new ensemble class that allows compact change of parameters

[dprovasi]

to run with multiple values of model parameters one would ideally have
a system class that contains the current trajectory class and the reactions associated with it.

we can have the ensemble class have a system generator (i.e. a function that takes parameters, and returns the system). the ensemble should then have a method to create a number of instances, run them, and the analysis methods we have right now.
the analysis methods should return a pd dtaframe that includes the model parameters to facilitate plotting

def setup_system(params):
    return system1

em = ensemble(setup_system, {diffusions={'mono':1, 'dimer'=10}, ...})
em.add_trajs(10)
em.runtrajs(frames=10000, which='all')
em.add.trajs(5)
em.runtrajs(frames=1000, which=arange(10,15))

em.lifetimes()