Added Doctests for QuadraticForms methods

[jonhanke]

Adding Doctests to bring coverage up to 100%.

CC: @sagetrac-mabshoff @williamstein @tornaria

Component: quadratic forms

Keywords: quadraticform

Issue created by migration from https://trac.sagemath.org/ticket/6040

[jonhanke]

comment:1

Attachment: patch-3__QF_misc_doctests__3.4.1.patch.gz

Note: There are currently two broken doctests in this patch (using the older routine IsPadic Square()), which should resolve themselves once Cremona's patch (Ticket #5834) is applied.

[jonhanke]

comment:3

Additional patch to bring QuadraticForm doctests to 100%.

Known Issues:
Several doctests fail because of the need for hilbert_symbol() to accept rational numbers, and similarly for IsPadicSquare(). These should be addressed by the changes made in Ticket #5834.

[sagetrac-mabshoff]

comment:4

Attachment: patch-4__QF_more_doctests__3.4.1.patch.gz

Together with #5954 this brings coverage in the QF code to 100%.

Cheers,

Michael

[jonhanke]

comment:5

Also the patch in Ticket #6037 (rewrite and careful documentation of local density routines) is related to getting the doctest coverage to 100%.

[tornaria]

comment:6

So, it turns out there are 4 patches in this series, and they must be applied in order. In particular, patch-3 depends on patch-2, which is at #6037, but I misunderstood that.
The correct sequence is to apply patch-1 in #5954, then patch-2 in #6037, then patch-3 and patch-4 in this ticket.

If that order is followed, the patch sequence applies cleanly to 3.4.1 as well as 4.0.alpha0.

[tornaria]

Attachment: patch-5__QF_reviewer__4.0.alpha0.patch.gz

fix doctests for 4.0.alpha0

[tornaria]

comment:7

Some doctests were broken on 4.0.alpha0 + patch-1 (#5954) + patch-2 (#6037) + patch-3 + patch-4.

All doctests pass for me when adding on top of that

• attachment: patch-5__QF_reviewer__4.0.alpha0.patch
• patch in speed regresion in hilbert_symbol after #5834 #6059
• patch in QQ method is_padic_square fails when argument is python int #6062
• patch in allow rationals in kronecker_symbol and legendre_symbol #6064

Note that the patch-3__QF_misc_doctests__4.0.alpha0.patch I uploaded earlier is a mistake, just ignore it.

[tornaria]

comment:8

Note: the patch-5 also adds a few "#long time" comments to file quadratic_form__local_representation_conditions.py. Skipping these reduces the time for doctesting the whole file from 48 s to 20 s.

[sagetrac-mabshoff]

comment:9

Add positive review due to Gonzalo.

Cheers,

Michael

[tornaria]

comment:10

I'm ok with the positive review. I'll list the renamed/removed functions as for the other tickets, and I'll post a ticket to add compatibility functions with deprecation warnings.

• hanke_mass__maximal was removed.
• GHY_mass_maximal was renamed GHY_mass__maximal.
• conway_generic_mass was removed.
• conway_p_mass_adjustment was removed.
• local_diagonal was removed.
• find_primitive_p_divisible_vector__all was removed.

[sagetrac-mabshoff]

comment:11

Merged in Sage 4.0.rc0.

Cheers,

Michael