neutronimaging/tofimaging

Tool package for neutron time of flight analysis

tofimaging's Issues

TOF - adapt BB module

adapt the BB module to handle TOF imaging dataset.

• OB normalization should be done one-by-one with projection
• handle the absence of dark current (not existing for the MCP)

Implement calibration module

Implement routine in C++ for calibrating the flight path on reference sample.

Restructuring modules and cleanup

Towards an alpha release the following items:

• code cleanup
• restructuring / renaming of the modules for more consistency
• Pdf with guides to the utilization of functions.

Implement running average

That is very often applied to process the data before fitting, the running average is done along the wavelength axis.

• on the data
• or on the signal?

Add list of requirements

We need to have a requirements.txt, and a requirements_dev.txt that will list the python packages needed to be able to use that library.

Normalization should be optional in the UI

User should have the option to normalize or not the data. If flag is off, the data loaded will be used directly. If flag is on, user will have to load the sample and the ob data.

Split coherent elastic cross section term in texture fitting model.

Implement binning routines in the wavelength axis

ToF to wavelength spectrum calibration with refrence using 2 edges

It is often required to calibrate ToF spectra from tof to wavelength, when the instrumental parameters are not known but the material is. Add a simple function that asks which edges to use for calibration and automatically match experimental and reference edges to perform calibration.

Edge fitting function - parameter constraints

Add constraints to parameters: t0, alpha and gamma should be non-negative

• Python
• C++

Implement the model and fitting function in C++.

Bring the model and fitting function in C++

• implement different options for edge shape (transmission, attenuation and options for functions before and after the edge)
• implement the gaussian of the gradient option, this is now issue #9
• implement L-M fitting for all cases and tests the result
• define strategies for parameter intialization

Jorgensen-type profile function

Implement edge fitting as in Sato 2017.
With the three-step fitting

Implement also a simplified version for finding the edge position

Instead of using the edge fitting,

• provide also a simplified version using gaussian and derivative, to get an estimated of the edge position
• Add pre smoothing before fitting.

TOF - module for computing ratio before and after the edge

Implement a module that compute the ratio before and after the edge in a given dataset, without the previous need of edge fitting.
It should work for

• 2D dataset
• 3D dataset

Build scripts for TOFImaging

Create build scripts:

• TOF_ImagingAlgorithm
• Dependencies
• Tests
• Add builds in imaging suite build_all (not supported by build server yet @anderskaestner )

TOF scripts - optimize edgefitting_2D

Optimize performance of edgefitting_2D.py and eventually change the way to call it, it is very slow from the notebook

Kernel size should be customized in ui

The kernel selection should have the following options:

• Selection using a combo box
• Box (Moving average), parameters box size in different directions or uniform
• Gauss, parameters sigma in different directions or uniform
• Custom-designed, the user provides kernel... is it needed?

ToFImaging scripts refinements

Following refinements are required for scripts:

• edgefitting_2D: change calibration parameters into dictionary, check outputs, the width is not assigned (check line 135), change est_parameters into dictionary and rename them initial_guess, add a check for saving data
• TOFCalibration, check for smarter solution for ROI extraction

FOBI algorithms

Implement a module for FOBI analysis.
Must include:

• chopper time delay models
• spectrum overlap merging
• Wiener decorrelation

Optional:

• full data reduction routine (from tifs)
• moving average options
• spectrum calibration (non manual)

ToFImaging - edge fitting in 2D

Write the routine to perform edge fitting in all chosen pixels in an image.

• ROI selection of the area to be fitted
• fitting parameters to be saved
• images with edge position
• evaluate performance and speed

Bragg Edge fitting - fit in attenuation

Update the code and make it flexible in order to switch smoothly the fitting routine from attenuation to transmission and back.
The Kropff model has to be updated accordingly

TOF - Bin in the TOF/wavelength dimension

Implement advanced binning given the d lambda / lambda of the different instruments

prepare project to library package convention

This steps will make the transition to a pip module easier.

