Changeset 635 for trunk

Timestamp:

May 26, 2012 12:56:38 PM (13 years ago)

Author:

vondreele

Message:

HKLF refinement "numeric" now OK; revamp output a bit for more generalized R-values.

Location:

trunk

Files:

• GSASII.py (modified) (2 diffs)
• GSASIIgrid.py (modified) (1 diff)
• GSASIIstruct.py (modified) (12 diffs)

trunk/GSASII.py ¶

@@ -2008 +2008 @@
             return
         #works - but it'd be better if it could restore plots
-        dlg = wx.ProgressDialog('Residual','Powder profile Rwp =',101.0, 
+        dlg = wx.ProgressDialog('Residual','All data Rw =',101.0, 
             style = wx.PD_ELAPSED_TIME|wx.PD_AUTO_HIDE|wx.PD_CAN_ABORT)
         screenSize = wx.ClientDisplayRect()
@@ -2015 +2015 @@
         dlg.SetPosition(wx.Point(screenSize[2]-Size[0]-305,screenSize[1]+5))
         dlg.SetSize(Size)
-        Rwp = 100.00
+        Rw = 100.00
         try:
-            Rwp = G2str.Refine(self.GSASprojectfile,dlg)
+            Rw = G2str.Refine(self.GSASprojectfile,dlg)
         finally:
             dlg.Destroy()        
-        dlg = wx.MessageDialog(self,'Load new result?','Refinement results, Rwp =%.3f'%(Rwp),wx.OK|wx.CANCEL)
+        dlg = wx.MessageDialog(self,'Load new result?','Refinement results, Rw =%.3f'%(Rw),wx.OK|wx.CANCEL)
         try:
             if dlg.ShowModal() == wx.ID_OK:

trunk/GSASIIgrid.py ¶

@@ -863 +863 @@
                 value = 1.0
             data['minF/sig'] = value
-            Factr.SetValue('%.2f'%(value))
+            minSig.SetValue('%.2f'%(value))
 
         LSSizer = wx.FlexGridSizer(cols=4,vgap=5,hgap=5)

trunk/GSASIIstruct.py ¶

@@ -1465 +1465 @@
             elif 'HKLF' in histogram:
                 pfx = str(pId)+':'+str(hId)+':'
+                print ' Final refinement RF, RF^2 = %.2f%%, %.2f%% on %d reflections'   \
+                    %(Histogram[pfx+'Rf'],Histogram[pfx+'Rf^2'],Histogram[pfx+'Nref'])
                 ScalExtSig = {}
                 for item in ['Scale','Ep','Eg','Es']:
@@ -1818 +1820 @@
             print '\n Histogram: ',histogram,' histogram Id: ',hId
             print 135*'-'
-            print ' Final refinement wRp = %.2f%% on %d observations in this histogram'%(Histogram['wRp'],Histogram['Nobs'])
+            print ' Final refinement wR = %.2f%% on %d observations in this histogram'%(Histogram['wR'],Histogram['Nobs'])
             if Print:
                 print ' Instrument type: ',Sample['Type']
@@ -2883 +2885 @@
                 varylist,Histogram,Phases,calcControls,pawleyLookup)
             if dlg:
-                dlg.Update(Histogram['wRp'],newmsg='Hessian for histogram %d Rwp=%8.3f%s'%(hId,Histogram['wRp'],'%'))[0]
+                dlg.Update(Histogram['wR'],newmsg='Hessian for histogram %d Rw=%8.3f%s'%(hId,Histogram['wR'],'%'))[0]
             if len(Hess):
                 Vec += np.sum(dMdvh*np.sqrt(w[xB:xF])*dy[xB:xF],axis=1)
@@ -2904 +2906 @@
             Histogram['Nobs'] = len(refList)
             dFdvDict = StructureFactorDerv(refList,G,hfx,pfx,SGData,calcControls,parmdict)
-            df = np.empty(len(refList))
+            df = np.zeros(len(refList))
             for i,ref in enumerate(refList):
                 if ref[6] > 0:
@@ -2911 +2913 @@
                     ref[8] = ref[5]/parmdict[phfx+'Scale']
                     df[i] = (ref[5]-ref[7])/ref[6]
-                    sumwYo += ref[5]/ref[6]**2
+                    sumwYo += (ref[5]/ref[6])**2
             M = np.concatenate((M,df))
 
