Ralf Stephan

``` cdef class Rational(Number): ... def _sympy_(self): rat = self.get_num_den() return rat[0]._sympy_() / rat[1]._sympy_() cdef class Integer(Number): ... def _sympy_(self): import sympy return sympy.Integer(deref(self.thisptr).__str__().decode("utf-8")) ```

What about this snippet in the `sympy.Rational` constructor? ``` if not isinstance(p, SYMPY_INTS + (Rational,)): raise TypeError('invalid input: %s' % p) ```

Ah ok, in `compatibility.py`: ``` if GROUND_TYPES == 'gmpy': SYMPY_INTS += (type(gmpy.mpz(0)),) ```

Rubi contains a lot of potential structure and matchpy is overkill for this problem, not only because parallel matching is overkill, but also because there are practically no sequential variables...

Ah got it, these were deprecated claims. Should they appear at all?

It is not JSON but a simple command language made for the purpose, please see https://www.wikidata.org/wiki/Help:QuickStatements#Command_sequence_syntax

reason is: on the first atom change the mass number is set from getMajorIsotope() but with the second change we assume the mass number set and try to get that...

crash is fixed, leaving the better error handling as enhancement

I'm no longer working on JCP. The best option you have at the moment is to compile the applet yourself from source. On 3 Feb 2014 14:46, "LaiaNavarro" [email protected] wrote:...

Thanks for the link!