SrMise - Peak extraction and peak fitting tool for atomic pair distribution functions.

## Project description

Tool for unbiased peak extraction from atomic pair distribution functions.

The diffpy.srmise package is an implementation of the ParSCAPE algorithm for peak extraction from atomic pair distribution functions (PDFs). It is designed to function even when a priori knowledge of the physical sample is limited, utilizing the Akaike Information Criterion (AIC) to estimate whether peaks are statistically justified relative to alternate models. Three basic use cases are anticipated for diffpy.srmise. The first is peak fitting a user-supplied collections of peaks. The second is peak extraction from a PDF with no (or only partial) user-supplied peaks. The third is an AIC-driven multimodeling analysis where the output of multiple diffpy.srmise trials are ranked.

The framework for peak extraction defines peak-like clusters within the data, extracts a single peak within each cluster, and iteratively combines nearby clusters while performing a recursive search on the residual to identify occluded peaks. Eventually this results in a single global cluster containing many peaks fit over all the data. Over- and underfitting are discouraged by use of the AIC when adding or removing (during a pruning step) peaks. Termination effects, which can lead to physically spurious peaks in the PDF, are incorporated in the mathematical peak model and the pruning step attempts to remove peaks which are fit better as termination ripples due to another peak.

Where possible, diffpy.srmise provides physically reasonable default values for extraction parameters. However, the PDF baseline should be estimated by the user before extraction, or by performing provisional peak extraction with varying baseline parameters. The package defines a linear (crystalline) baseline, arbitrary polynomial baseline, a spherical nanoparticle baseline, and an arbitrary baseline interpolated from a list of user-supplied values. In addition, PDFs with accurate experimentally-determined uncertainties are necessary to provide the most reliable results, but historically such PDFs are rare. In the absence of accurate uncertainties an ad hoc uncertainty must be specified.

## REQUIREMENTS

The diffpy.srfit package requires Python 2.6 or 2.7 and the following software:

• setuptools - software distribution tools for Python

• NumPy - numerical mathematics and fast array operations for Python

• SciPy - scientific libraries for Python

• matplotlib - python plotting library

On Ubuntu Linux, the required software can easily be installed using the system package manager:

sudo apt-get install \
python-setuptools python-numpy python-scipy python-matplotlib

For Mac OS X systems with the MacPorts package manager, the required software can be installed with

sudo port install \
python27 py27-setuptools py27-numpy py27-scipy py27-matplotlib

When installing for MacPorts, make sure the MacPorts bin directory is the first in the system PATH and that python27 is selected as the default Python version in MacPorts:

sudo port select --set python python27

python get-pip.py

It is recommended to install all other dependencies using prebuilt binaries. Visit http://www.scipy.org and http://www.matplotlib.org for instructions. Alternately, install a full Python distribution such as Python(x,y) or Enthought Canopy which already includes the required components.

## INSTALLATION

The simplest way to obtain diffpy.srmise on Unix, Linux, and Mac systems is using easy_install or pip to download and install the latest release from the Python Package Index.

sudo pip diffpy.srmise

If you prefer to install from sources, make sure all required software packages are in place and then run

sudo python setup.py install

This installs diffpy.srmise for all users in the default system location. If administrator (root) access is not available, see the usage info from python setup.py install --help for options to install to user-writable directories.

To install on Windows run either of the commands above omitting sudo.

## DEVELOPMENT

diffpy.srmise is open-source software developed with support of the Center of Research Excellence in Complex Materials at Michigan State University, in cooperation with the DiffPy-CMI complex modeling initiative at the Brookhaven National Laboratory. The diffpy.srmise sources are hosted at https://github.com/diffpy/diffpy.srmise.

Feel free to fork the project and contribute. To install diffpy.srmise in a development mode, with its sources being directly used by Python rather than copied to a package directory, use

python setup.py develop --user

## ACKNOWLEDGEMENT

The source code of pdfdataset.py was derived from diffpy.pdfgui.

## CONTACTS

http://www.diffpy.org

or email Prof. Simon Billinge at sb2896@columbia.edu.

## Project details

Uploaded source