Drug binding macro that mimics experimental settings when a perturbation is added after a reaction is initiated

[ortega2247]

This macro generates the reversible binding reaction DRUG + SUBSTRATE | DRUG:SUBSTRATE that only gets triggered when the simulation reaches the time point t_action. The idea of this macro is to mimic experimental settings when a reaction is started and later on some kind of perturbation is added to the system. This type of reaction has been used fairly frequent in the Lopez lab and I thought it could be added as a macro.

[coveralls]

Coverage decreased (-0.4%) to 79.198% when pulling 85d280e on LoLab-VU:drug_binding_macro into b8e036c on pysb:master.

[alubbock]

I like the idea. Some thoughts which come to mind

• Perhaps the components for generating t_obs should be factored out into a separate macro, since having a time observable is a common requirement. I'd probably rename t_obs to t, k_t to __k_t and Monomer t to __t. This helps to separate "internal" components from the component of interest to the modeler, t (the observable).
• The time tracking approach won't work on stochastic simulations. This should be clearly stated in the docs.
• The component names are hard-coded, which limits multiple applications of the macro within a model. I'd suggest generating these names dynamically, e.g. kf_<drug name>_<substrate name>

[alubbock]

I would suggest upgrading the Sphinx documentation notes to warnings, since it's important to be clear that using e.g. SSA with the macro will lead to incorrect results.

Otherwise, I think this looks good to merge.

[ortega2247]

I just made the change from notes to warnings

[alubbock]

Looks good to me. Thanks!