- Added Lab.dataframe() method to save going to the notebook
- Added updateResults() at Lab level, removed now-unnecessary
overriding at ClusterLab level
- Optimised updateResults() to grab all completed pending results
in a single network transaction
- Added global wait() to ClusterLab, coding around the
ipyparallel.Client code's wait() method (which can't be used
directly -- see notes in the method)