[Monitor] Ignore alternative, excursion and approach roles in segment computation

[StC-OSM]

I propose to temporarily ignore all members of relations and sub-relations that have roles such as alternative, excursion and approach when computing the number of segments in a route and displaying interruptions in said routes.

It is understood that it would be best to deal with such members in a more considerate way, but currently the way these members are managed hampers our ability to monitor routes. When displaying the list of routes in a group, the number of segments goes from 1 to 20 or more depending on the number of alternative/approach/Excursion segments; it makes it difficult to detect variations in these numbers. Ignoring these segments would allow us to maintain a list where all segment numbers are 1, and changes would be immediately noticed.

[StephaneChatty]

@pyrog asks me in private to provide examples. The fr-nwn group provides us with a perfect one (or two, depending): one of the following routes has been broken in recent days, and the other is perfect in terms of continuity. Can you tell which is which?