This project serves as a basis for the practical part of my bachelor thesis at the Vienna University of Technology. Currently, it is still a work in progress.
This project implements various pseudorandom and random number generators (RNGs). In most cases, I attempted to closely follow the given description, where available. In other cases, I implemented my own solutions, e.g. for the true random number generators (TRNGs) that utilize some of the sensors available on the computer (camera, audio input).
Please note that the code here is for research purposes. Both good and bad RNGs are included here. I make no guarantees about the quality. Be careful before using any of the generators for scientific or cryptographic purposes!
Generators are in the module generator. All generators implement the default Python's random.Random class. Additionally, they also provide (at least) methods random_bit() and random_byte(). See abstract classes in generator.generator for signatures of various types of generators (in particular BitGenerator and NumberGenerator).
Statistical tests for (pseudo)random number generators are in the module randomness_test. These can be used to test whether a given generator (or a sequence) can be considered "random".
Unittests of the generators and statistical tests are provided in the test module.
In the root directory, scripts rng.py and run_all_tests.py can be executed in shell to generate random numbers/bits and to test the generators. See their usage for details.
This project requires Python version 3. It has been developed and tested with version 3.5.
Additionally, following libraries are required:
- NumPy and SciPy for numerical calculations and analysis
- pyDes for 3DES encryption
- Pygame for camera access for TRNG
On linux systems, you can install easily with pip:
pip install numpy scipy pydes pygame
For other systems, refer to the links.