@@ -2925 +2927 @@
     Histograms,Phases = HistoPhases
     M = np.empty(0)
-    sumwYo = 0
+    SumwYo = 0
     Nobs = 0
     histoList = Histograms.keys()
@@ -2944 +2946 @@
             Nobs += Histogram['Nobs']
             Histogram['sumwYo'] = np.sum(w[xB:xF]*y[xB:xF]**2)
-            sumwYo += Histogram['sumwYo']
+            SumwYo += Histogram['sumwYo']
             yc[xB:xF],yb[xB:xF] = getPowderProfile(parmdict,x[xB:xF],
                 varylist,Histogram,Phases,calcControls,pawleyLookup)
@@ -2951 +2953 @@
             Histogram['sumwYd'] = np.sum(np.sqrt(w[xB:xF])*(yd[xB:xF]))
             wdy = -np.sqrt(w[xB:xF])*(yd[xB:xF])
-            Histogram['wRp'] = min(100.,np.sqrt(np.sum(wdy**2)/Histogram['sumwYo'])*100.)
+            Histogram['wR'] = min(100.,np.sqrt(np.sum(wdy**2)/Histogram['sumwYo'])*100.)
             if dlg:
-                dlg.Update(Histogram['wRp'],newmsg='For histogram %d Rwp=%8.3f%s'%(hId,Histogram['wRp'],'%'))[0]
+                dlg.Update(Histogram['wR'],newmsg='For histogram %d Rw=%8.3f%s'%(hId,Histogram['wR'],'%'))[0]
             M = np.concatenate((M,wdy))
         elif 'HKLF' in histogram[:4]:
@@ -2969 +2971 @@
             Histogram['Nobs'] = len(refList)
             refList = StructureFactor(refList,G,hfx,pfx,SGData,calcControls,parmdict)
-            df = np.empty(len(refList))
+            df = np.zeros(len(refList))
+            sumwYo = 0
+            sumFo = 0
+            sumFo2 = 0
+            sumdF = 0
+            sumdF2 = 0
             for i,ref in enumerate(refList):
                 if ref[6] > 0 and ref[5] > 0:
@@ -2976 +2983 @@
                     if calcControls['F**2']:
                         if ref[5]/ref[6] >= calcControls['minF/sig']:
+                            sumFo2 += ref[5]
+                            Fo = np.sqrt(ref[5])
+                            sumFo += Fo
+                            sumFo2 += ref[5]
+                            sumdF += abs(Fo-np.sqrt(ref[7]))
+                            sumdF2 += abs(ref[5]-ref[7])
                             Nobs += 1
                             df[i] = -(ref[5]-ref[7])/ref[6]
-                            sumwYo += ref[5]/ref[6]**2
+                            sumwYo += (ref[5]/ref[6])**2
                     else:
                         Fo = np.sqrt(ref[5])
@@ -2984 +2997 @@
                         sig = ref[6]/(2.0*Fo)
                         if Fo/sig >= calcControls['minF/sig']:
+                            sumFo += Fo
+                            sumFo2 += ref[5]
+                            sumdF += abs(Fo-Fc)
+                            sumdF2 += abs(ref[5]-ref[7])
                             Nobs += 1
                             df[i] = (Fo-Fc)/sig
-                            sumwYo += Fo/sig**2
+                            sumwYo += (Fo/sig)**2
+            Histogram['sumwYo'] = sumwYo
+            SumwYo += sumwYo
+            Histogram['wR'] = min(100.,np.sqrt(np.sum(df**2)/Histogram['sumwYo'])*100.)
+            Histogram[phfx+'Rf'] = 100.*sumdF/sumFo
+            Histogram[phfx+'Rf^2'] = 100.*sumdF2/sumFo2
+            Histogram[phfx+'Nref'] = Nobs
+            if dlg:
+                dlg.Update(Histogram['wR'],newmsg='For histogram %d Rw=%8.3f%s'%(hId,Histogram['wR'],'%'))[0]
             M = np.concatenate((M,df))
-
-
-    Histograms['sumwYo'] = sumwYo
+    Histograms['sumwYo'] = SumwYo
     Histograms['Nobs'] = Nobs
-    Rwp = min(100.,np.sqrt(np.sum(M**2)/sumwYo)*100.)
-    print 'Rwp',sumwYo,np.sum(M**2),np.sqrt(np.sum(M**2)),Rwp,Nobs
+    Rw = min(100.,np.sqrt(np.sum(M**2)/SumwYo)*100.)
     if dlg:
-        GoOn = dlg.Update(Rwp,newmsg='%s%8.3f%s'%('Powder profile Rwp =',Rwp,'%'))[0]
+        GoOn = dlg.Update(Rw,newmsg='%s%8.3f%s'%('All data Rw =',Rw,'%'))[0]
         if not GoOn:
             parmDict['saved values'] = values
@@ -3093 +3115 @@
         print ' Number of function calls:',result[2]['nfev'],' Number of observations: ',Histograms['Nobs'],' Number of parameters: ',len(varyList)
         print ' Refinement time = %8.3fs, %8.3fs/cycle, for %d cycles'%(runtime,runtime/ncyc,ncyc)
-        print ' wRp = %7.2f%%, chi**2 = %12.6g, reduced chi**2 = %6.2f'%(Rvals['Rwp'],Rvals['chisq'],Rvals['GOF'])
+        print ' wR = %7.2f%%, chi**2 = %12.6g, reduced chi**2 = %6.2f'%(Rvals['Rwp'],Rvals['chisq'],Rvals['GOF'])
         try:
             covMatrix = result[1]*Rvals['GOF']