Optimize Similarity using in-memory mutation

[JohT]

🎉 Reports

• Add summary for all community detection results: After each projection, all detected communities will be written to a new report file to be able to compare them and have them all in one CSV.
• Add query for same package names across Artifacts: Helps to find duplicate package names across artifacts that might lead to confusion and allow unintended access to package-protected classes, methods or members.

⚙️ Optimization

• Optimize Similarity using in-memory mutate: The results are then streamed or written from the in-memory projection without recalculating them.
• Resolve duplicate class paths last: Assure that e.g. Class nodes with the same full qualified name are identified and marked as such as a last step after detecting dependencies between Packages and Artifacts.
• Refine type classification and projection filter: Not all types should be taken into account for queries and graph algorithms. Build-in Java Types, Primitives, Duplicates and external Types should be filtered out to get more precise results.

🛠 Fixes

• Replace deprecated graph data science procedures: This assures that future updates won't break because of procedures that were already deprecated.