ENH: Schechter sampler resolution should be a required argument

[rrjbca]

We should consider making resolution a required argument of the schechter sampler and other functions that call it. Currently default values are being used throughout the library e.g. schechter_lf but a single fixed default value can lead to vastly different levels of numerical accuracy depending on the other parameters alpha, x_min and x_max. By requiring the user to choose the resolution we avoid them blindly using a default value that may be unsuitable for their use case and instead implicitly ask them to consider their own numerical requirements.