Split isochrone into isochrone + isotrips functions

[mpadge]

Following on from discussions in #56 & #57, the current updates to the isochrone function effectively (only) return the best isotrip to each end station. There may have been several candidate trips along other routes, but the eventual selection of a single best route effectively drops these intermediate stations, leaving the midpoints objects containing only a restricted subset of all stations which were actually visited on the way to calculating the isochrone.

In effect, the current version generates some more like "isotrips", representing the single fastest trips out to each end station, and only those midpoint stations visited along the paths of these single best trips. This could be exposed in a few isotrips() function, while the isochrone function itself could return a structure similar to the current version, except which contains all midpoint stations visited along the way, along with associated statistics.

[mpadge]

Closing now, because the whole isochrone function(s) will be deprecated anyway, to be replaced by the new gtfs_traveltimes() function.