v2.0.15

• Adds NHS recogniser if the entity is specified
• Adds NINO recogniser if the entity is specified
• Corrects type hints on (you can't have None as the default value for a list type)
• Adds extra boilerplate because of the above correction
• Adds tests for NHS and NINO functionality

v2.0.14

• Small type, test and README fixes

v2.0.13

• Use native Pandas type "string" instead of Python base "str" for DataFrame type conversion

v2.0.12

• Adds **kwargs argument to IDScrub.dataframe(), which allows other keyword arguments in IDScrub() to be modified when using IDScrub.dataframe()
• Fixes ID column bug in IDScrub.dataframe() when no id_col argument is passed

v2.0.11

• Add custom_methods.ipynb to give further examples on customisation and contribution
• Add extra information and advice to README

v2.0.1

• Handle whitespace-only strings if passed to IDScrub.presidio_entities by stripping whitespace

v2.0.0

• This release contains non-backward compatible (breaking) changes. This is required for improving the package.
• scrub.scrub() now takes a pipeline dictionary that defines the scrub methods.
• Each method now outputs a list of IDEnt (identified entity) objects instead of the scrubbed text. These objects contain all of the information required to find and replace the entity identified.
• The list is then passed to a resolve_overlaps , which selects the entity with a higher priority score if multiple entities are identified in the same text e.g. john@madeupemail.com is both an email and a handle (@madeupemail), but we can score the email higher so an email is retracted.
• The final de-duplicated list is then passed to scrub_texts , which removes the text and replaces it.
• Add an exclude argument to exclude certain strings from being scrubbed.
• Improve error handling.
• This is a big improvement because every method sees the same text, then the text is scrubbed in one step once all of the conflicts have been reconciled.

v1.1.1

• Update README to emphasise development
• Pin Pandas <3.0 until changes resolved

v.1.1.0

• Allows users to specify which SpaCy entities to scrub
• Changes all NER scrubbing methods to <method>_entities e.g. IDScrub.spacy_entities()
• Adds a uk addresses scrubbing method