WIP support split service

[gcamp]

Support split service by adding

• a boolean to define if the service is split
• a secondary shape to define the second part of the shape

I think this might be ok for our usecase, but that does mean that there can be only a single split service modification. I don't think that will be ok for the general spec.

[cmonagle]

Nit: would no_through_travel through_travel_prohibited work? it feels a bit clearer to me.

[juanborre]

If the flag is true, how do we know the two trips that result?

Is the detour defined in the trip_modifications->modification the second trip? Or the stops in the modificationdefine a part that is not served and then that is the part that does not have service?

I think we could clarify the behaviour.

[cmonagle]

If the flag is true, how do we know the two trips that result?

There are two undocumented requirements with this:

1. There must not be any replacement stops.
2. There must be stops remaining before the startStopSelector, and stops after the endStopSelector.

The resulting trips are (first stop) to startStop, and then endStop to (last stop).

[Edit]:

I think I understand your question now: how do we know which trip is associated with which shape. I don't think that's defined in the current implementation.

[juanborre]

It feels to me we are stuffing this here with a shoehorn. When we define a detour in TM, we actually define a segment that is not served and there is a lot of implicits or extra behaviour. There are also fields that don't mean anything in the current spec when the flag is set.

I'm not sure what's best though.

• Making another element similar to a Modification that can represent this better could be too much but could be cleaner.
• Also adding in an enum at the Modification level instead of a boolean to define what the modification is doing: detour, split, etc...