Merge branch 'master' of github.com:sandialabs/InterSpec

Author: wcjohns

CMakeLists.txt

@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.1...3.20 FATAL_ERROR)
 
-project(InterSpec VERSION 1.0.11)
+project(InterSpec VERSION 1.0.12)
 
 set(CMAKE_CXX_STANDARD 14)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
@@ -168,7 +168,7 @@ if( CMAKE_COMPILER_IS_GNUCC AND CMAKE_CXX_COMPILER_VERSION VERSION_LESS 5.0 )
   message( FATAL_ERROR, "GCC less than 5 not supported." )
 endif()
 
-# Make a cached variable avaiable to parent cmake projects
+# Make a cached variable available to parent cmake projects
 set( INTERSPEC_VERSION "${PROJECT_VERSION}" CACHE STRING "InterSpec version" )
 
 if(NOT CMAKE_BUILD_TYPE)

InterSpec/RelActCalcAuto.h

@@ -62,40 +62,22 @@ namespace RelActCalc
 
 
 /*
- We need to take in:
- 
- foreground_spectrum;
- background_spectrum; //optional
- 
- struct RoiRanges{ double lower_energy, upper_energy; continuum_type; bool force_full_range; }
- vector<RoiRanges>: roi_ranges;
- 
- struct NucInputInfo{ const SandiaDecay::Nuclide *nuc; double age; bool fit_age; vector<double> gammas_to_release; }
- vector<NucInputInfo> nuclides;
- 
- struct FreeFloatPeak{ double energy; bool release_fwhm; }
- vector<FreeFloatPeak> free_floating_peaks;
- 
- options{ bool fit_energy_cal; RelEffEqnType; size_t RelEffEqnOrder; FwhmFcnForm/Order; }
- 
- 
- For deciding peak ROIs:
- - First use whole specified ROI, and fit for a solution
- - Eliminate statistically insignificant gamma lines; then use remaining gammas to decide peak ROI widths (and if should be further broken up?)
- 
- To get rid of the degeneracy of rel-eff and rel-act:
- - Add an extra term to the Chi2 that forces the rel-act equation to be 1.0 at the lowest energy; maybe multiple difference of current rel-eff equation from 1.0 by the area of the largest peak in any of the ROIs
- 
- Parameters to fit:
- - Energy offset; Energy gain adjust //will be fixed to begin with
- - Fwhm Parameters  // Will be `num_parameters(FwhmForm)` number of these
- - RelEff coefs     // There will be one more than the "order" of relative eff eqn
- - {Activity,Age}   // One pair for each nuclide.  Age will be mostly fixed, but can also be tied to another nuclide of the same elements age
- - Free float peaks {amplitude, width-multiple} // The width multiple is of what the Fwhm Parameters predict (e.g., 1.0 says use parameter value).
- 
- 
  Further things to consider:
- - Sum peaks - not totally sure how to handle - can there just be one parameter that controls strength of peak random summing, and then another for cascade?
+ - For the relative efficiency chart, currently all markers for peaks info fall exactly on the
+    RelEff line. What we should do is, after finding the solution, cluster the peaks together
+    (similar to Act/Shield fit), and then make a template for each clustered peak (sum all peaks into
+    a channel count array, with separate array for each clustered peak), with some threshold (like
+    99% detection threshold) that if the clustered peaks area is less than, it just gets subtracted
+    from the data, unless its the only peak for that ROI.
+    Then fit the templates to data, and use the fit values to make the markers on the RelEff plot.
+    I think the plotting JS can handle showing info for multiple nuclides in a single plots
+    mouse-over.  We will need to create new variable that is analogous to
+    #RelActAutoSolution::m_fit_peaks, but maps multiple source nuclides to a single peak.
+ - Sum peaks - not totally sure how to handle - can there just be one parameter that controls
+    strength of peak random summing, and then another for cascade?  For cascades we dont have
+    gamma-x-ray coincidences (or is it the case we can just assume constant fraction between any
+    gamma and any nuclide?), so we might need two parameters, one for gamams, one for gamma-neutrons
+    and I guess maybe xray-xray?
  */
 
 namespace RelActCalcAuto
@@ -298,7 +280,13 @@ struct Options
   RelActCalc::PuCorrMethod pu242_correlation_method;
 
 
-  /** Under development */
+  /** Peak skew to apply to the entire spectrum.
+   
+   Under development: currently, if total energy range being fit is less than 100 keV, then all peaks will share the same skew.
+   Otherwise a linear energy dependance will be assumed, where the fitting parameters will be for the spectrums lower
+   energy, and the spectrums upper energy; not all skew parameters are allowed to vary with energy; e.g., the Crystal Ball
+   power law is not allowed to have an energy dependence (see `PeakDef::is_energy_dependent(...)`).
+   */
   PeakDef::SkewType skew_type;
 
   static const int sm_xmlSerializationVersion = 0;

external_libs/SpecUtils

@@ -1353,10 +1353,10 @@ bool PeakDef::skew_parameter_range( const SkewType skew_type, const CoefficientT
       if( coef != CoefficientType::SkewPar0 )
         return false;
 
-      starting_value = 0.2;
-      step_size = 0.05;
+      starting_value = 2.0;
+      step_size = 1.0;
       lower_value = 0.0; //Below 0.005 would be numerically bad, but the Bortel function should protect against it.
-      upper_value = 10;
+      upper_value = 15;
 
       break;
     }//case SkewType::Bortel:

src/PeakDef.cpp

@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.15 FATAL_ERROR)
 
-project( InterSpecApp VERSION 1.0.11 )
+project( InterSpecApp VERSION 1.0.12 )
 
 set(BUILD_AS_OSX_APP ON CACHE INTERNAL "Build a OSX native app" )
 set(USE_OSX_NATIVE_MENU ON CACHE INTERNAL "Use macOS native menu-system" )

target/osx/CMakeLists.txt

@@ -100,7 +100,7 @@
                 </menuItem>
             </items>
         </menu>
-        <window title="InterSpec OSX" allowsToolTipsWhenApplicationIsInactive="NO" autorecalculatesKeyViewLoop="NO" releasedWhenClosed="NO" frameAutosaveName="MainWindow" animationBehavior="default" id="371" userLabel="MainWindow">
+        <window title="InterSpec" allowsToolTipsWhenApplicationIsInactive="NO" autorecalculatesKeyViewLoop="NO" releasedWhenClosed="NO" frameAutosaveName="MainWindow" animationBehavior="default" id="371" userLabel="MainWindow">
             <windowStyleMask key="styleMask" titled="YES" closable="YES" miniaturizable="YES" resizable="YES"/>
             <windowPositionMask key="initialPositionMask" leftStrut="YES" rightStrut="YES" topStrut="YES" bottomStrut="YES"/>
             <rect key="contentRect" x="335" y="390" width="1214" height="734"/>