local field unramified field failing tests

[edgarcosta]

on https://www.lmfdb.org/padicField/11.6.4.2 we have x^2 - x + 7 but on https://beta.lmfdb.org/padicField/11.6.4.2 we have x^2 + 7*x + 2 and the tests expect the former: https://github.com/LMFDB/lmfdb/blob/master/lmfdb/local_fields/test_localfields.py#L27-L30 which @jwj61 agrees that are not robust

Furthermore, the data for both fields seem to differ, in particular the entry for unram

{'aut': 3, 'c': 4, 'coeffs': [847, 0, 0, -11, 0, 0, 1], 'e': 3, 'eisen': 'y^3 - 11t', 'f': 2, 'gal': 5, 'galT': 5, 'galois_label': '6T5', 'gms': '2/3', 'gsm': [847, 0, 0, -11, 0, 0, 1], 'hw': '$1$', 'ind_of_insep': [0], 'inertia': ['i', [3, 1]], 'inertia_gap': [3, 1], 'is_completion': True, 'label': '11.6.4.2', 'n': 6, 'p': 11, 'ram_poly_vert': [[1, 0], [3, 0]], 'residual_polynomials': ['z^2 + 3z + 3'], 'rf': [1, 1], 'slopes': '[]', 'subfield': ['11.2.0.1'], 'subfield_mult': [1], 'subfields': [['7,-1,1', 1]], 't': 3, 'top_slope': '01.0000000001', 'u': 6, 'unram': 't^2 - t + 7', 'visible': '[]', 'wild_gap': [1, 1]}

vs

{'associated_inertia': [1], 'aut': 3, 'c': 4, 'coeffs': [847, 0, 0, -11, 0, 0, 1], 'e': 3, 'eisen': 'y^3 - 11t', 'f': 2, 'gal': 5, 'galT': 5, 'galois_label': '6T5', 'gms': '2/3', 'gsm': [847, 0, 0, -11, 0, 0, 1], 'hw': '$1$', 'ind_of_insep': [0], 'ind_of_insep_tmp': ['0A'], 'inertia': ['i', [3, 1]], 'inertia_gap': [3, 1], 'is_completion': True, 'label': '11.6.4.2', 'n': 6, 'p': 11, 'ram_poly_vert': [[1, 0], [3, 0]], 'residual_polynomials': ['z^2 + 3z + 3'], 'rf': [1, 1], 'slopes': '[]', 'slopes_tmp': [], 'subfield': ['11.2.0.1'], 'subfield_mult': [1], 'subfields': [['7,-1,1', 1]], 't': 3, 'top_slope': '01.0000000001', 'u': 6, 'unram': 't^2 - t + 7', 'visible': '[]', 'visible_tmp': [], 'wild_gap': [1, 1]}

[roed314]

Yeah, @jwj61 has been changing the data as part of our AIM SQuaRE this week, using different polynomials for unramified extensions.

[roed314]

I've been making changes to the LMFDB code as part of it, so I can work on updating the tests as part of that upcoming PR.

[jwj61]

We should comment out that test for the time being. The data changed, and will probably change again. We can reinstate it when that process is done.