prefer "X in Fields()" rather than "X.is_field()"

[videlec]

X.is_field() fails for many parent X. The matrix code relies a lot on the distinction ring/field. We change the code in the matrix folder to make the functions work for more base rings.

The ticket also

• simplifies the code of adjugate
• handles more matrix inversion (as reported in Sage can not invert some known invertible matrix #27869)

Component: algebra

Author: Vincent Delecroix

Branch: 4981d44

Reviewer: Frédéric Chapoton

Issue created by migration from https://trac.sagemath.org/ticket/28189

[videlec]

New commits:

9997ea9

28189: matrices TMP COMMIT

[videlec]

Commit: 9997ea9

[videlec]

Branch: u/vdelecroix/28189

[sagetrac-git]

Changed commit from 9997ea9 to 2a2bbb4

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

2a2bbb4

28189: matrices

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

becd03b	fix a doctest in judson-abstract-algebra
2d16f89	also avoid is_integral_domain

[sagetrac-git]

Changed commit from 2a2bbb4 to 2d16f89

[videlec]

comment:5

Kwankyu: this branch introduces a serious speedup in function_field code that I hardly explain... you might want to investigate.

Git branch: matrices2                             Git branch: develop
sage -t rings/function_field/differential.py      sage -t rings/function_field/differential.py
    [220 tests, 2.46 s]                               [220 tests, 2.81 s]
sage -t rings/function_field/divisor.py           sage -t rings/function_field/divisor.py
    [200 tests, 3.42 s]                               [200 tests, 3.72 s]
sage -t rings/function_field/ideal.py             sage -t rings/function_field/ideal.py
    [915 tests, 8.90 s]                               [915 tests, 9.51 s]
sage -t rings/function_field/valuation_ring.py    sage -t rings/function_field/valuation_ring.py
    [52 tests, 0.23 s]                                [52 tests, 0.33 s]
sage -t rings/function_field/order.py             sage -t rings/function_field/order.py
    [421 tests, 1.71 s]                               [421 tests, 2.18 s]
sage -t rings/function_field/maps.py              sage -t rings/function_field/maps.py
    [378 tests, 3.77 s]                               [378 tests, 5.44 s]
sage -t rings/function_field/constructor.py       sage -t rings/function_field/constructor.py
    [42 tests, 0.08 s]                                [42 tests, 0.10 s]
sage -t rings/function_field/element.pyx          sage -t rings/function_field/element.pyx
    [275 tests, 1.27 s]                               [275 tests, 1.88 s]
sage -t rings/function_field/function_field.py    sage -t rings/function_field/function_field.py
    [743 tests, 33.02 s]                              [743 tests, 34.50 s]
sage -t rings/function_field/function_field_valua sage -t rings/function_field/function_field_valua
    [338 tests, 1.97 s]                               [338 tests, 1.97 s]
sage -t rings/function_field/place.py             sage -t rings/function_field/place.py
    [197 tests, 4.61 s]                               [197 tests, 5.24 s]

[fchapoton]

comment:6

ok, looks good.

[fchapoton]

Reviewer: Frédéric Chapoton

[videlec]

comment:7

Merci Frédéric!

[vbraun]

comment:8

Merge conflict

[fchapoton]

comment:9

probably with #27473

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

c4a4dc9	28189: use "in Fields()" rather than ".is_field()"
fac455d	28189: doctest fix for judson-abstract-algebra
4ef1429	28189: use "in IntegralDomains()" rather than ".is_integral_domain()"

[sagetrac-git]

Changed commit from 2d16f89 to 4ef1429

[fchapoton]

comment:12

Not clear to me what conflict was resolved and how. Could you explain, please ?

[videlec]

comment:13

Oh I see. The conflict is not merged yet.

[videlec]

comment:23

I can rebase but I will not spend my life doing it. Volker, could you merge it first in the 9.0.beta0 if I rebase?

[vbraun]

comment:24

Can do...

[DeRhamSource]

comment:25

Nice, thank you very much! :)

[videlec]

New commits:

5d92db3	28189: prefer R in Fields() rather than R.is_field()
7767b30	28189: fix doctest in judson-abstract-algebra book

[videlec]

Changed branch from public/ticket/28189 to u/vdelecroix/28189

[videlec]

Changed commit from 63ad0a3 to 7767b30

[fchapoton]

comment:28

New: a failing doctest in src/sage/matrix/matrix0.pyx

[jhpalmieri]

comment:29

Is this a py2-only failure? From my limited experimentation, it looks like build_inverse_from_augmented_sparse behaves differently in Python 2 vs. 3.

[mwageringel]

comment:30

Replying to @jhpalmieri:

Is this a py2-only failure? From my limited experimentation, it looks like build_inverse_from_augmented_sparse behaves differently in Python 2 vs. 3.

Yes, see also #28402. It looks like the fix there was accidentally undone here.

[sagetrac-git]

Changed commit from 7767b30 to 4981d44

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

4981d44

fix accidental merge with #28402

[videlec]

comment:32

now fixed

[fchapoton]

comment:33

ok, bien

[vbraun]

Changed branch from u/vdelecroix/28189 to 4981d44

[fchapoton]

Changed commit from 4981d44 to none

[fchapoton]

comment:35

This has broken the inversion of matrices over ZZ, see sage-devel thread.

[jhpalmieri]

comment:36

I think the problem is that the new matrix inverse_of_unit is shadowing the old one from categories.rings.

[fchapoton]

comment:37

follow-up ticket in #28570 to try and fix the broken things ; please help