akiss-xyz / certainpy

A short and sweet library handling uncertainty in calculations. Can use both standard, probabilistic uncertainties and maximal uncertainties for arbitrary functions over arbitrary variables.

Certainpy - a library to deal with uncertainty in all your calculations!

What is this?

This library provides two basic classes: an "Uncertain_Value" class that represents a number with uncertainty, for example something like 3 +- 0.1, and an associated "Vec" class that represents a 2D vector where each co-ordinate has an associated, independent uncertainty.

These classes then have a lot of the standard operations you expect from floats (basic arithmetic operations, exponentiation, and arbitrary function execution) and vectors (addition, subtraction, conversion between Polar and Cartestian) implemented (the Vec class needs a lot of work - any help would be much appreciated).

How do I use this?

Simply from certainpy import *.

And from there you can create your values and start doing whatever you like with them:

Using this project

This project is under the GNU GPL 3, with no warranty or fitness for any particular purpose. See LICENSE for details. Don't hesitate to get in touch for anything!

certainpy's Issues

Evaluating arbitrary functions - design a process for checking whether f has a turning point

When a function is evaluated by nonprob_evalfun(fun, args, no_turning=False), the optional parameter indicates whether the function has a turning point within the uncertain region.

Instead of asking the user to provide this information, we could attempt to figure it out, hence making function evaluation more powerful.

Keeping track of the use of the same variable

If you have a function with terms of the same variable, we currently have no way of realising that their uncertainties aren't independent.