Entity duplication in the de-anonymization process

[lamerentertainment]

Thanks again for this astonishing release. I use it nearly every day heavily! My issue:

During anonymization, multi-word PII entities such as “Hans Müller” are replaced with multiple identical placeholders (e.g., two instances of 1_person). In the sidebar view, these placeholders span across occurrences of the same entity, being uniformly represented as 1_person (Hans Müller, “Will replace [1_person] → Hans Müller”). However, this placeholder duplication during anonymization introduces an inconsistency at the de-anonymization stage: each occurrence of the duplicated placeholder is independently restored to Hans Müller. As a result, the process yields duplicate mentions of the same entity, despite their original equivalence ("Hans Müller Hans Müller").

See pull request for a possible solution (add appendices to the placeholders marking the words)

[colin-r-carter]

Thank you very much! #2 will resolve this, will review in the coming days and merge, thank you very much!

[colin-r-carter]

Fixed with d1f9c6d hope this works as well.