Truncate printing of large expressions

[odow]

Closes #3574

I don't know if this is the right approach, but we should do something.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (283a99d) 98.38% compared to head (22bd2a4) 98.38%.
Report is 5 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #3575   +/-   ##
=======================================
  Coverage   98.38%   98.38%           
=======================================
  Files          37       38    +1     
  Lines        5629     5635    +6     
=======================================
+ Hits         5538     5544    +6     
  Misses         91       91           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pulsipher]

It would be nice to also truncate terms in flattened nonlinear sums. Should be straightforward to add for expressions of the form: NonlinearExpr(:+, [nlp_sum_terms...]).

[odow]

Truncated the nonlinear stuff.

My main hesitation is the use of TERM_LIMIT_FOR_PRINTING as a global constant. But maybe that's okay. It's really a debugging flag. Maybe we should just document that extensions should not modify this, leave it to the user. And that extensions should never rely on the output of function_string remaining stable.

[pulsipher]

My main hesitation is the use of TERM_LIMIT_FOR_PRINTING as a global constant. But maybe that's okay. It's really a debugging flag. Maybe we should just document that extensions should not modify this, leave it to the user. And that extensions should never rely on the output of function_string remaining stable.

This seems reasonable from an extension point of view, so long as function_string doesn't have significant changes that impact testing frequently. In this case, having the default term limit be 60 is plenty large and likely will not impact any testing that relies on printing. For InfiniteOpt, I don't think this will break any tests since the string checks are for small expressions.

[odow]

@blegat thoughts?

[blegat]

I think it's yet another proof that we should unify printing of JuMP and MOI. In MOI we have _PrintOptions and this would be a good candidate of additional option for _PrintOption. This would be much cleaner than modifying a global constant.

[odow]

Should or shouldn't?

[blegat]

Indeed, I meant should

[odow]

Should we make it a private constant for now then? Merge this PR, and open an issue to unify the printing?

[blegat]

Yes, if it's private then I don't have any objection to merge this

[odow]

Made it a private constant

[odow]

https://github.com/jump-dev/JuMP.jl/actions/runs/7009559319