walberla integration + bugfix in GPU block indexing

jinja_filters.py

 import sympy as sp
 import jinja2
 import copy
+
+from pystencils import TypedSymbol
 from pystencils.astnodes import ResolvedFieldAccess
 from pystencils.data_types import get_base_type
 from pystencils.backends.cbackend import generate_c, CustomSympyPrinter
@@ -291,7 +293,7 @@ def generate_call(ctx, kernel_info, ghost_layers_to_include=0, cell_interval=Non
         spatial_shape_symbols = []
         for param in ast_params:
             if param.is_field_shape_argument:
-                spatial_shape_symbols = [sp.Symbol("%s_cpu[%d]" % (param.name, i))
+                spatial_shape_symbols = [TypedSymbol("%s_cpu[%d]" % (param.name, i), get_base_type(Field.SHAPE_DTYPE))
                                          for i in range(field.spatial_dimensions)]
 
         if not spatial_shape_symbols:
@@ -301,7 +303,8 @@ def generate_call(ctx, kernel_info, ghost_layers_to_include=0, cell_interval=Non
                 field = fields[param.field_name]
                 if field.field_type == FieldType.GENERIC:
                     kernel_call_lines += create_field_shape_code(field, '_size', gpu_copy=False)
-                    spatial_shape_symbols = [sp.Symbol("_size[%d]" % (i, )) for i in range(field.spatial_dimensions)]
+                    spatial_shape_symbols = [TypedSymbol("_size[%d]" % (i, ), get_base_type(Field.SHAPE_DTYPE))
+                                             for i in range(field.spatial_dimensions)]
                     break
 
         indexing_dict = ast.indexing.call_parameters(spatial_shape_symbols)