As part of a project, I am translating some MAGMA code to SAGE. The relevant piece of code computes the maximal order of the relative extension of Qa12, a number field of degree 20, by the polynomial y^2 - kappa12:
subOrderK:=ext<OO | y^2-kappa12>;
subOrderK:=AbsoluteOrder(subOrderK);
D:=Discriminant(subOrderK);
for p in PrimeDivisors(D) do
subOrderK:=pMaximalOrder(subOrderK,p);
end for;
OOK:=subOrderK;
Here OO is the ring of integers (i.e. the maximal order) of Qa12.
As I did not see a way to translate this word for word (if I am missing something, please point it out), I tried a different approach. Here is my code (in SAGE):
L.<c> = Qa12.extension(y^2-kappa12)
L.<alpha> = L.absolute_field()
subOrderK = L.order(alpha)
D = subOrderK.discriminant()
for p in factor(D):
subOrderK = L.maximal_order(p[0])
OOK = subOrderK
Note that y^2-kappa12 has coefficients coerced in Qa12 and we make sure that it is indeed irreducible in Qa12 (i.e. we make sure that kappa12 is not the square of an element of Qa12).
You may notice immediately that my for-loop is a rather clumsy translation of the corresponding loop in the MAGMA code. I was hoping that SAGE would "remember" the previous value of subOrderK, thus having the same effect as the MAGMA-command pMaximalOrder(subOrderK,p). However, my problem arises even earlier than that:
the number computed by subOrderK.discriminant() is absurdly huge - too big, in fact, for there to be any hope to factorise it in any reasonable amount of time.
The obvious alternative of simply writing
L.<c> = Qa12.extension(y^2-kappa12)
L.<alpha> = L.absolute_field()
OOK = L.maximal_order()
is also extremely time-intensive; I have not yet seen this finish.
I was hoping someone could help either with improving my code, or with a radically different approach to the problem.
NOTE: Though I have tried to make this question understandable, if some parts remain opaque, please do say so. I'll do my best to correct it.
