vmenger
Improve PATIENT/PERSOON processing and more
This is the result of my first encounter with this codebase (Docdeid and Deduce), the second part (Deduce). My goal was to understand the inner workings of it and then make sure that capitalized street names are pseudonymized (all-caps or titlecased, and covering also the special case of the "IJ" digraph in Dutch). When at it, I noticed unexpected behaviour for patient names v. other person names and improved that as well.
This depends on changes in Docdeid, filed as https://github.com/vmenger/docdeid/pull/20.
To use that Docdeid version, I checked out the two repos side by side and added the following configuration in Deduce's pyproject.toml:
[tool.poetry.dependencies]
docdeid = {path = "../docdeid", develop = true}
FWIW, I also see a diff in my local (non-committed) version of base_config.json affecting "initiaal_patient" mentions but it's been 4 months since I intensively worked on this codebase so I don't remember anymore whether it's useful or even necessary anymore. But if some tests fail without it for you, let me know, this may well be the reason.
