Add more UnitTest functions similar to those in MonoidalCategories

[kamalsaleh]

CAP_project/MonoidalCategories/gap/MonoidalCategoriesTensorProductOnObjectsAndTensorUnitTest.gi

Lines 7 to 31 in 9c6986f

	InstallGlobalFunction( "MonoidalCategoriesTensorProductOnObjectsAndTensorUnitTest",
	function( cat, opposite, a, b )
	local verbose,
	u, u_op,
	a_op, b_op,
	a_tensor_b, a_tensor_b_op,
	b_tensor_a, b_tensor_a_op;
	a_op := Opposite( opposite, a );
	b_op := Opposite( opposite, b );
	verbose := ValueOption( "verbose" ) = true;
	if CanCompute( cat, "TensorUnit" ) then
	if verbose then
	# COVERAGE_IGNORE_NEXT_LINE
	Display( "Testing 'TensorUnit' ..." );
	fi;

It would be nice to have similar functions for the various doctrines (ab-categories, linear categories, additive categories, Abelian categories, triangulated categories, ...) and probably would be even better if they somehow expose the tests and the equalities that should hold in the manual file by generating examples (maybe using RandomObject/Morphism etc ... methods).

[kamalsaleh]

and also get rid of ValueOption for Julia :)

[TKuh]

It would be nice to have similar functions for the various doctrines (ab-categories, linear categories, additive categories, Abelian categories, triangulated categories, ...) and probably would be even better if they somehow expose the tests and the equalities that should hold in the manual file by generating examples (maybe using RandomObject/Morphism etc ... methods).

I couldn't agree more. The idea in MonoidalCategories was once to someday generate tests from the method records. This was back then the reason for these various comments:

CAP_project/MonoidalCategories/gap/MonoidalCategoriesTensorProductOnObjectsAndTensorUnitMethodRecord.gi

Line 14 in 9c6986f

# Test in MonoidalCategoriesTensorProductOnObjectsAndTensorUnitTest

[kamalsaleh]

Does the method record contain what equalities and properties should hold for each operation?

[TKuh]

Does the method record contain what equalities and properties should hold for each operation?

No, sadly not, so there would be much more work to do for math related checks among operations. But the method records are at least a starting point and they would suffice for simple function calls with random data and for consistency checks with the opposite category (which I guess is already not everywhere the case).