Add a partitioning strategy based on ObjectID format as described in the TD

This strategy will use the timestamp of the ObjectID to get the time range of the collection's minId and maxId. It will then create partitions that are equal time-sized.

So say that we want to create two partitions and the collection min and max timestamps (from the min and max IDs) are tsMin and tsMax.

Then first create a dummy ObjectID where the time component is tsMin + (tsMax-tsMin)/2 .

Then find max id in the collection that is less than the above dummy ID. Use this to create the two partitions. This ensures that we use existing docs as the partition boundaries and so will work well with the other strategies we have (bucketAuto/random)