[SciPy-User] Fwd: Can scipy.sparse add support to other numeric type?

[cat fa]

Hi everyone,
    In the source code of ..\scipy\sparse\sputils.py, the upcast
function only support these types

int8, uint8,
short, ushort,
intc, uintc,
longlong, ulonglong,
single, double, longdouble,
csingle, cdouble, clongdouble.

Now I want to use sparse matrix to store ad(an automatic
differentiation object) type. When I do some calculation operation, I
got TypeError exception:

TypeError: no supported conversion for types: object


Is there a easy way to get sparse matrix to support other numeric type?