To fix issue with Betti Tables

[Lax202]

This is an attempt to fix Issue #3798 @wdecker

• I have removed the extra line at the end of the show function of the Betti table.
• I am unsure about the negative degrees in the Betti table. I've made some kind of change but don't have an example yet to test it on. Help with an example with negative cohomological degrees would be great.

[wdecker]

@Lax202 Please, your changes with respect to catching non-gradedness are not helpful. You ONLY need to erase as well as copy and paste as I described in the issue:

Please remove test in the loop and, instead, start the function body with something like

@Assert is_standard_graded(base_ring(F)) "resolution must be defined over a standard graded ring"
@Assert is_graded(F) "resolution must be graded"

That is all.

[Lax202]

@wdecker sorry for misunderstanding. I have added the assert statements.

[wdecker]

Please delete this line:

module_degrees === nothing && error("One of the modules in the graded free resolution is not graded.")

[benlorenz]

joschmitt added the backport 1.0.x label yesterday

I would rather not backport this as this is not really a bugfix and it does change the output of the book examples (even if it is just whitespace). If I understand it correctly this is about doing extra checks to catch invalid input?

[joschmitt]

joschmitt added the backport 1.0.x label yesterday

I would rather not backport this as this is not really a bugfix and it does change the output of the book examples (even if it is just whitespace). If I understand it correctly this is about doing extra checks to catch invalid input?

Ah, okay, sorry. I just read "fix issue".

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.30%. Comparing base (7d851c2) to head (0935040).
Report is 32 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3800      +/-   ##
==========================================
+ Coverage   81.25%   81.30%   +0.05%     
==========================================
  Files         580      579       -1     
  Lines       79289    79229      -60     
==========================================
- Hits        64426    64419       -7     
+ Misses      14863    14810      -53     

Files	Coverage Δ
src/Modules/ModulesGraded.jl	73.35% <100.00%> (-0.03%)	⬇️

... and 27 files with indirect coverage changes

[wdecker]

does change the output of the book examples

Indeed? In the latex version of the book it looks okay.

[benlorenz]

The CI job for the book tests is complaining about the removed lines, e.g.:

  julia> betti(free_resolution(defining_ideal(D)))
  0  1
  -----------
  2    : 3  2
  -----------
  total: 3  2
- 
  
  julia> Oscar.rat_normal_curve_anticanonical_map(D)

But it turns out in the book there is no extra new line. This was probably not flagged because the old Documenter based checking needed some extra filtering which removed the extra newline. So backporting this PR would change the output of 1.0.x but on the other hand adjust it to match the output in the book.

[benlorenz]

So on second thought I think it does make sense to backport this since it will bring 1.0 and the book closer together even if it does change the output of 1.0 slightly.

Please adjust the remaining jlcons, I think exres2.jlcon and param.jlcon, and once it is merged I will check if it applies cleanly to the release branch.