Add string methods for inverse semigroups/monoids

[james-d-mitchell]

There were no specific methods for String installed for semigroups that happen to be inverse semigroups but which fail to have generators as an inverse semigroup. For example, previously:

gap> S := Semigroup(Transformation([1, 2, 3, 4, 5, 6, 7, 7, 7]), 
>                   Transformation([4, 6, 3, 6, 6, 6, 7, 7, 7]),
>                   Transformation([4, 5, 6, 1, 6, 6, 7, 7, 7]), 
>                   Transformation([6, 6, 3, 1, 6, 6, 7, 7, 7]), 
>                   Transformation([4, 6, 6, 1, 2, 6, 7, 7, 7]));
gap> IsInverseSemigroup(S);
true
gap> String(S);
"InverseSemigroup( ... )"

After this pull request:

gap> S := Semigroup(Transformation([1, 2, 3, 4, 5, 6, 7, 7, 7]), 
>                   Transformation([4, 6, 3, 6, 6, 6, 7, 7, 7]),
>                   Transformation([4, 5, 6, 1, 6, 6, 7, 7, 7]), 
>                   Transformation([6, 6, 3, 1, 6, 6, 7, 7, 7]), 
>                   Transformation([4, 6, 6, 1, 2, 6, 7, 7, 7]));;
gap> IsInverseSemigroup(S);
true
gap> String(S);
"Semigroup( [ Transformation( [ 1, 2, 3, 4, 5, 6, 7, 7, 7 ] ), Transformation(\
 [ 4, 6, 3, 6, 6, 6, 7, 7, 7 ] ), Transformation( [ 4, 5, 6, 1, 6, 6, 7, 7, 7 \
] ), Transformation( [ 6, 6, 3, 1, 6, 6, 7, 7, 7 ] ), Transformation( [ 4, 6, \
6, 1, 2, 6, 7, 7, 7 ] ) ] )"

[codecov-io]

Current coverage is 49.11% (diff: 100%)

Merging #882 into master will increase coverage by 0.04%

@@             master       #882   diff @@
==========================================
  Files           422        422          
  Lines        228711     228715     +4   
  Methods        3447       3447          
  Messages          0          0          
  Branches          0          0          
==========================================
+ Hits         112222     112327   +105   
+ Misses       116489     116388   -101   
  Partials          0          0          

Powered by Codecov. Last update 55d9ff2...af2ee30

[wilfwilson]

This looks good to me.

[markuspf]

Mhm. This looks correct (in that one can copy-and-paste the string to get a GAP object).

For future reference, could you expand why you don't print this as an InverseSemigroup? I assume there is a good reason.

[james-d-mitchell]

There is no (unary) inverse operation for transformations, and so they cannot be created using InverseSemigroup or InverseMonoid. You cannot create a group out of arbitrary transformations for the same reason. So, it wouldn't be valid if it was InverseSemigroup instead of Semigroup.

[markuspf]

I see. Going to merge this then.