[Discovery] Avoid Duplicate Discoveries

[pr0me]

Problem description:
For patterns that have multiple instances that all show to the same discovery rule, there will be multiple identical entries in the discovery file. This is expected but confusing, I think.

Having multiple patterns with the same discovery rule is the outcome of a past discussion: In such cases, the instances and their code snippets act a examples of different manifestations of a testability pattern but have the same underlying problem on a code/CPG level.

Suggested solution:
During pattern discovery, collect individual discoveries as a set instead of a list, eliminating exact duplicates in the discovery report.