Add capability to recover data to Time Series Wizard

[pjcross]

Currently, when creating a time series model using the wizard, if the session id is lost or the workflow is interrupted after the parallel processing is complete but before the dendrogram is created, the entire process has to be restarted from the beginning even though the distance matrices and binned timeseries exist as files on the cluster. We need to add a capability to restart from this point in the process. The user could point to the directory where these files were stored and the process could resume (initially, moving the files to the server and generating a dendrogram). Eventually, we could offer a selection of alternative clustering/dimensionality reduction techniques and associated visualization that require distance matrices as a starting point.