Improve or clarify the reportNum parameter

[yxngl]

Currently because cluster representatives of redundancy reduction methods are always added to the top results shown in the webpage, the actual number of gene sets could exceed the reportNum threshold. Or the information will not be available when AP or WSC is chosen. Either we can store a separate list for them or figure out a better way since the redundancy reduction need some changes in the new version.

Another related problem is when representatives are rescued, they are just appended to the end without sorting again. This could be fixed quickly.

[iblacksand]

In the new package, users can enable and disable the redundancy removal, so the size of this effect will change depending on what they enable.

I'll fix this in my fork. I think not showing the representatives in the All table could be a nice middle ground. Although it may be confusing for new users when they see gene sets that are not found in the top $n$ list.

I mirrored this issue in my fork at iblacksand#7 and experiment with different options.