private static List<Tuple<int, List<TimeSpan>>> InterpolateSchedule(ConnexionzRoute connexionzRoute, List<GoogleStopSchedule> schedule)
{
var adherencePoints = connexionzRoute.Path
.Select((val, idx) => new { value = val, index = idx })
.Where(a => a.value.IsScheduleAdherancePoint)
.ToList();
// some invariants we can rely on:
// the first stop is an adherence point.
// the last stop is not listed as an adherence point, even though it has a schedule in google.
// therefore, we're going to add the last stop in manually so we can use the last schedule and interpolate.
adherencePoints.Add(new { value = connexionzRoute.Path.Last(), index = connexionzRoute.Path.Count });
var results = new List<Tuple<int, List<TimeSpan>>>();
for (int i = 0; i < adherencePoints.Count - 1; i++)
{
int sectionLength = adherencePoints[i + 1].index - adherencePoints[i].index;
var stopsInBetween = connexionzRoute.Path.GetRange(adherencePoints[i].index, sectionLength);
var differences = schedule[i].Times.Zip(schedule[i + 1].Times,
(startTime, endTime) => endTime.Subtract(startTime));
// we're simply going to add an even time span the schedules of consecutive stops.
// there doesn't appear to be a more reliable way of estimating than this.
var stepSizes = differences.Select(d => d.Ticks / sectionLength);
results.AddRange(stopsInBetween.Select(
(val, idx) => Tuple.Create(
val.PlatformId,
schedule[i].Times.Zip(stepSizes, (time, step) => RoundToNearestMinute(time.Add(TimeSpan.FromTicks(step * idx))))
.ToList())));
}
return results;
}
