Huffman Coding based Compression Library
Background
Huffman code is used to compress data file, where the data is represented as a sequence of characters. Huffman's greedy algorithm uses a table giving how often each character occurs; it then uses this table to build up an optimal way of representing each character as a binary string. We call the binary string the codeword for that character. A property of Huffman code is that it is a prefix code, i.e., in Huffman coding, no codeword is a prefix of some other codeword. The advantage of prefix code is that it makes decoding easier, as we do not need to use delimiter between two successive codewords. Given the frequency of each of the character, we can devise a greedy algorithm for finding the optimal Huffman codeword of each of the characters. For details of the greedy algorithm, INTRODUCTION TO ALGORITHMS 3rd edition by CORMEN please read Section 16.3 of the textbook.
In this assignment, we will build a compression library that compress text files using Huffman coding scheme. This library will have two programs: compress, and decompress; compress accepts a text file and produces a compressed representation of that text file; decompress accepts a file that was compressed with the compress program, and recovers the original file.
Implementation
Input to the compress is a text le with arbitrary size, but for this assignment we will sippose that data structure of the file fits in the main memory of a computer. Output of the program is a compressed representation of the original file. You will have to save the codetable in the header of the compressed file, so that you could use the codetable for decompressing the compressed file. Input to the decompress is a compressed file, from that the program recovers the original file. For sanity check, you must have a specific magic word at some position in the header of the compressed file, so that decompress can recognize whether the given file is a valid Huffman compressed file.
You must pay attention to the following issues:
The file that we would use for testing can be very large, having size in Gigabytes, so make sure that your program is bug-free and it works for large input file.
Write efficient algorithm, we will take off as much as 20 points if we feel that the program is taking unusually long time.
You must make sure that your program runs on a Linux Machine, and identically follows the formatting instructions. For formatting error, as much as 15 points can be taken off .
You should provide a Make file to compile your programs. Also, a README.txt file must be provided that will have the instruction to compile and run the programs.
Command-Line options
Compression:
C++: ./compress -f myfile.txt [-o myfile.hzip -s
Java: sh compress.sh -f myfile.txt [-o myfile.hzip -s]
Decompression:
C++: ./decompress -f myfile.hzip[-r -v]
Java: sh decompress.sh -f myfile.hzip [-o myfile.txt -s]
The command-line options that are within the square bracket are optional. The option \-f" precedes the input le name, which always has a .txt extension. The \-s" option prints statistics, such as for compression it prints, how many distinct characters are there, what is the compression ratio, and the wall clock time that it took for performing the compression task. For decompression, it prints how many character were written, and the wall clock time it took for performing the decompression task. The \-o" option precedes the name of an output le. If the output file name is not given, then we will append .hzip at the end of the input filename to create the output filename.