UMAP and HDBSCAN for cytoflow

[CaRniFeXeR]

I added UMAP to CytoFlow. This Pull Request contains three major adaptations:

• BulkConditionOp
  • Allows adding multiple conditions to an experiment at once. Source is either a pandas.Dataframe or an csv File. Added conditions can be combined into one column. This operation is intended to be used when manual gating ground truth is available.
• BaseDimensionalityReductionOp
  • Represents the (abstract) base class for dimensionality reduction now used by PCAOp and UMAPOp. I optionally added the possibility to rescale data before fitted and transformed with an dimensionality reduction algorithm. (Should be beneficial for both PCA and UMAP)
• UMAPOp
  • The Operation using the umap python package. Implementation is very similar to that of PCAOp.

I added a couple of unittests for the adaptations. As well as added an example of using UMAP to the Machine Learning Notebook. I did not managed to build the docs, so this is still missing.