v3 marks the release of major improvements to the internal NLP preprocessing code. It has been written from the ground up to:
- be easier to debug
- be easier to deploy
- faster to execute
- and to more quickly add new fields-to-be-pseudonymized to the ETL process
v3 will live alongside v2 while we slowly migrate completely away from the old preprocessing code to the new.
v3.0 is identical to v2.10 for the end user. See release notes for that version for additional information regarding end-user-features.