " + m_HuffmanDictionary[key]);
}
Utility.PrintDebug();
}
private void CreateHuffmanMap(ref BinaryTree i_tree, uint i_code)
{
if (i_tree.Left != null)
{
BinaryTree tempTree = i_tree.Left;
CreateHuffmanMap(ref tempTree, 2 * i_code + 0);
}
if (i_tree.Right != null)
{
BinaryTree tempTree = i_tree.Right;
CreateHuffmanMap(ref tempTree, 2 * i_code + 1);
}
if (i_tree.Left == null && i_tree.Right == null)
{
if (m_HuffmanDictionary.ContainsValue(i_code))
{
throw new System.ArgumentException("Code alrady in the Huffman map");
}
m_HuffmanDictionary[i_tree.Node.Byte] = i_code;
}
}
static void Main(string[] args)
{
bool Compress(ref HuffmanData data)
{
FrequencyTable frequencyTable = new FrequencyTable();
foreach (byte b in data.uncompressedData)
{
frequencyTable.Add(b);
}
/*
* Creation of the forest which is represented by a list of binary trees.
* We add one BinaryTree containing one Node representing a symbol for all the different possible symbols (byte).
* Each Node will thus have its symbol (byte) and its frequency (int).
*/
Forest forest = new Forest();
foreach (byte b in frequencyTable.Keys)
{
forest.Add(new BinaryTree(new Node(b, frequencyTable[b])));
}
/*
* This function is in charge of getting a unique binary tree from all the binary trees contained in the Forest.
* While the Forest has several trees, we remove the two trees containing the two nodes that have the lowest frequency
* and assign them to the left and right nodes of the root of a new BinaryTree that we finally had to the forest.
*/
BinaryTree binaryTree = forest.GetUniqueTree();
/*
* This is where the Huffman codes are created.
* We go all through our BinaryTree thanks to the classic preorder method.
* When we reach a leaf, we add our code to dictionaries that we'll be using to compress our data.
* A code is represented by a List of Boolean.
*/
forest.Preorder(binaryTree.Root, new Code());
Dictionary <Byte, Code> codeTable = forest.getCodeTable();
Dictionary <Code, Byte> decodeTable = forest.getDecodeTable();
/*
* We get the number of bytes that are required to store our compressed data on a array of bytes.
* To do it, for each symbol (byte), we multiply its number of occurrences in our uncompressed data by the number of Boolean
* used to code the symbol. If the result is divisible by eight, we proceed to the division and this is our seeked number. If not,
* we add one extra byte.
*/
int GetRequiredBytesNumber()
{
int size = 0;
foreach (Byte b in frequencyTable.Keys)
{
size += frequencyTable[b] * codeTable[b].Count;
}
return(size % 8 == 0 ? size / 8 : size / 8 + 1);
}
int RequiredBytesNumber = GetRequiredBytesNumber();
/*
* Here starts the compressing.
*/
Byte[] compressedData = new Byte[RequiredBytesNumber];
/*
* First, for all our symbols in our uncompressed data, let's add their Huffman code to a List of Booleans.
* This List only contains the bits that are very significant according to our Huffman Code.
*/
List <bool> compressedDataBoolean = new List <bool>();
foreach (Byte b in data.uncompressedData)
{
compressedDataBoolean.AddRange(codeTable[b]);
}
int sizeDataUncompressed = compressedDataBoolean.Count;
/*
* This function will put all the booleans contained in our list, in bytes (eight by eight) to store them into
* an array of bytes.
*/
Byte[] BooleanListToByteArray(List <bool> boolList)
{
Byte[] _compressedData = new Byte[RequiredBytesNumber];
Byte curr = 0;
int i = 0, j = 0;
Boolean b;
while (boolList.Count > 0)
{
b = boolList.Last();
boolList.RemoveAt(boolList.Count - 1);
if (i == 0)
{
// If we iterate over a boolean that must be the first to be put in a byte ...
curr = new byte();
curr = 0x0;
}
if (b)
{
// If our boolean is true, then we need to put a one in our byte at the right position.
curr = (Byte)((1 << i) | curr);
}
else
{
// Same if it's false.
curr = (Byte)((0 << i) | curr);
}
if (i == 7)
{
// If we iterate over a boolean that will fill our byte, then we add the byte to our array.
