Patch bugs of computation and input of the regulator.

[willian-m]

There were a couple of issues in the first iteration of my implementation of the regulator. Namely:

• The input.cpp was not reading the threshold value from the .yaml
• The Reynolds number calculation in (3+1)D was taking into account the pi^{\mu 3} twice.
• The place where the regulator was being called was wrong and was causing an effect which was unexpected. I changed its location (Reynolds number was not limited). Changing its call to inside the main loop in BSQHydro::run solved the issue. It was also necessary to separate the check of positive entropy from the regulator function.

[willian-m]

@surkhii Review this and see if you understand what is going on and if you agree. If so, just merge the PR (pull request) and delete the patch_regulator branch. If you don't agree, you can reject the patch and we can talk about it.