EntropyVUE

Visualise the structure and entropy of files.

It uses Tkinter, has no third-party dependencies, and requires Python 3.10 or newer.

Features

• Interactive Visualisation (Vue): The application displays normalised symbol-count profiles, log-scaled count profiles, and Shannon entropy for the selected file.
• Cross-Platform: EntropyVUE uses cross-platform Python and the embedded Tkinter library for the GUI.
• Customisability: Users can configure the MAX_HEIGHT, ALPHABET, and Scale parameters. ALPHABET controls the number of symbol bins. Values up to 256 map individual bytes into bins; larger values use fixed-width multi-byte symbols.

Displayed measurements

• Normalised symbol-count profile: Raw symbol/bin counts scaled to the chart height. This is a distribution profile, not an entropy metric.
• Normalised log2/log10 count profile: Log-scaled count profiles for exposing low-frequency structure.
• Shannon entropy of input: Shannon entropy over the observed symbol/bin distribution, reported in bits per symbol.
• Shannon entropy of log2-transformed counts: Shannon entropy over the log2-transformed count distribution.

Usage

You can start the program by double-clicking on vue.py, or by passing vue.py to Python,

python vue.py

The program will start in "demo" mode, and you can specify a file to open using the "Select File" button.

Alternatively, you can run the program from the command line with the -f or --file option

python vue.py -f /path/to/file

Discover EntropyVUE interactivity

• How does a text file's symbol-count profile compare to a ZIP archive?
• How does a deflate ZIP archive compare to a compressed RAR or 7-zip archive?
• What changes in distribution shape and Shannon entropy between an MP3 file and a WAV file?
• How does a JPG file compare to a PNG file?

What additional structure can be extracted from these distribution and entropy readings?

---

Tim Abdiukov