api: cleanup weight kwargs

devito/finite_differences/derivative.py

@@ -218,17 +218,15 @@ def _process_weights(cls, **kwargs):
         else:
             return as_tuple(weights)
 
-    def __call__(self, x0=None, fd_order=None, side=None, method=None,
-                 weights=None, w=None):
+    def __call__(self, x0=None, fd_order=None, side=None, method=None, **kwargs):
+        weights = kwargs.get('weights', kwargs.get('w'))
         rkw = {}
         if side is not None:
             rkw['side'] = side
         if method is not None:
             rkw['method'] = method
         if weights is not None:
             rkw['weights'] = weights
-        if w is not None:
-            rkw['weights'] = w
 
         if x0 is not None:
             x0 = self._process_x0(self.dims, x0=x0)

devito/finite_differences/differentiable.py

@@ -293,7 +293,7 @@ def laplace(self):
         """
         return self.laplacian()
 
-    def laplacian(self, shift=None, order=None, method='FD', w=None):
+    def laplacian(self, shift=None, order=None, method='FD', **kwargs):
         """
         Laplacian of the Differentiable with shifted derivatives and custom
         FD order.
@@ -309,7 +309,13 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
         order: int, optional, default=None
             Discretization order for the finite differences.
             Uses `func.space_order` when not specified
+        method: str, optional, default='FD'
+            Discretization method. Options are 'FD' (default) and
+            'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite differences.
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         order = order or self.space_order
         space_dims = [d for d in self.dimensions if d.is_Space]
         shift_x0 = make_shift_x0(shift, (len(space_dims),))
@@ -318,7 +324,7 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
                                       method=method, fd_order=order, w=w)
                      for i, d in enumerate(derivs)])
 
-    def div(self, shift=None, order=None, method='FD', w=None):
+    def div(self, shift=None, order=None, method='FD', **kwargs):
         """
         Divergence of the input Function.
 
@@ -334,15 +340,18 @@ def div(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite difference coefficients.
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         space_dims = [d for d in self.dimensions if d.is_Space]
         shift_x0 = make_shift_x0(shift, (len(space_dims),))
         order = order or self.space_order
         return Add(*[getattr(self, 'd%s' % d.name)(x0=shift_x0(shift, d, None, i),
                                                    fd_order=order, method=method, w=w)
                      for i, d in enumerate(space_dims)])
 
-    def grad(self, shift=None, order=None, method='FD', w=None):
+    def grad(self, shift=None, order=None, method='FD', **kwargs):
         """
         Gradient of the input Function.
 
@@ -358,11 +367,14 @@ def grad(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite
         """
         from devito.types.tensor import VectorFunction, VectorTimeFunction
         space_dims = [d for d in self.dimensions if d.is_Space]
         shift_x0 = make_shift_x0(shift, (len(space_dims),))
         order = order or self.space_order
+        w = kwargs.get('weights', kwargs.get('w'))
         comps = [getattr(self, 'd%s' % d.name)(x0=shift_x0(shift, d, None, i),
                                                fd_order=order, method=method, w=w)
                  for i, d in enumerate(space_dims)]

devito/finite_differences/operators.py

@@ -1,4 +1,4 @@
-def div(func, shift=None, order=None, method='FD', weights=None, w=None):
+def div(func, shift=None, order=None, method='FD', **kwargs):
     """
     Divergence of the input Function.
 
@@ -14,10 +14,10 @@ def div(func, shift=None, order=None, method='FD', weights=None, w=None):
     method: str, optional, default='FD'
         Discretization method. Options are 'FD' (default) and
         'RSFD' (rotated staggered grid finite-difference).
-    weights: list, tupe, or dict, optional, default=None
+    weights/w: list, tuple, or dict, optional, default=None
         Custom weights for the finite difference coefficients.
     """
-    w = weights or w
+    w = kwargs.get('weights', kwargs.get('w'))
     try:
         return func.div(shift=shift, order=order, method=method, w=w)
     except AttributeError:
@@ -41,7 +41,7 @@ def div45(func, shift=None, order=None):
     return div(func, shift=shift, order=order, method='RSFD')
 
 
-def grad(func, shift=None, order=None, method='FD', weights=None, w=None):
+def grad(func, shift=None, order=None, method='FD', **kwargs):
     """
     Gradient of the input Function.
 
@@ -57,10 +57,10 @@ def grad(func, shift=None, order=None, method='FD', weights=None, w=None):
     method: str, optional, default='FD'
         Discretization method. Options are 'FD' (default) and
         'RSFD' (rotated staggered grid finite-difference).
-    weights: list, tupe, or dict, optional, default=None
+    weights/w: list, tuple, or dict, optional, default=None
         Custom weights for the finite difference coefficients.
     """
-    w = weights or w
+    w = kwargs.get('weights', kwargs.get('w'))
     try:
         return func.grad(shift=shift, order=order, method=method, w=w)
     except AttributeError:
@@ -84,7 +84,7 @@ def grad45(func, shift=None, order=None):
     return grad(func, shift=shift, order=order, method='RSFD')
 
 
-def curl(func, shift=None, order=None, method='FD', weights=None, w=None):
+def curl(func, shift=None, order=None, method='FD', **kwargs):
     """
     Curl of the input Function. Only supported for VectorFunction
 
@@ -100,10 +100,10 @@ def curl(func, shift=None, order=None, method='FD', weights=None, w=None):
     method: str, optional, default='FD'
         Discretization method. Options are 'FD' (default) and
         'RSFD' (rotated staggered grid finite-difference).
-    weights: list, tupe, or dict, optional, default=None
+    weights/w: list, tuple, or dict, optional, default=None
         Custom weights for the finite difference coefficients.
     """
-    w = weights or w
+    w = kwargs.get('weights', kwargs.get('w'))
     try:
         return func.curl(shift=shift, order=order, method=method, w=w)
     except AttributeError:
@@ -128,7 +128,7 @@ def curl45(func, shift=None, order=None):
     return curl(func, shift=shift, order=order, method='RSFD')
 
 
-def laplace(func, shift=None, order=None, method='FD', weights=None, w=None):
+def laplace(func, shift=None, order=None, method='FD', **kwargs):
     """
     Laplacian of the input Function.
 
