An educational compression experimentation toolkit built in Go. This CLI tool allows you to experiment with different compression algorithms, analyze their performance, and understand compression theory through hands-on examples.

• 🔧 Multiple Compression Algorithms:

  • RLE (Run-Length Encoding) - Best for data with consecutive repeated characters
  • Huffman Coding - Variable-length encoding based on character frequency
  • LZ77 - Dictionary-based compression using sliding windows
  • LZW (Lempel-Ziv-Welch) - Dictionary-based compression with dynamic dictionary

• 📊 Comprehensive Metrics:

  • Compression ratio
  • Space savings percentage
  • Shannon entropy analysis
  • Compression/decompression performance timing

• 📈 Visualization:

  • ASCII bar charts for compression ratios
  • Entropy visualization with interpretation
  • Side-by-side algorithm comparison
  • Color-coded output for clarity

• 🎯 Educational Focus:

  • Clean, readable algorithm implementations
  • Detailed metrics to understand compression behavior
  • Support for comparing multiple algorithms

• Go 1.16 or higher

git clone https://github.com/BaseMax/go-compress-lab.git cd go-compress-lab go build -o compress-lab ./cmd/compress-lab

Optionally, install it globally:

go install ./cmd/compress-lab

# Compress text with all algorithms (comparison mode) ./compress-lab -text="Hello World" -compare # Compress a file with a specific algorithm ./compress-lab -input=file.txt -algo=huffman # Compress and save to file ./compress-lab -input=file.txt -algo=lzw -output=compressed.bin # Decompress a file ./compress-lab -input=compressed.bin -algo=lzw -decompress -output=original.txt

• -text - Text string to compress (alternative to input file)
• -input - Input file path
• -output - Output file path for compressed/decompressed data
• -algo - Algorithm to use: rle, huffman, lz77, lzw, or all
• -compare - Compare all algorithms (shows detailed metrics)
• -decompress - Decompress mode instead of compress

./compress-lab -text="AAAABBBCCCCCCDDDDDD" -compare

Output:

Comparing all compression algorithms... Input size: 19 bytes ==================================================================================================== COMPRESSION RESULTS ==================================================================================================== Algorithm | Original | Compressed | Ratio | Savings % | Entropy | Comp Time | Decomp Time ---------------------------------------------------------------------------------------------------- RLE | 19 B | 8 B | 2.38 | 57.89% | 1.94 | 2.284µs | 360ns Huffman | 19 B | 23 B | 0.83 | -21.05% | 1.94 | 13.244µs | 8.937µs LZ77 | 19 B | 25 B | 0.76 | -31.58% | 1.94 | 1.503µs | 1.353µs LZW | 19 B | 22 B | 0.86 | -15.79% | 1.94 | 40.856µs | 36.188µs ==================================================================================================== COMPRESSION RATIO VISUALIZATION ------------------------------------------------------------ RLE | ████████████████████████████████████████ 2.38:1 Huffman | █████████████ 0.83:1 LZ77 | ████████████ 0.76:1 LZW | ██████████████ 0.86:1 ------------------------------------------------------------ ENTROPY ANALYSIS ------------------------------------------------------------ Shannon Entropy: 1.9440 bits/byte Randomness: 24.30% [▓▓▓▓▓▓▓▓▓░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░] Interpretation: → Very low entropy: Highly repetitive data, excellent for compression ------------------------------------------------------------ 

Analysis: RLE performs best on this data because it efficiently encodes consecutive runs of identical characters.

./compress-lab -text="The quick brown fox jumps over the lazy dog." -compare

Analysis: LZW and Huffman typically perform better on natural language text due to repeated patterns and character frequency distribution.

# Compress a file ./compress-lab -input=document.txt -algo=lzw -output=document.lzw # Decompress it ./compress-lab -input=document.lzw -algo=lzw -decompress -output=restored.txt

./compress-lab -input=data.txt -algo=huffman

This shows detailed metrics for just the Huffman algorithm.

• Formula: Original Size / Compressed Size
• Interpretation: Higher is better. Ratio > 1 means compression, < 1 means expansion.

• Formula: (1 - Compressed/Original) × 100%
• Interpretation: Positive percentage = compression, negative = expansion.

• Range: 0 to 8 bits/byte
• Interpretation:
  • 0-2: Very low entropy, highly compressible
  • 2-4: Low entropy, good compression potential
  • 4-6: Medium entropy, moderate compression
  • 6-8: High entropy, difficult to compress

• Best for: Data with long runs of repeated characters
• How it works: Replaces sequences of repeated bytes with (count, byte) pairs
• Example: "AAAA" → (4, 'A')

• Best for: Data with non-uniform character distribution
• How it works: Assigns shorter codes to frequent characters
• Example: 'e' might be encoded as "01" while 'z' as "1101100"

• Best for: Data with repeated patterns at various distances
• How it works: Uses a sliding window to find and reference previous occurrences
• Example: References previous data with (offset, length, next_char) triplets

• Best for: Data with repeated substrings
• How it works: Builds a dictionary of patterns on-the-fly
• Example: Commonly used in GIF and TIFF formats

go-compress-lab/ ├── cmd/ │ └── compress-lab/ │ └── main.go # CLI application entry point ├── pkg/ │ ├── algorithms/ │ │ ├── compressor.go # Compressor interface │ │ ├── rle.go # RLE implementation │ │ ├── huffman.go # Huffman coding implementation │ │ ├── lz77.go # LZ77 implementation │ │ └── lzw.go # LZW implementation │ ├── metrics/ │ │ └── metrics.go # Compression metrics and calculations │ └── visualization/ │ └── display.go # Terminal output formatting ├── go.mod └── README.md 

1. Understanding Compression Theory: Compare algorithms on different types of data to see which performs best
2. Entropy Analysis: Learn about information theory and data randomness
3. Performance Benchmarking: Measure compression speed vs. ratio tradeoffs
4. Algorithm Behavior: Observe how each algorithm handles different data patterns

Contributions are welcome! This is an educational project, so please keep implementations clear and well-documented.

MIT License - See LICENSE file for details

Max Base

This project is designed for educational purposes to help understand compression algorithms and their behavior on different types of data.