README.md

Python implementation of Lempel-Ziv Complexity

This repository contains a small, simple and efficient implementation of the Lempel-Ziv complexity algorithm.

Examples

Simple usage

If the lempel_ziv_complexity.py file is accessible in your PATH or in Python's path:

>>> from lempel_ziv_complexity import lempel_ziv_complexity
>>> s = '1001111011000010'
>>> lempel_ziv_complexity(s)  # 1 / 0 / 01 / 11 / 10 / 110 / 00 / 010
8

Documentation

See this file.

If you are really interested about the details of how the algorithm works, please see this GIF screencast from PythonTutor.com.

• Reference: this short lecture note by Peter Shor (Oct.2005).

With the C extension

If the lempel_ziv_complexity.so file is accessible in your PATH or in Python's path:

Small benchmark

There is also a Cython version, to have a faster implementation:

$ ipython
...
>>> s = '1001111011000010'
>>> %timeit lempel_ziv_complexity(s)
6.1 µs ± 33.6 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
>>> %timeit lempel_ziv_complexity_cython(s)
132 ns ± 2.55 ns per loop (mean ± std. dev. of 7 runs, 10000000 loops each)

The speedup is typically between ×50 and ×100.

Demo on a Jupyter notebook

See this notebook: on nbviewever, which also compares with the Julia version.

---

Install and build

Manually ?

Easy!

Clone this repository, go in the folder, compile, test, and if it works, install it.

cd /tmp/
git clone https://GitHub.com/Naereen/Lempel-Ziv_Complexity
cd Lempel-Ziv_Complexity/src/
make build
make test     # should pass
make install  # mv the build/lib*/*.so files where you need them

Be sure to include the dynamic library when you need it, or in a folder accessible by your Python interpreter (somewhere in sys.path). The file is lempel_ziv_complexity_cython.so (for Python 2) or the lempel_ziv_complexity_cython.cpython-35m-x86_64-linux-gnu.so (for Python 3.5, or higher, adapt the name).

With pip ?

This project is hosted on the Pypi package repository.

sudo pip install lempel_ziv_complexity
# test it
python -c "from lempel_ziv_complexity import lempel_ziv_complexity; print(lempel_ziv_complexity('1001111011000010') == 8)"  # test

---

Julia implementation ?

I was curious and wanted to write the same algorithm in Julia. Here it is: lempel_ziv_complexity.jl. See this part of the notebook for a benchmark between Julia and Python.

The Julia package is published here: Naereen/LempelZiv.jl, and see here for its documentation.

---

About

Language?

Python v2.7+ or Python v3.1+.

• Numba can be used to speed up the pure Python version.
• Cython is needed to build the C extension (faster).

📜 License ?

MIT Licensed (file LICENSE). © Lilian Besson, 2017-2019.