Merge pull request #791 from mkstoyanov/update_diff

Updated the documentation for the new python method and added some testing.

Author: mkstoyanov

InterfacePython/TasmanianSG.py

@@ -110,6 +110,7 @@
 pLibTSG.tsgGetPointsStatic.argtypes = [c_void_p, POINTER(c_double)]
 pLibTSG.tsgGetQuadratureWeightsStatic.argtypes = [c_void_p, POINTER(c_double)]
 pLibTSG.tsgGetInterpolationWeightsStatic.argtypes = [c_void_p, POINTER(c_double), POINTER(c_double)]
+pLibTSG.tsgGetDifferentiationWeightsStatic.argtypes = [c_void_p, POINTER(c_double), POINTER(c_double)]
 pLibTSG.tsgLoadNeededValues.argtypes = [c_void_p, POINTER(c_double)]
 pLibTSG.tsgGetLoadedValuesStatic.argtypes = [c_void_p, POINTER(c_double)]
 pLibTSG.tsgEvaluate.argtypes = [c_void_p, POINTER(c_double), POINTER(c_double)]
@@ -1091,11 +1092,11 @@ def getDifferentiationWeights(self, lfX):
         lfX: a 1-D numpy.ndarray with length iDimensions
              the entries indicate the points for evaluating the weights
 
-        output: a 1-D numpy.ndarray of length getNumPoints()
+        output: a 2-D numpy.ndarray of size (getNumPoints(), iDimensions)
             the order of the weights matches the order in getPoints()
 
-        TODO: This, and the C++ call, would probably benefit from using a
-        sparse matrix data structure.
+        Similar to the other Weights methods, those are generally slower than
+        the internal alternatives, e.g., differentiate() or evaluate().
         '''
         iNumX = len(lfX)
         if (iNumX != self.getNumDimensions()):

InterfacePython/testExceptions.py

@@ -209,6 +209,8 @@ def getSparseGridTests(self):
                    ["grid1 = Tasmanian.SparseGrid(); grid1.setGPUID(grid1.getNumGPUs());", "iGPUID"],
                    ["grid.makeLocalPolynomialGrid(1, 1, 1, 1, 'localp'); Tasmanian.loadNeededPoints(lambda x, tid : x, grid, 1);", "notError"],
                    ["grid.makeLocalPolynomialGrid(1, 1, 1, 1, 'localp'); Tasmanian.loadNeededPoints(lambda x, tid : np.ones((2,)) * x, grid, 1);", "loadNeededValues"],
+                   ["grid.makeLocalPolynomialGrid(1, 1, 1, 1, 'localp'); grid.getDifferentiationWeights((1, 2));", "lfX"],
+                   ["grid.makeLocalPolynomialGrid(2, 0, 2, 1, 'localp'); grid.getDifferentiationWeights(np.array((1.0, 2.0)));", "notError"],
                    ]
 
     def getCustomTabulatedTests(self):

InterfacePython/testMakeUpdate.py

@@ -94,7 +94,7 @@ def checkMakeAgainstKnown(self):
         aW1 = grid.getInterpolationWeights(aX1)
         aW2 = grid.getInterpolationWeights(aX2)
         aW3 = grid.getInterpolationWeights(aX3)
-        aWW = np.row_stack([aW1, aW2, aW3])
+        aWW = np.vstack([aW1, aW2, aW3])
         np.testing.assert_equal(aW, aWW, "Batch weights", True)
 
         grid.makeGlobalGrid(3, 0, 1, 'level', 'chebyshev')
@@ -116,6 +116,11 @@ def checkMakeAgainstKnown(self):
         aA = np.array([[0.0, 0.0, 0.0], [0.0, 0.0, -0.707106781186548], [0.0, 0.0, 0.707106781186548], [0.0, -0.707106781186548, 0.0], [0.0, 0.707106781186548, 0.0], [-0.707106781186548, 0.0, 0.0], [0.707106781186548, 0.0, 0.0]])
         np.testing.assert_almost_equal(aA, grid.getPoints(), 14, "Original equal", True)
 
+        grid = Tasmanian.makeLocalPolynomialGrid(2, 1, 1, 2, "localp")
+        aW = grid.getDifferentiationWeights(np.array((0.5, 0.5)))
+        aA = np.array(((-1.0, -1.0), (0.0, 0.0), (0.0, 1.0), (0.0, 0.0), (1.0, 0.0)))
+        np.testing.assert_almost_equal(aA, aW, 14, "Differentiation weights", True)
+
         # number of points
         grid = Tasmanian.makeLocalPolynomialGrid(2, 1, 2, 2, "localp")
         iNN = grid.getNumNeeded()