iminuit 1.2

Interactive Minimization Tools based on MINUIT

iminuit

Interactive IPython-Friendly Minimizer based on SEAL Minuit2. (A slightly modified version is included in the package no need to install it separately)

It is designed from ground up to be fast, interactive and cython friendly. iminuit extracts function signature very permissively starting from checking func_code down to last resort of parsing docstring(or you could tell iminuit to stop looking and take your answer). The interface is inspired heavily by PyMinuit and the status printout is inspired by ROOT Minuit. iminuit is mostly compatible with PyMinuit(with few exceptions). Existing PyMinuit code can be ported to iminuit by just changing the import statement.

In a nutshell:

from iminuit import Minuit
def f(x, y, z):
    return (x - 2) ** 2 + (y - 3) ** 2 + (z - 4) ** 2
m = Minuit(f)
m.migrad()
print(m.values)  # {'x': 2,'y': 3,'z': 4}
print(m.errors)  # {'x': 1,'y': 1,'z': 1}

Install

First install cython (pip install cython), then

python setup.py install

or from pip:

pip install iminuit

For Windows, Christoph Gohlke made a nice windows binary to save you all from Windows compilation nightmare:

http://www.lfd.uci.edu/~gohlke/pythonlibs/#iminuit

Tutorial

All the tutorials are in tutorial directory. You can view it online too.

• Quick start
• Hard Core Cython tutorial. If you need to do a huge likelihood fit that needs speed, or you want to learn how to parallelize your stuff, this is for you.

Documentation

http://iminuit.readthedocs.org/

Technical Stuff

Using it as a black box is a bad idea. Here are some fun reads; the order is given by the order I think you should read.

• Wikipedia for Quasi Newton Method and DFP formula. The magic behind migrad.
• Variable Metric Method for Minimization William Davidon 1991
• A New Approach to Variable Metric Algorithm. (R.Fletcher 1970)
• Original Paper: MINUIT - A SYSTEM FOR FUNCTION MINIMIZATION AND ANALYSIS OF THE PARAMETER ERRORS AND CORRELATIONS by Fred James and Matts Roos.

You can help

Github allows you to contribute to this project very easily just fork the repository, make changes and submit a pull request.

Here’s the list of concrete open issues and feature requests: https://github.com/iminuit/iminuit

More generally any contribution to the docs, tests and package itself is welcome!

• Documentation. Tell us what’s missing, what’s incorrect or misleading.
• Tests. If you have an example that shows a bug or problem, please file an issue!
• Performance. If you are a C/cython/python hacker go ahead and make it faster.