Work will be done in branch jean_python_packaging

Import lmfit in the repository

Add the source code in the external repo

March-Dollase texture fitting

Implement module for fitting March-Dollase-type texture. Example:

Implement python bindings to ToFLib

• Set up CMake file
• Implement binding code in separate lib
• Setup python unit test

Add unit test to reduction_tools

TOF scripts rename AdvancedBraggEdgeFitting_v2

Rename AdvancedBraggEdgeFitting_v2 to AdvancedBraggEdgeFitting, and update all calls from different notebooks

Create a python binding for NXS lib

• Implement a C++ version if the NXS library
• Implement python bindings for NXS lib using PyBind11
• Implement CMake file to compile the lib
• Test on different python interpreters
  • MacOS
  • Linux Ubuntu
  • Windows
• Create PIP installer

Documentation/wiki for the ToF Imaging project

Comment the code and write wiki/documentation on what is done and what will have to be done

Reference spectra database

Untill the NCrystal can be integrated in the software package for accurate modeling of reference Bragg pattern, it would be useful to add reference spectra of commonly used elements and alloys with specific crystal structures.

Import and test lmfit in all OS

• Linux
• Mac, possibly already working
• Windows

Provide unit test and use new Filter algorithm instead of moving average

Module that goes from .fits to ToF transmission data

Adding routine that loads data from .fits folders, output from MPC and tremsin correction, and normalize into transmission after handling infs, nans, and optionally applying moving average.

Segmentation tools for regions to fit and texture mapping.

• Prior to the 2D Bragg edge fitting routines it would be useful to detect the sample region to avoid undesired fitting.

• Based on the spectrum correlation, detect sample regions with a similar spectrum hence, texture.

Peak profile module

Implement module able to reproduce peak profiles such as:

Rename and reformat

Update names to conform with style guide.

Edgefitting function

Convert from C array to std vectors

TOFImaging - add armadillo in Win

Update the pro file including armadillo lib and dependencies for Win10

Kropff model - remove exponential decay

Add the option to remove the exponential decay from the Kropff model, because for ODIN (long pulse) this should not be present.

TOF - normalization module

Add a usual normalization module: divinding by the OB and dose correcting for TOF imaging dataset.
This could be a special case of the BB module for TOF datasets (withouth BB correction)

Move clean up code into the TOFImaging repository.

Clean up the code of ToF_notebooks and bring the cleaned files with history into the TOFImaging respository

• TOF calibration from IMATexperiment branch in scripts
• TOFBinImages from IMATexperiment branch in scripts
• CalculateSummedImgs from IMATexperiment branch in scripts
• ShutterFileGen_IMAT from IMATexperiment branch in scripts
• RunMuhrec from IMAT_TOF_tomography in scripts
• WriteXMLConfigFiles from IMAT_TOF_tomography in scripts
• LSP_data_analysis from LSP_edgefitting branch in scritps, to be renamed in something like 2D_edgefitting

Normalization mode

Flag should be present to allow 2 normalization:

• normalize pixel by pixel
• normalize by ROI (mean of ROI counts)

TOF - multi-edge fitting

Implement a simplified version of multi-edge fitting. Parameters to be fitted: lattice parameter

TOF - iterative fitting

Implement schemes for iterative fitting

MD docs.

Add docs inspired by https://github.com/neutronimaging/ImagingQuality/tree/master/docs/

parsing more than one spectrum in a directory

Dear developer(s),

it is often the case that in one directory we have more than one ToF radiography (for example to split a long exposure time). These files are normally named with a _00x at the end of their given name (e.g. test_000_0nnnn.fits and test_001_0nnnn,fits). It would be great if with your program one would be able to correct each of them with its own spectrum file.
Thanks a lot and keep up with the great work

M

Try total variation as spectrum ID

Total variation seems to be a useful scalar metric to compare different pixels and to identify regions with different spectra. I tried with different smoothing and noise levels and it looks promising.

Eq id=log(abs(sum(s(lambda))/N))

TOF ImagingAlgorithms UnitTests in Win and Mac

UnitTests failing in Mac and Win

• in Mac they fail to run because nexlib is not found by kipl
• in Windows fail to compile due to unresolved externals

TOF - Bragg Edge fitting, moderator decay parameter

• Provide robust routine to fit the moderator decay at the calibration step, this parameters only depends on the instrument and should be not be fitted during the sample analysis

• Provide moderator decay formula for different instruments, as described in the instruments paper, see https://doi.org/10.1107/S1600576719001730 for IMAT