Skip to content

Latest commit

 

History

History
62 lines (43 loc) · 3.04 KB

readme.md

File metadata and controls

62 lines (43 loc) · 3.04 KB
Raw

Archiver

I realised archiver, which can can compress any number of files into one. Also, my program can decompress one archive into original files.

Algorithm

I used Huffman coding to compress and decompress my files. Also, my algorithm haven't loss of data during compression. I was reading char from files, encoded this char by binary sequence in trie. Then every encoded char is vertex. Moreover, no one terminate vertex isn't parent of other terminate vertex. Because of it we can unambiguously decode out data. Using this trie we create Canonical Huffman code. Also, we encode special symbols:

  • FILENAME_END = 256
  • ONE_MORE_FILE = 257
  • ARCHIVE_END = 258

They help us to separate different files and data from filename, which encodes too. Length of our char is 9 bits. Also, we put into archive information about alphabet and lengths of encoded chars. Using this information, we can retrieve binary sequence for every char. After we read a data and decode it.

Program build

After copying of data into local directory we need write in terminal following:

  • cmake PROJECT_ROOT
  • make

Now we can use our executable file PROJECT_NAME. If you want to run your program then write in terminal following:

  • ./PROJECT_NAME

Program usage

You can launch this program with three different flags:

  1. ./PROJECT_NAME -c ARCHIVE_NAME INPUT_FILE_NAME_1 INPUT_FILE_NAME_2 ...
  2. ./PROJECT_NAME -d ARCHIVE_NAME
  3. ./PROJECT_NAME -h

First variant compress your files INPUT_FILE_NAME_1, INPUT_FILE_NAME_2 ... into ARCHIVE_NAME. if file ARCHIVE_NAME already exists then this file will be overwritten. ARCHIVE_NAME is necessary parameter. Also, you should write at least one INPUT_FILE_NAME parameter.

Second variant decompress ARCHIVE_NAME into files with their original names. If these files already exists then they will be overwritten. ARCHIVE_NAME is necessary parameter.

Third variant writes into standard output stream usage help.

Benchmarks

Filename Input file size Output file size Ratio Compression time Decompression time Compression speed Decompression speed
document.docx 13,85 KB 13,04 KB 1,0621 0 min 0,055 sec 0 min 0,176 sec 251,85 KB/sec 74,1 KB/sec
music.mp3 7,15 MB 7,11 MB 1,0045 0 min 17,815 sec 1 min 29,054 sec 410,76 KB/sec 81,8 KB/sec
photo.jpg 379,3 KB 369,66 KB 1,0261 0 min 0,963 sec 0 min 4,34 sec 393,88 KB/sec 85,18 KB/sec
video.mkv 653,92 MB 382,55 MB 1,7094 25 min 32,402 sec 89 min 29,937 sec 436,97 KB/sec 72,95 KB/sec
bigdocument.doc 4,07 MB 2,42 MB 1,6796 0 min 9,603 sec 0 min 30,321 sec 433,83 KB/sec 81,81 KB/sec

We can see that decompression speed is roughly equal, but compression speed increases due to grow of input data. Also, big natural text and video is files with the best coefficient of compression.