@@ -143,10 +143,10 @@ def laplace(func, shift=None, order=None, method='FD', weights=None, w=None):
         Uses `func.space_order` when not specified
     method: str, optional, default='FD'
         Discretization method. Options are 'FD' (default) and 'RSFD'
-    weights: list, tupe, or dict, optional, default=None
+    weights/w: list, tuple, or dict, optional, default=None
         Custom weights for the finite difference coefficients.
     """
-    w = weights or w
+    w = kwargs.get('weights', kwargs.get('w'))
     try:
         return func.laplacian(shift=shift, order=order, method=method, w=w)
     except AttributeError:

devito/types/tensor.py

@@ -188,7 +188,7 @@ def values(self):
         else:
             return super().values()
 
-    def div(self, shift=None, order=None, method='FD', w=None):
+    def div(self, shift=None, order=None, method='FD', **kwargs):
         """
         Divergence of the TensorFunction (is a VectorFunction).
 
@@ -202,7 +202,10 @@ def div(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite differences.
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         comps = []
         func = vec_func(self)
         ndim = len(self.space_dimensions)
@@ -222,7 +225,7 @@ def laplace(self):
         """
         return self.laplacian()
 
-    def laplacian(self, shift=None, order=None, method='FD', w=None):
+    def laplacian(self, shift=None, order=None, method='FD', **kwargs):
         """
         Laplacian of the TensorFunction with shifted derivatives and custom
         FD order.
@@ -238,7 +241,13 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
         order: int, optional, default=None
             Discretization order for the finite differences.
             Uses `func.space_order` when not specified
+        method: str, optional, default='FD'
+            Discretization method. Options are 'FD' (default) and
+            'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         comps = []
         func = vec_func(self)
         order = order or self.space_order
@@ -251,7 +260,7 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
                               for i, d in enumerate(self.space_dimensions)]))
         return func._new(comps)
 
-    def grad(self, shift=None, order=None, method=None, w=None):
+    def grad(self, shift=None, order=None, method=None, **kwargs):
         raise AttributeError("Gradient of a second order tensor not supported")
 
     def new_from_mat(self, mat):
@@ -314,7 +323,7 @@ def __str__(self):
 
     __repr__ = __str__
 
-    def div(self, shift=None, order=None, method='FD', w=None):
+    def div(self, shift=None, order=None, method='FD', **kwargs):
         """
         Divergence of the VectorFunction, creates the divergence Function.
 
@@ -328,7 +337,10 @@ def div(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite difference coefficients.
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         shift_x0 = make_shift_x0(shift, (len(self.space_dimensions),))
         order = order or self.space_order
         return sum([getattr(self[i], 'd%s' % d.name)(x0=shift_x0(shift, d, None, i),
@@ -342,7 +354,7 @@ def laplace(self):
         """
         return self.laplacian()
 
-    def laplacian(self, shift=None, order=None, method='FD', w=None):
+    def laplacian(self, shift=None, order=None, method='FD', **kwargs):
         """
         Laplacian of the VectorFunction, creates the Laplacian VectorFunction.
 
@@ -353,7 +365,13 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
         order: int, optional, default=None
             Discretization order for the finite differences.
             Uses `func.space_order` when not specified
+        method: str, optional, default='FD'
+            Discretization method. Options are 'FD' (default) and
+            'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         func = vec_func(self)
         shift_x0 = make_shift_x0(shift, (len(self.space_dimensions),))
         order = order or self.space_order
@@ -363,7 +381,7 @@ def laplacian(self, shift=None, order=None, method='FD', w=None):
                  for s in self]
         return func._new(comps)
 
-    def curl(self, shift=None, order=None, method='FD', w=None):
+    def curl(self, shift=None, order=None, method='FD', **kwargs):
         """
         Gradient of the (3D) VectorFunction, creates the curl VectorFunction.
 
@@ -377,10 +395,13 @@ def curl(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite difference coefficients.
         """
         if len(self.space_dimensions) != 3:
             raise AttributeError("Curl only supported for 3D VectorFunction")
         # The curl of a VectorFunction is a VectorFunction
+        w = kwargs.get('weights', kwargs.get('w'))
         dims = self.space_dimensions
         derivs = ['d%s' % d.name for d in dims]
         shift_x0 = make_shift_x0(shift, (len(dims), len(dims)))
@@ -400,7 +421,7 @@ def curl(self, shift=None, order=None, method='FD', w=None):
         func = vec_func(self)
         return func._new(3, 1, [comp1, comp2, comp3])
 
-    def grad(self, shift=None, order=None, method='FD', w=None):
+    def grad(self, shift=None, order=None, method='FD', **kwargs):
         """
         Gradient of the VectorFunction, creates the gradient TensorFunction.
 
@@ -414,7 +435,10 @@ def grad(self, shift=None, order=None, method='FD', w=None):
         method: str, optional, default='FD'
             Discretization method. Options are 'FD' (default) and
             'RSFD' (rotated staggered grid finite-difference).
+        weights/w: list, tuple, or dict, optional, default=None
+            Custom weights for the finite difference coefficients.
         """
+        w = kwargs.get('weights', kwargs.get('w'))
         func = tens_func(self)
         ndim = len(self.space_dimensions)
         shift_x0 = make_shift_x0(shift, (ndim, ndim))