Split construction and statistics

There is no consistency in which functions provide statistics and which add information to the graph database. group and grouping_overview both produce similar output; add_functions, add_pav and add_variants have separate overview functions in function_overview and variation_overview while add_annotations and add_phenotypes have no separate overview function at all. My suggestion is to split all statistics from construction altogether. While increasing the number of commands this provides the following benefits:

• Subcommands have a more singular purpose and are easier to read, refactor and run faster.
• Construction methods from two versions can be compared with the overview function of the newer version in the CI pipeline.