Fix problems with decode_as_matrix when rep_type=GLFq

[roed314]

Compare

• https://beta.lmfdb.org/GaloisGroup/15T34 (which gives a server error)
• https://localhost:37777/GaloisGroup/15T34

I also removed ^1 in the display of matrices over Fq and added the minimal polynomial of $$\alpha$$.

There will be compatibility problems between Sage and Magma if we end up outside the range where Conway polynomials are available (we can't just use GF(q, modulus="primitive") since we need to match the polynomial chosen when defining the field as the data is computed, and we don't currently store that polynomial). Fortunately, the current set of qs is

4,8,9,16,25,27,32,49,64,81,121,125,128,169,243,256,289,343,361,529,625,729,841,961

and Conway polynomials are available in all of these cases.

[jenpaulhus]

I tested a bunch of group pages which have GLFq representatives and they all seem to work. Code looks good to me so I will go ahead and merge.