_compressedData[j] = curr;
j++;
i = 0; // We get the 'byte pointer' ready for the next one.
}
else
{
i++;
}
}
// If we didn't get enough booleans to fill the final byte,
// we had the byte to the array (the other bits of the byte are zero by default).
if (i != 0)
{
_compressedData[j] = curr;
}
return(_compressedData);
}
compressedData = BooleanListToByteArray(compressedDataBoolean);
/*
* We need to pass our frequency table as a List<KeyValuePair<Byte, int>>.
* It's required by the struct used by the tool.
*/
List <KeyValuePair <Byte, int> > frequency = frequencyTable.ToList();
data.compressedData = compressedData;
data.frequency = frequency;
data.sizeOfUncompressedData = data.uncompressedData.Length;
return(true);
}
bool Decompress(ref HuffmanData data)
{
/*
* We start by receiving the frequency table already created and we create (again) our forest from it.
* We then re-generate the Huffman codes for each symbol.
*/
Forest forest2 = new Forest();
int necessarybytes = 0;
foreach (KeyValuePair <Byte, int> kvp in data.frequency)
{
forest2.Add(new BinaryTree(new Node(kvp.Key, kvp.Value)));
necessarybytes += kvp.Value;
}
BinaryTree binaryTree2 = forest2.GetUniqueTree();
forest2.Preorder(binaryTree2.Root, new Code());
/*
* To decompress our data, we must parse our compressed data. For an easier writing of the code, We will parse
* it from a BitArray, so that we don't have to jump from byte to byte.
*/
BitArray bitArray = new BitArray(data.compressedData);
/*
* In some cases, the first byte isn't full of significant bits for our Huffman encoding. It means, that the first
* zeros must be ignored (because they're only here to fill the byte). So we need to start parsing after the non-significant
* bits. In order to do that, we first retrieve the number of bits that are significant (sizeDataUncompressed).
* We just need to adjust our index (start) according to that number.
*/
int sizeDataUncompressed = 0;
foreach (KeyValuePair <Byte, int> kvp in data.frequency)
{
sizeDataUncompressed += kvp.Value * forest2.getCodeTable()[kvp.Key].Count;
}
int start = sizeDataUncompressed % 8 == 0 ? 0 : 8 - sizeDataUncompressed % 8;
/*
* This is the function that will parse the full stream of bits, and using the Huffman Table, will get each Symbol.
*/
forest2.ProduceByteArrayNew(bitArray, binaryTree2.Root, necessarybytes, start);
Byte[] finalArray = forest2.GetByteArray();
data.uncompressedData = finalArray;
return(true);
}
/*
* Actual Main instructions used to test the proper functioning of the program.
