A toy implementation of file compression with Canonical Huffman Encoding.
Compiling and running the program:
This program was written and tested on Ubuntu 16.04 with the mono compiler.
Assuming you have a similar setup, simply running make should work.
If you have a different setup/compiler the relevant files for compression are:
Compressor.cs (Main method),
SymbolNode.cs,
InternalNode.cs,
LeafNode.cs,
HuffmanTree.cs,
SymbolPriorityQueue.cs,
CodeBook.cs
The files for decompression are:
Decompressor.cs (Main method),
BitBuffer.cs
Note that because of potential instability, no files are ever destroyed or overwritten during the compression/decompression process, so a failure on the part of the program shouldn't result in the loss/corruption of data. Having said that, the program shouldn't have any trouble compressing/decompressing Unicode encoded text, which is its intended use.
Format for the outputted file:
section format: [(byte count)contents]
if Huffman compression is used:
[(1) 1, indicating Huffman compression][(4) number of input symbols (not necessarily the length of the file in bytes however)][(1) length of the codebook (and therefore number of unique symbols)][(var)code book][(var)encoded file]
The codebook is stored as follows:
(...[(1-2)Unicode Character][(1)length of the character's code]...)
Note that because the length of a given character's code is expressed by a single byte, the maximum length (and therefore maximum count of unique symbols) is limited to 255.
if count compression is used:
[(1) 0, indicating count compression][(1-2)symbol][(4)count]