Edsnlp

Added

- Now possible to provide regex flags when using the RegexMatcher
- New `ContextualMatcher` pipe, aiming at replacing the `AdvancedRegex` pipe.
- New `as_ents` parameter for `eds.dates`, to save detected dates as entities

Changed

- Faster `eds.sentences` pipeline component with Cython
- Bump version of Pydantic in `requirements.txt` to 1.8.2 to handle an incompatibility with the ContextualMatcher
- Optimise space requirements by using `.csv.gz` compression for verbs

Fixed

- `eds.sentences` behaviour with dot-delimited dates (eg `02.07.2022`, which counted as three sentences)

Added

- Complete revamp of the measurements detection pipeline, with better parsing and more exhaustive matching
- Add new functionality to the method `Span._.date.to_datetime()` to return a result infered from context for those cases with missing information.
- Force a batch size of 2000 when distributing a pipeline with Spark
- New patterns to pipeline `eds.dates` to identify cases where only the month is mentioned
- New `eds.terminology` component for generic terminology matching, using the `kb_id_` attribute to store fine-grained entity label
- New `eds.cim10` terminology matching pipeline
- New `eds.drugs` terminology pipeline that maps brand names and active ingredients to a unique [ATC](https://en.wikipedia.org/wiki/Anatomical_Therapeutic_Chemical_Classification_System) code

Added

- Support for strings in the example utility
- [TNM](https://en.wikipedia.org/wiki/TNM_staging_system) detection and normalisation with the `eds.TNM` pipeline
- Support for arbitrary callback for Pandas multiprocessing, with the `callback` argument

Added

- Support for chained attributes in the `processing` pipelines
- Colour utility with the category20 colour palette

Fixed

- Correct a REGEX on the date detector (both `nov` and `nov.` are now detected, as all other months)

Fixed

- Updated Numpy requirements to be compatible with the `EDSPhraseMatcher`

Added

- New `eds` language to better fit French clinical documents and improve speed
- Testing for markdown codeblocks to make sure the documentation is actually executable

Changed

- Complete revamp of the date detection pipeline, with better parsing and more exhaustive matching
- Reimplementation of the EDSPhraseMatcher in Cython, leading to a x15 speed increase