api: Minor fixes to arithmetic operations with scalar and tensors

devito/ir/support/utils.py

@@ -267,7 +267,7 @@ def pull_dims(exprs, flag=True):
     """
     dims = set()
     for e in as_tuple(exprs):
-        dims.update({i for i in e.free_symbols if i.is_Dimension})
+        dims.update({i for i in e.free_symbols if isinstance(i, Dimension)})
     if flag:
         return set().union(*[d._defines for d in dims])
     else:

devito/passes/iet/misc.py

@@ -219,7 +219,7 @@ def abridge_dim_names(iet):
         # Find SubDimensions or SubDimension-derived dimensions used as indices in
         # the expression in the innermost loop
         indexeds = FindSymbols('indexeds').visit(tree.inner)
-        dims = set().union(*[pull_dims(i, flag=False) for i in indexeds])
+        dims = pull_dims(indexeds, flag=False)
         dims = [d for d in dims if any([dim.is_Sub for dim in d._defines])]
         dims = [d for d in dims if not d.is_SubIterator]
         names = [d.root.name for d in dims]

devito/types/basic.py

@@ -8,6 +8,7 @@
 import sympy
 
 from sympy.core.assumptions import _assume_rules
+from sympy.core.decorators import call_highest_priority
 from cached_property import cached_property
 
 from devito.data import default_allocator
@@ -704,6 +705,15 @@ def adjoint(self, inner=True):
         # Real valued adjoint is transpose
         return self.transpose(inner=inner)
 
+    @call_highest_priority('__radd__')
+    def __add__(self, other):
+        try:
+            # Most case support sympy add
+            return super().__add__(other)
+        except TypeError:
+            # Sympy doesn't support add with scalars
+            return self.applyfunc(lambda x: x + other)
+
     def _eval_matrix_mul(self, other):
         """
         Copy paste from sympy to avoid explicit call to sympy.Add

devito/types/equation.py

@@ -106,7 +106,12 @@ def _flatten(self):
         """
         if self.lhs.is_Matrix:
             # Maps the Equations to retrieve the rhs from relevant lhs
-            eqs = dict(zip(as_tuple(self.lhs), as_tuple(self.rhs)))
+            try:
+                eqs = dict(zip(self.lhs, self.rhs))
+            except TypeError:
+                # Same rhs for all lhs
+                assert not self.rhs.is_Matrix
+                eqs = {i: self.rhs for i in self.lhs}
             # Get the relevant equations from the lhs structure. .values removes
             # the symmetric duplicates and off-diagonal zeros.
             lhss = self.lhs.values()