PyXRF is a python-based sophisticated fluorescence analysis package for fitting and visualizing X-ray fluorescence data. This package contains a high-level fitting engine, comprehensive command-line/GUI design, rigorous physics calculation and a powerful visualization interface. The theoretical part of PyXRF is based on MAPS, developed by Stefan Vogt at APS. PyXRF offers some of the unique features as follows.
- Automatic elements finding: Users do not need to spend extra time selecting elements manually.
- Forward calculation: Users can observe the spectrum from forward calculation at real time while adjusting input parameters. This will help users perform sensitivity analysis, and find an appropriate initial guess for fitting.
- Construct your own fitting algorithm: An advanced mode was created for advanced users to construct their own fitting strategies with a full control of each fitting parameter.
- Batch mode: Users can easily perform quick fitting of multiple fluorescence datasets or XANES datasets.
- Interface with NSLS-II database: A specific I/O interface was designed to obtain data directly from BNL/NSLS-II experimental database.
Currently PyXRF only works for python 2.7. pyxrf is not compatible with
python 3+ because it relies on the [enaml] (http://github.com/nucleic/enaml) library, which is not python 3 compatible as of this writing (11/2015).
First you need to install [conda] (http://continuum.io/downloads). We suggest anaconda because it is a complete installation of the entire scientific python stack, but is ~400 MB. For advanced users, consider downloading [miniconda] (http://conda.pydata.org/miniconda.html) because it is a smaller download (~20 MB).
Then create a conda environment(say pyxrf_test) with python2.7.
$ conda create -n pyxrf_test python=2.7
Then go to the environment named pyxrf_test
$ source activate pyxrf_test
At the same environment, install pyxrf by simply typing
$ conda install pyxrf pyqt=4.11 -c licode
Also install ipython-notebook in the same environment (recommended by not required)
$ conda install ipython-notebook
$ pyxrf
$ conda update -c licode pyxrf
Every time you open a new terminal, make sure to go to pyxrf_test environment first, then launch the software.
$ source activate pyxrf_test
$ pyxrf
To leave this environment, just type
$ source deactivate
The blocking issue for getting pyxrf running on windows is creating a conda
package for [xraylib] (https://github.com/tschoonj/xraylib). If enough users
request this feature, we will make it happen. Until then it is on the back
burner.
Please download a standard h5 file from the link https://drive.google.com/file/d/0B45Mm22EF9TNQzFkSW0xa01mbkE/view This h5 file mainly contains spectrum from 3 detectors, positions of x,y motors and scalers for normalization. Please create h5 file with a similar structure in order to use PyXRF.
For test purposes, a parameter file (https://drive.google.com/file/d/0B45Mm22EF9TNYW11OXozRXVic1E/view) is also provided for users to do fitting for this standard h5 file. However, you should never create parameter file manually. Parameter file can be easily created during the step of automatic peak finding.
Users can transfer spec file to hdf file that pyxrf can take. Please see examples at https://github.com/NSLS-II/PyXRF/blob/master/examples/specfile_to_hdf.ipynb
The latest youtube tutorial of pyxrf version 0.0.4 is at https://www.youtube.com/watch?v=2nFLccehaHI
You may also refer to http://nbviewer.ipython.org/gist/licode/06654b079fd617aaaeca
More documentation will be ready soon!
The core fitting functions are a part of the [scikit-xray] (https://github.com/scikit-beam/scikit-beam) data analysis library for xray data analysis. The design philosophy is to separate fitting and gui, so it is easy to maintain. For more questions, please submit issues through github, or contact Li at [email protected].