GitLab

It is no longer necessary, because adding the Analyses is actually not symmetrical:
collections of the "native" pipeline know about the newly added objects,
but these objects still refer to their "foreign" hive_pipeline, which makes it possible
to derive their original AnalysisStats dynamically.
Keep this in mind until adding these objects is no longer necessary (desired redesign plan).

Moved the caching from Hive.pm into the specific storable classes themselves (via Cacheable.pm) to better conform to Active Record pattern

Graph.pm uses cache collections (no change in interface) ; A.control_rules_collection and A.dataflow_rules_collection lazy-load from collections

a big switch from using cluster labels in _allocate_to_subgraph() to setting {'_funnel_dfr'} of each graph's node and edge

simplified table_node_name generation and merged the two instances of DuplicateTables-based decision to a single point