Compression Project

This project provides a platform for testing and benchmarking compression algorithms using trained dictionaries.

Features

• Training of compression dictionaries.
• Compression and decompression of files using trained dictionaries.
• Measurement of the execution time for dictionary training, compression, and decompression.
• Reporting of average, minimum, and maximum performance metrics and compression ratios.

Getting Started

Prerequisites

• Ensure you have a C compiler, preferably GCC or Clang.
• Install the ZSTD, LZ4, libdeflate libraries. You can use make install if you use Linux.

Compilation

make clean build

Usage

After compilation, run the program using:

./compr_alg

The program will search for data samples in the ./data directory, train a dictionary using these samples, and then compress and decompress each file using the trained dictionary.

Output

The program provides a comprehensive report on the compression and decompression processes, including:

• Time taken for dictionary training, compression, and decompression.
• Compression ratio statistics.
• Any discrepancies between the original and decompressed files.

+----------------------------------+--------------+--------------+--------------+
| Metric                           | Average      | Min          | Max          |
+----------------------------------+--------------+--------------+--------------+
| Dictionary training time (s)     |     0.000624 |     0.005911 |     0.017405 |
| Compression time (s)             |     0.000018 |     0.000007 |     0.000062 |
| Decompression time (s)           |     0.000012 |     0.000009 |     0.000036 |
| Compression ratio                |     0.101090 |     0.021484 |     1.013672 |
+----------------------------------+--------------+--------------+--------------+

For generating graph report you can use view/analysis.py