Tools library for inelastic neutron spectroscopy array analyzer MultiFLEXX.
Type %matplotlib qt in console bevore importing multiflexxlib.
The code is being worked on to keep up with changes in its dependencies. Please open an issue or drop me an email if something is off.
multiflexxlib is a Python package for visualization and treatment of neuron spectroscopy data acquired with cold neutron array detector MultiFLEXX. A detailed description on the detector can be found on the HZB website.
multiflexxlib requires python3 version > 3.5. For installation of Python environment under Windows it is recommended to install a scientific Python package such as Anaconda to make your life easier.
Support for MacOS is questionable at the moment due to peculiarities on how tkinter interacts with GUI routines of MacOS. The code should work though - and please report if it doesn't.
Alternately, Python2 > 2.7 is partially supported. Most of the code is written with Python 2 compatibility in mind, but might contain (additional) bugs.
>1GB of free memory recommended.
If you are unsure about how to do something or are getting unexpected behaviour, open an issue or drop me an email!
run command pip install multiflexxlib from windows or linux command console.
Data files for measurement of excitations in antiferromagnet MnF2 can be found here. Please download the files into a folder.
Download run.py. Double-click on saved file and select the folder containing your MultiFLEXX scan files when asked for data folder. All possible 2D const-E plots will be shown. Double-click on a plot to open the plot in its own window.
Please also refer to docstrings of classes and functions, accessible through help() function, e.g. help(mfl.UBMatrix).
It is possible and recommended to use this package in an interactive Python interpreter such as IPython or a Matlab-esque Python development environment like Spyder, which both come with a default Anaconda installation.
import multiflexxlib as mfl
This is required before you can start using the package. mfl is a shorthand that you can also choose for yourself.
alldata = mfl.read_and_bin()
Alternately, use
alldata = mfl.read_and_bin(processes=4)to load data with multiple CPU cores simultaneously. This does not always work depending on the platform and python runtime in use.
UB-matrix and plotting axes will be determined automatically if not provided. x and y axes are scaled to be equal in terms of absolute reciprocal length. Non-orthogonal axes is supported. If lattice parameter or plotting axes is to be overridden from scan files metadata, create a UBMatrix object as follows:
my_ubmatrix = mfl.UBMatrix([a, b, c, alpha, beta, gamma], hkl1, hkl2, plot_x, plot_y)hkl1, hkl2, plot_x, plot_yshould be given as 3-element list [h, k, l].plot_xandplot_yparameters can be omitted ifplot_x == hkl1andplot_y == hkl2. Pass this customUBMatrixintoread_and_binasmfl.read_and_bin(ub_matrix=my_ubmatrix)
You will be asked for a folder containing data like in minimal usage. All data from the folder will be loaded, scans done under identical conditions will be binned together.
print(alldata)
Prints a tabular summary as follows.
| ei | en | tt | mag | locus_a | locus_p | points | |
|---|---|---|---|---|---|---|---|
| 0 | 8.000747 | 3.500747 | 294.8735 | 0 | 1p 180v | 1p 180v | 3782 pts |
| 1 | 8.000747 | 4.000747 | 294.8735 | 0 | 1p 180v | 1p 180v | 3538 pts |
| 2 | 8.000747 | 4.500747 | 294.8735 | 0 | 1p 180v | 1p 180v | 3538 pts |
| 3 | 8.000747 | 5.000747 | 294.8735 | 0 | 1p 180v | 1p 180v | 3660 pts |
| 4 | 8.000747 | 5.500747 | 294.8735 | 0 | 1p 180v | 1p 180v | 3782 pts |
The binning process considers any two values that are different by less than tolerance threshold to be identical. Defaults are energy: 0.05meV, temperature: 1K, angles 0.2° and magnetic field 0.05T. Partially and fully repeating and overlapping scans will be dealt with in a sensible manner. A3-A4 scans with very small A4 step size will be binned correctly.
p = alldata.plot(subset=[0,1,2,3,4]) Plots data with index 0 ~ 4.
Add, remove or replace numbers in
[0,1,2,3,4]to select data entries you want. Thesubsetparameter can be omitted if you want to plot all possible plots (like in this case). APlot2Dobject is returned to name p. A graph will be generated. Plots can be panned and zoomed using controls in graph window.
The returned Plot2D object can be used to access graph customization methods and 1D-cuts.
Accidental Bragg scattering spurions can have high intensity that drown out inelastic signals. To Crop out such intensities:
p.auto_lim()
This attempts to detect if there is a couple of pixels that have way higher counts than the rest, and cut these out.
p.set_plim(pmin=0, pmax=99.7)
This limits color map scale to 99.7th percentile of pixels, leaving out the highest 0.3%.
It might be interesting extract an 1-D subset (1D cut), which is done as follows:
c = p.cut_voronoi([1, 1, 1], [1.1, 1.1, 1], subset=[0, 2])
This does a cut through Voronoi partition of const-E data. Regions crossed by cut line are picked up and included in the cut. subset parameter can be omitted. [1, 1, 1] here is [h, k, l] values. subset=[0, 2] instead of [1, 3] because python index starts with 0. This generates a cut from [1, 1, 1] to [1.1, 1.1, 1]. The cut method draws a line segment between specified cut start and end points, and each data point corresponding to crossed regions is subsequently projected onto cut axis.
It makes more sense to do a traditional cut with rectangular bins if amount of data is sufficiently high:
c = p.cut_bins([1, 1, 1], [1.1, 1.1, 1], subset=[0, 2], no_points=21, ytol=[0, 0, 0.02])
To plot a Q-E dispersion plot:
df.dispersion([1, 1, 1], [1.1, 1.1, 1], no_points=21)
This generates a vertical stacking of 1D const-E cuts. It should be noted that data from all final energies is used, thus intensity might zigzag from resolution effects.
To export all data into CSV files:
df.to_csv()
This writes a series of CSV files into folder containing source scan files.