DevNotes_Projects_BetaApproximation

Beta Approximation Work Package

Incorporation of the 'Beta Approximation' or 'Beta Correction' for S(Q) has long been identified as an important, yet often ignored, correction in model fitting. However, implementing it in SasView is non-trivial.

As of Summer 2018 we are starting to lay the groundwork, with the help of two vacation students, one in the US and one in the UK.

They will approach the implementation from both the bottom-up (changing the code kernel to compute amplitudes instead of intensities, etc) and the top-down (the UI, model parameterisation, choice of effective radius, etc).

As part of the implementation process, computations from SasView are also being validated against those from SASfit, FISH and Matlab.

Tasks

• Understand exactly what the Beta Approximation calculation requires
• Figure out how the SasView 'middle layer' needs to change to implement the Beta Approximation calculation
• Decide how best to redesign the UI to allow a User to make use of the Beta Approximation calculation
• Decide how best to implement the Effective Radius (ER) of interaction in S(Q) calculations (may involve an overhaul of constraints in SasView)
• Provide unit tests
• Provide documentation

Guiding References

• Analysis of small angle neutron scattering spectra from polydisperse interacting colloids. M Kotlarchyk & S‐H Chen. Journal of Chemical Physics 79, 2461 (1983) ​https://doi.org/10.1063/1.446055
• A critical examination of the decoupling approximation for small-angle scattering from hard ellipsoids of revolution. DG Greene, DV Ferraro, AM Lenhoff & NJ Wagner. Journal of Applied Crystallography 49, 1734 (2016) ​https://doi.org/10.1107/S1600576716012929
• SANS from concentrated dispersion. S Kline. NCNR Summer School. Neutron Small Angle Scattering and Reflectometry from Submicron Structures . June 5 - 9, (2000) ​https://www.ncnr.nist.gov/programs/sans/pdf/sans_conc_disp.pdf

Contributors

• Paul Kienzle & Greg Suczewski (student)
• Richard Heenan & Torin Cooper-Bennun (student)
• Yun Liu

• Paul Butler
• Wojtek Potrzebowski
• William Heller
• Steve King

Tickets

Milestone: SasView 5.0.0 (7 matches)

Ticket	Resolution	Summary	Owner	Reporter
#1149		qt5 - Separate composition of mixture models in GUI	tcbennun	tcbennun
#1136	fixed	QT5 GUI - rename FitPage0 to Fitpage1 or rename M1 to M0	richardh	richardh
#1135		QT GUI - complex constraints (master ticket)	tcbennun	tcbennun
#1130	fixed	QT GUI - cannot fit any S(Q) params if using P(Q)S(Q) model		tcbennun
#1119	duplicate	external calls to sasmodels needing beta(Q)	richardh	richardh
#1116		qt5 user level, changes to model tab	tcbennun	richardh
#1115	wontfix	qt5 user level S(Q) tab		richardh
Milestone: SasView Next Release +1 (14 matches)
Ticket	Resolution	Summary	Owner	Reporter
#1153		Separate composition of mixture models in GUI		tcbennun
#1126		Passing additional computed, non-fitting, parameter values back from models	richardh	richardh
#1122		General documentation for beta(Q) modification of S(Q)	richardh	richardh
#1117	fixed	qt5 user level Data window		richardh
#1114	fixed	qt5 gui - model parameter table not displaying well in Windows	tcbennun	richardh
#1113	wontfix	qt5 gui - constraints not working?	richardh	richardh
#1065		Check the parameter names for complex mixture models		pkienzle
#1063		Need unit tests for combined models P(Q)S(Q) etc	awashington	awashington
#1001		allow mixtures with some components being P*S		pkienzle
#780		let the user set effective radius in P(Q)*S(Q)		pkienzle
#746	wontfix	Consider enhanced fitting constraints		richardh
#587		Refactor sum/product model		pkienzle
#487		Constraints should be relational operator not assignment operators		butler
#475		Add volume fraction explicitly in all relevant models		smk78
Milestone: sasmodels 1.0 (5 matches)
Ticket	Resolution	Summary	Owner	Reporter
#1202	fixed	beta approx branch doesn't support ER/VR tests	GitHub <[email protected]>	pkienzle
#1201	fixed	default effective radius type	GitHub <[email protected]>	pkienzle
#1121		validation of beta(Q) and other calculations against external code		richardh
#1120		beta(Q) approx models to edit	richardh	richardh
#491		adding support for F(Q) in sasmodels	wojciech	butler
Milestone: sasmodels Next Release +1 (2 matches)
Ticket	Resolution	Summary	Owner	Reporter
#1118		middle level - sort out P(Q)S(Q) including beta(Q)	richardh	richardh
#1096		improve structure factor calculations with beta approximation		pkienzle