Introduction

PteRedactyl is a redaction package for personally identifiable information (PII) in text, that combines NER models with regex matching to identify and mask sensitive information. Custom transformers-based NER models can be swapped in,

This guide will cover usage of PteRedactyl's main functions: create_analyser, analyse, anonymise and anonymise_df:

• create_analyser(): Creates a presidio AnalyserEngine that can be reused across high-level functions
• analyse(): Analyses a string for PII, returning a list of recognised NER or regex-based results with scores
• anonymise(): Analyses and anonymises text, replacing PII with placeholders or hide-in-plain-sight (HIPS) replacements
• anonymise_df(): As above, but by looping over a given column (or columns) in a DataFrame