Better UX for CoclusteringReport

[popescu-v]

Some queries in the CoclusteringReport class are difficult to program. For example:

• Obtain the parts of a cell for a specific dimension by name
• Given a cluster in a dimension, obtain all cells in another one that have that cluster as a marginal

The idea is to extend the classes for the coclustering results, possibly by creating indexes to ease these kind of queries. However, this must be done carefully as the memory consumption can explose.

Implementation Ideas

• A "lever" method index that would create the indexes on request