fixed create_kernel parameter data_type="float" to procucde single precision

Currently if create_kernel(assignments, data_type="float") is used then the untyped symbols are typed with float64, since the np.dtype("float") creates this during the construction of a new TypedSymbol.

Since data_type or as it is called in cpu.create_kernel type_info can be an string of an C type, At least following the documentation of cpu.create_kernel this behavior is a bit confusing, since typical the C type specifier "float" is meant to be single precision.

So I added a small function that just replaces "float" with "single" in the symbol_to_type dict, so the untyped symbols get the single precision type.

pystencils_tests/test_kernel_data_type.py

@@ -5,9 +5,18 @@ from pystencils import Assignment, create_kernel, fields
 
 
 def test_single_precision():
-    s = fields("s: float32[1D]")
+    s = fields("s: double[1D]")
     assignments = [Assignment(x, y), Assignment(s[0], x)]
     ast = create_kernel(assignments, data_type="float")
     assert ast.body.args[0].lhs.dtype.numpy_dtype == np.dtype("float32")
     assert ast.body.args[0].rhs.dtype.numpy_dtype == np.dtype("float32")
     assert ast.body.args[1].body.args[0].rhs.dtype.numpy_dtype == np.dtype("float32")
+
+
+def test_double_precision():
+    s = fields("s: double[1D]")
+    assignments = [Assignment(x, y), Assignment(s[0], x)]
+    ast = create_kernel(assignments, data_type="double")
+    assert ast.body.args[0].lhs.dtype.numpy_dtype == np.dtype("float64")
+    assert ast.body.args[0].rhs.dtype.numpy_dtype == np.dtype("float64")
+    assert ast.body.args[1].body.args[0].rhs.dtype.numpy_dtype == np.dtype("float64")