AjayMT / text-compressor

text-compressor

This is a very simple text compression algorithm based on Huffman coding.

Build

git clone http://github.com/AjayMT/text-compressor.git
cd text-compressor
make

Usage

There are two components to this program: the encoder and the decoder.

encoder

encoder <tree-destination>

encoder will read text from stdin, compress it and write it to stdout. Additionally, it will store the produced Huffman tree (which is required for decoding) in a file (tree-destination).

Example: This will compress harry-potter.txt into harry-potter-compressed and store the tree in harry-potter-tree.

encoder harry-potter-tree < harry-potter.txt > harry-potter-compressed

decoder

decoder <tree-path>

decoder will read compressed text from stdin, decompress it (using data from the tree file at tree-path) and write it to stdout.

Example: This will decompress harry-potter-compressed into harry-potter.txt using data from harry-potter-tree.

decoder harry-potter-tree < harry-potter-compressed > harry-potter.txt

Bugs and limitations

• decoder sometimes writes an extraneous character at the end. This is because encoder zeroes out leftover bits (since the compressed text may not fit into 8-bit characters perfectly), which are sometimes read as another character.
• Having to store the tree separately is cumbersome.
• Compression using Huffman coding is not particularly effective when all the characters in the text occur at the same (or very similar) frequency.
• The current implementation is not as efficient as it could (or should) be.

Author(s)

• Ajay Tatachar ([e-mail])