Allow for filling small gaps instead of restarting the batch for every gap

[jds485]

In #218, using the fill_batch argument checks for any gaps in the timeseries, and each gap is used to end one batch and start a new batch. There is a check if the gap length is less than the sequence length, but nothing is currently done when that happens. We could add a function argument that specifies the maximum gap length that should be filled, and create a function that edits to the array to fill in the missing dates. This functionality is not needed for current work, so it's being filed as an enhancement,