Clean imports related to coding theory

[sagetrac-dlucas]

For now, a lot of modules are imported from sage.coding when Sage starts.
These modules can be lazy imported instead for better performances.

In all.py there's also several deprecated imports related to a 2-year old ticket.
These deprecation warnings can be removed too.

CC: @nathanncohen

Component: coding theory

Author: David Lucas

Branch: bc6ecf9

Reviewer: Vincent Delecroix, Jeroen Demeyer

Issue created by migration from https://trac.sagemath.org/ticket/19315

[sagetrac-dlucas]

Branch: u/dlucas/clean_imports

[sagetrac-dlucas]

comment:3

I removed deprecation warnings in all.py, passed everything in lazy imports and fixed a few doctests that broke in the operation.

---

New commits:

8e42c26

Removed deprecation warnings and lazy imported everything

[sagetrac-dlucas]

Commit: 8e42c26

[sagetrac-dlucas]

Author: David Lucas

[sagetrac-git]

Changed commit from 8e42c26 to dadb126

[sagetrac-git]

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

dadb126

Removed lazy import for self_dual_codes_binary

[sagetrac-dlucas]

comment:5

As putting self_dual_codes_binary in a lazy import seemed to mess up with the garbage collector doctests in categories/homset.py and structure/coerce.pyx I changed it for a hard import instead.

[videlec]

comment:6

Salut David,

1. The function cyclotomic_cosets was deprecated in cyclotomic_cosets for any finite ring #16464 which is also more than one year old. Could you remove the function and the import as well?

2. Do you really want to keep codesize_upper_bound, dimension_upper_bound, etc in the global namespace?

Vincent

[videlec]

Reviewer: Vincent Delecroix

[sagetrac-dlucas]

comment:7

Hello Vincent,

The function cyclotomic_cosets was deprecated in #16464 which is also more than one year old. Could >you remove the function and the import as well?

Sure, I'll do that.

Do you really want to keep codesize_upper_bound, dimension_upper_bound, etc in the global namespace?

Not really. I can put them in a catalog so one could access them typing sage.codes.bounds or something like that. It would be better. But if I do that, I guess I need to put a new deprecation warning over these bounds, isn't it?

David

[jdemeyer]

comment:8

Can you please also remove the ugly backslashes, they are not needed inside (...) or [...].

[sagetrac-git]

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

9110eaf	Removed deprecated cyclotomic_cosets
bc6ecf9	New catalog for bounds. Some changes in all.py, fixed doctests.

[sagetrac-git]

Changed commit from dadb126 to bc6ecf9

[sagetrac-dlucas]

comment:10

I removed cyclotomic_cosetsmethod, add a new catalog for bounds so they won't be available in the global namespace... When the new deprecation warning I added will be deleted in a year. I fixed the doctests that went down in the process.

I also removed the backslashes.

[videlec]

comment:11

I am good with this ticket. Jeroen?

[videlec]

Changed reviewer from Vincent Delecroix to Vincent Delecroix, ​Jeroen Demeyer

[jdemeyer]

comment:12

I haven't actually reviewed, I just looked a diff and saw too many backslashes hurting my eyes :-)

[vbraun]

Changed branch from u/dlucas/clean_imports to bc6ecf9

[jdemeyer]

Changed reviewer from Vincent Delecroix, ​Jeroen Demeyer to Vincent Delecroix, Jeroen Demeyer

[jdemeyer]

Changed commit from bc6ecf9 to none