*/
byte[] testString = System.IO.File.ReadAllBytes(Path.GetDirectoryName(Process.GetCurrentProcess().MainModule.FileName) + @"\..\..\testString.txt");
HuffmanData mainData = new HuffmanData();
mainData.uncompressedData = testString;
Compress(ref mainData);
Decompress(ref mainData);
}
static void Main(string[] args)
{
//class variables
#region Variables
StreamReader txtIn;
StreamWriter txtOut = new StreamWriter(String.Format("encoding{0}", args[1]));
byte b = 0;
int pow = 7;
char ch;
string encoding;
string line;
char delimiter = 'ยป';
#endregion
//check for proper usage
#region Proper Usage
if (args.Length != 2)
{
Console.WriteLine("Proper usage is: program.exe inFile outFile");
Console.ReadKey();
return;
}
//
try
{
txtIn = new StreamReader(args[0]);
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
Console.ReadLine();
return;
}
#endregion
//read the file and count the character frequencies
#region List of Character Frequencies
LinkedList <CharacterFrequency> list = new LinkedList <CharacterFrequency>();
char c = (char)txtIn.Read();
while (!txtIn.EndOfStream)
{
findFrequency(c, list);
c = (char)txtIn.Read();
}
findFrequency(c, list);
txtIn.Close();
#endregion
//create a sorted linked list of binary tree nodes
#region Sorted List of binary Tree Nodes of Character Frequencies
SortedLinkedList <BinaryTreeNode <CharacterFrequency> > sll = new SortedLinkedList <BinaryTreeNode <CharacterFrequency> >();
foreach (CharacterFrequency charf in list)
{
sll.Add(new BinaryTreeNode <CharacterFrequency>(charf, charf.Frequency));
}
#endregion
//while the list is not empty, build the binary tree nodes, adding the two smallest values
#region Building the Tree
while (sll.Count > 1)
{
BinaryTreeNode <CharacterFrequency> left;
BinaryTreeNode <CharacterFrequency> right;
BinaryTreeNode <CharacterFrequency> phNode;
left = sll.First.Value;
sll.RemoveFirst();
right = sll.First.Value;
sll.RemoveFirst();
phNode = new BinaryTreeNode <CharacterFrequency>(null, (left.Value + right.Value));
phNode.Left = left;
phNode.Right = right;
sll.Add(phNode);
}
BinaryTree tree = new BinaryTree(sll.First.Value);
tree.inOrder(tree.Root, "");
//write the encoding table to a new file
foreach (EncodingData d in tree.EncodingTable)
{
txtOut.WriteLine(String.Format("{0}", d.ToString()));
}
#endregion
//compress the file
#region Compression
txtIn = new StreamReader(args[0]);
StreamWriter compressedFile = new StreamWriter(args[1]);
while (!txtIn.EndOfStream)
{
ch = (char)txtIn.Read();
encoding = findEncoding(ch, tree.EncodingTable);
foreach (char character in encoding)
{
if (character == '1')
{
b = (byte)(b | (byte)Math.Pow(2, pow));
}
pow--;
if (pow < 0)
{
pow = 7;
compressedFile.Write((char)b);
b = 0;
}
}
}
if (pow != 7)
{
compressedFile.Write((char)b);
}
txtOut.Close();
txtIn.Close();
compressedFile.Close();
#endregion
//decompress the file
#region Decompression
txtIn = new StreamReader(String.Format("encoding{0}", args[1]));
//decompression
BinaryTree charTree = new BinaryTree(new BinaryTreeNode <CharacterFrequency>(new CharacterFrequency((char)0, 0), 0));
BinaryTreeNode <CharacterFrequency> node = charTree.Root;
while (!txtIn.EndOfStream)
{
line = txtIn.ReadLine();
string[] holder;
holder = line.Split(delimiter);
if (holder.Length > 1)
{
foreach (char letter2 in holder[1])
{
if (letter2 == '0')
{
if (node.Left == null)
{
node.Left = new BinaryTreeNode <CharacterFrequency>(new CharacterFrequency((char)0, 0), 0);
}
node = node.Left;
}
else
{
if (node.Right == null)
{
node.Right = new BinaryTreeNode <CharacterFrequency>(new CharacterFrequency((char)0, 0), 0);
}
node = node.Right;
}
}
byte holdingByte = 0;
Byte.TryParse(holder[0], out holdingByte);
node.Data.Ch = (char)holdingByte;
node = charTree.Root;
}
}
txtIn.Close();
txtIn = new StreamReader(args[1]);
StreamWriter decompressedFile = new StreamWriter(String.Format("decompressed{0}", args[0]));
byte decomp;
byte decompVar;
pow = 7;
node = charTree.Root;
while (!txtIn.EndOfStream)
{
decomp = (byte)txtIn.Read();
while (pow >= 0)
{
decompVar = (byte)(decomp & (byte)(Math.Pow(2, pow)));
if (decompVar > 0)
{
//if the variable is greater than 0, go to the right
node = node.Right;
}
else
{
node = node.Left;
}
if (node.isLeaf())
{
decompressedFile.Write(node.Data);
node = charTree.Root;
}
pow--;
}
pow = 7;
}
decompressedFile.Close();
txtIn.Close();
txtOut.Close();
#endregion
Console.ReadKey();
}