Regional data for clustering

This MR is a first tentative idea on how to implement regional clusterization within our ACTS algorithms. This MR allows the clusterization to run on detector elements within an RoI. This MR however does not implement any mechanism to update the cluster collection, so multiple passes will not be allowed yet. That will be addressed in other MRs.

The idea of the code is the following:

• The clusterization algorithms always process an RoI, may that be the entire detector (FS) or specific area(s) of the detector
• At offline level, a FS RoI will be produced by an upstream algorithm (still to be written, will add soon) that will then be passed to the downstream algorithms. Alternatively, the user can provide their own RoIs collection.
• The clusterization algorithms retrieve the hash ids of the detector elements from the regions of the detector they have to process.
• A check is performed to verify the hash id is not already in the cluster collection in SG (this functionality is currently incomplete since we only create the cluster collection in the code, so it will always be empty)
• We run on the list of hash ids we have to process, retrieve the corresponding RDO collection and run clusterization on it

I still have to check if there is a potential overhead w.r.t. default code, or if there are smarter solutions

@sutt @keli @ncalace @pagessin @lgagnon @tbold @jmasik what is your opinion on this? Would this structure be ok?

By defining a dummy RoI, the code now performs clusterization only on specific regions of the detector: