ALSchwalm / foresight

A tool for predicting the output of random number generators

Foresight

foresight is a python library for predicting the output of random number generators across a variety of platforms and languages including:

• glibc
• MSVC
• PHP
• Java
• MySQL

This repository also contains a simple wrapper (called foresee) around this library that provides a CLI.

Installation

To install foresight, clone this repository and run python setup.py install.

Library Usage

The library is divided into subpackages for each supported platform. These subpackages then provide submodules for each supported function. For example,the following sample will predict the output of the glibc rand_r from some prior outputs:

from foresight.glibc import rand_r

prior_outputs = [int(i) for i in input().split()]
future_outputs = rand_r.from_outputs(prior_outputs)

for output in future_outputs:
    print(output)

Generally, each module provides the following functions:

• from_seed(seed): get a generator to the future values returned by the RNG when seeded with the given value.

• from_outputs(prev_values): attempt to predict the internal state of the RNG from the given prior outputs. If successful, returns a generator to the future values returned by the RNG.

CLI Usage

Command line usage of the library has the following general form:

foresee <platform> <function> [<options>]

For example, the above script is equivalent to the following foresee command:

foresee glibc rand_r -o <the prior outputs>

By default, an infinite stream is generated. The output may be limited to the next N outputs with -c N.

Tests

Tests can be run with nose by first installing the package with pip install nose, and then running nosetest in the project root.

License

This project is licensed under the terms of the MIT license. See the LICENSE file for details

Contributing

Keeping in mind that this project is still fairly unstable, contributions are welcome.