Changeset 2f1a0dc in sasview for sansmodels

Timestamp:

May 27, 2010 11:16:29 AM (15 years ago)

Author:

Jae Cho <jhjcho@…>

Branches:

master, ESS_GUI, ESS_GUI_Docs, ESS_GUI_batch_fitting, ESS_GUI_bumps_abstraction, ESS_GUI_iss1116, ESS_GUI_iss879, ESS_GUI_iss959, ESS_GUI_opencl, ESS_GUI_ordering, ESS_GUI_sync_sascalc, costrafo411, magnetic_scatt, release-4.1.1, release-4.1.2, release-4.2.2, release_4.0.1, ticket-1009, ticket-1094-headless, ticket-1242-2d-resolution, ticket-1243, ticket-1249, ticket885, unittest-saveload

Children:

beb374a

Parents:

06c7fcc

Message:

2d input for eval is now qr instead of list (qx,qr), this way we can use numpy.vertorize and it is much faster; So we don't support anisotropic 2D for python models.

File:

• sansmodels/src/sans/models/BaseComponent.py (modified) (3 diffs)

sansmodels/src/sans/models/BaseComponent.py

@@ -75 +75 @@
             qy_prime = [ qy[0], qy[1], qy[2], ....] 
             
+            Then get
+            q = numpy.sqrt(qx_prime^2+qy_prime^2)
+            
+            that is a qr in 1D array;
+            q = [q[0], q[1], q[2], ....] 
+            Note: Due to 2D speed issue, no anisotropic scattering is supported for python models, 
+            thus C-models should have their own evalDistribution methods.
+            
             The method is then called the following way:
             
-            evalDistribution([qx_prime, qy_prime])
+            evalDistribution(q)
+            where q is a numpy array.
             
             @param qdist: ndarray of scalar q-values or list [qx,qy] where qx,qy are 1D ndarrays 
@@ -92 +101 @@
             qy = qdist[1]
             
-            # Create output array
-            iq_array = numpy.zeros((len(qx)))
-            
-            for i in range(len(qx)):
-                    iq_array[i] = self.runXY([qx[i],qy[i]])
+            # calculate q_r component for 2D isotropic
+            q = numpy.sqrt(qx**2+qy**2)
+            # vectorize the model function runXY
+            v_model = numpy.vectorize(self.runXY,otypes=[float])
+            # calculate the scattering
+            iq_array = v_model(q)
+
             return iq_array
                 
         elif qdist.__class__.__name__ == 'ndarray':
                 # We have a simple 1D distribution of q-values
-                iq_array = numpy.zeros(len(qdist))
-                for i in range(len(qdist)):
-                    iq_array[i] = self.runXY(qdist[i])
+                v_model = numpy.vectorize(self.runXY,otypes=[float])
+                iq_array = v_model(qdist)
+
                 return iq_array
             
@@ -110 +121 @@
             mesg += " or a list [qx,qy] where qx,qy are 2D ndarrays."
             raise RuntimeError, mesg
+        
+    
     
     def clone(self):