Improved support for pad in TimeSeries.read and TimeSeries.find
This PR improves the support for padding gaps in data read using .find() or .read(). The main changes are:
- borrowed logic from the nds2 reader to ensure that all series are returned with the requested boundaries, padding the start or end as required
- allow gaps in datafind queries if
padis given
This should allow users to read, for example, months of trend data directly from GWF even if gaps are present.