Changeset 35e36fd in sasview for src/sas/qtgui/Perspectives/Fitting

Timestamp:

Sep 8, 2018 9:41:33 AM (6 years ago)

Author:

GitHub <noreply@…>

Branches:

ESS_GUI, ESS_GUI_batch_fitting, ESS_GUI_bumps_abstraction, ESS_GUI_iss1116, ESS_GUI_iss879, ESS_GUI_opencl, ESS_GUI_ordering, ESS_GUI_sync_sascalc

Children:

c0de493

Parents:

c8536d6c (diff), f3e5956 (diff)
Note: this is a merge changeset, the changes displayed below correspond to the merge itself.
Use the (diff) links above to see all the changes relative to each parent.

git-author:

Torin Cooper-Bennun <40573959+tcbennun@…> (09/08/18 09:41:33)

git-committer:

GitHub <noreply@…> (09/08/18 09:41:33)

Message:

Merge pull request #173 from SasView?/ESS_GUI_iss1033

SASVIEW-1033: Support for sasmodels' new 'lazy' intermediate results

Location:

src/sas/qtgui/Perspectives/Fitting

Files:

• FittingLogic.py (modified) (2 diffs)
• FittingWidget.py (modified) (1 diff)
• ModelThread.py (modified) (4 diffs)

src/sas/qtgui/Perspectives/Fitting/FittingLogic.py

@@ -161 +161 @@
         Create a new 1D data instance based on fitting results
         """
-
         return self._create1DPlot(tab_id, return_data['x'], return_data['y'],
                                   return_data['model'], return_data['data'])
@@ -212 +211 @@
         (pq_plot, sq_plot). If either are unavailable, the corresponding plot is None.
         """
-
-        pq_plot = None
-        sq_plot = None
-
-        if return_data.get('pq_values', None) is not None:
-            pq_plot = self._create1DPlot(tab_id, return_data['x'],
-                    return_data['pq_values'], return_data['model'],
-                    return_data['data'], component="P(Q)")
-        if return_data.get('sq_values', None) is not None:
-            sq_plot = self._create1DPlot(tab_id, return_data['x'],
-                    return_data['sq_values'], return_data['model'],
-                    return_data['data'], component="S(Q)")
-
-        return pq_plot, sq_plot
+        plots = []
+        for name, result in return_data['intermediate_results'].items():
+            plots.append(self._create1DPlot(tab_id, return_data['x'], result,
+                         return_data['model'], return_data['data'],
+                         component=name))
+        return plots
 
     def computeDataRange(self):

src/sas/qtgui/Perspectives/Fitting/FittingWidget.py

@@ -2428 +2428 @@
 
         # Create plots for intermediate product data
-        pq_data, sq_data = self.logic.new1DProductPlots(return_data, self.tab_id)
-        if pq_data is not None:
-            pq_data.symbol = "Line"
-            self.createNewIndex(pq_data)
-            # self.communicate.plotUpdateSignal.emit([pq_data])
-            new_plots.append(pq_data)
-        if sq_data is not None:
-            sq_data.symbol = "Line"
-            self.createNewIndex(sq_data)
-            # self.communicate.plotUpdateSignal.emit([sq_data])
-            new_plots.append(sq_data)
+        plots = self.logic.new1DProductPlots(return_data, self.tab_id)
+        for plot in plots:
+            plot.symbol = "Line"
+            self.createNewIndex(plot)
+            new_plots.append(plot)
 
         for plot in new_plots:

src/sas/qtgui/Perspectives/Fitting/ModelThread.py

@@ -164 +164 @@
         index = (self.qmin <= self.data.x) & (self.data.x <= self.qmax)
 
+        intermediate_results = None
+
         # If we use a smearer, also return the unsmeared model
         unsmeared_output = None
@@ -174 +176 @@
             mask = self.data.x[first_bin:last_bin+1]
             unsmeared_output = numpy.zeros((len(self.data.x)))
-            unsmeared_output[first_bin:last_bin+1] = self.model.evalDistribution(mask)
+
+            return_data = self.model.calculate_Iq(mask)
+            if isinstance(return_data, tuple):
+                # see sasmodels beta_approx: SasviewModel.calculate_Iq
+                # TODO: implement intermediate results in smearers
+                return_data, _ = return_data
+            unsmeared_output[first_bin:last_bin+1] = return_data
             output = self.smearer(unsmeared_output, first_bin, last_bin)
 
@@ -193 +201 @@
                 unsmeared_error=unsmeared_error
         else:
-            output[index] = self.model.evalDistribution(self.data.x[index])
-
-        sq_values = None
-        pq_values = None
-        s_model = None
-        p_model = None
-        if isinstance(self.model, MultiplicationModel):
-            s_model = self.model.s_model
-            p_model = self.model.p_model
-        elif hasattr(self.model, "calc_composition_models"):
-            results = self.model.calc_composition_models(self.data.x[index])
-            if results is not None:
-                pq_values, sq_values = results
-
-        if pq_values is None or sq_values is None:
-            if p_model is not None and s_model is not None:
-                sq_values = numpy.zeros((len(self.data.x)))
-                pq_values = numpy.zeros((len(self.data.x)))
-                sq_values[index] = s_model.evalDistribution(self.data.x[index])
-                pq_values[index] = p_model.evalDistribution(self.data.x[index])
+            return_data = self.model.calculate_Iq(self.data.x[index])
+            if isinstance(return_data, tuple):
+                # see sasmodels beta_approx: SasviewModel.calculate_Iq
+                return_data, intermediate_results = return_data
+            output[index] = return_data
+
+        if intermediate_results:
+            # the model returns a callable which is then used to retrieve the data
+            intermediate_results = intermediate_results()
+        else:
+            # TODO: this conditional branch needs refactoring
+            sq_values = None
+            pq_values = None
+            s_model = None
+            p_model = None
+
+            if isinstance(self.model, MultiplicationModel):
+                s_model = self.model.s_model
+                p_model = self.model.p_model
+
+            elif hasattr(self.model, "calc_composition_models"):
+                results = self.model.calc_composition_models(self.data.x[index])
+                if results is not None:
+                    pq_values, sq_values = results
+
+            if pq_values is None or sq_values is None:
+                if p_model is not None and s_model is not None:
+                    sq_values = numpy.zeros((len(self.data.x)))
+                    pq_values = numpy.zeros((len(self.data.x)))
+                    sq_values[index] = s_model.evalDistribution(self.data.x[index])
+                    pq_values[index] = p_model.evalDistribution(self.data.x[index])
+
+            if pq_values is not None and sq_values is not None:
+                intermediate_results  = {
+                    "P(Q)": pq_values,
+                    "S(Q)": sq_values
+                }
+            else:
+                intermediate_results = {}
 
         elapsed = time.time() - self.starttime
@@ -223 +250 @@
             source = self.source, unsmeared_output = unsmeared_output,
             unsmeared_data = unsmeared_data, unsmeared_error = unsmeared_error,
-            pq_values = pq_values, sq_values = sq_values)
+            intermediate_results = intermediate_results)
 
         if LocalConfig.USING_TWISTED: