Docdeid

Added
* some internal speedups for `SingleTokenLooupAnnotator`, `MultiTokenLookupAnnotator` and `LookupTrie`
* caching for sorting annotations, which helps with speed
* the `pre_match_words` attribute for `RegexpAnnotator`
* the option to provide a `LookupTrie` to a `MultiTokenAnnotator` directly
* a method for getting all words or, for looking up tokens with specific text values in a `TokenList`, with options for `matching_pipeline`
* automated build/publish on merge to main

Changed
* sorting `Annotation` and `AnnotationSet` now requires sort key to be provided as a `tuple`, and callbacks as a `frozendict`
* renamed `docdeid.tokenize` to `docdeid.tokenizer`
* renamed `docdeid.process.doc` to `docdeid.process.doc_processor`
* renamed `docdeid.process.annotation_set` to `docdeid.process.annotation_processor`
* `Annotation` and `Token` now only include `int`/`str` fields when serializing
* formatting and linting settings
* moved the logic for linking tokens to `TokenList` rather than `Tokenizer`
* use `casefold()` instead of `lower()` for lowercasing

Fixed
* a bug with overlapping annotations in `MultiTokenLookupAnnotator`

Removed
* automated coverage reporting

Added
* `RegexpAannotator` accepts regexp strings in addition to compiled regexp patterns

Changed
* consisent use of `args` and `kwargs` in `Annotator` class tree
* `RegexpAnnotator` now offers function to validate matches, implementable by subclassing

Changed
* made the `priority` attribute of an `Annotation` non-Optional
* multi token lookup now sets the `start_token` and `end_token` fields of an `Annotation`

Fixed
* a bug with determnistic sort, when `Optional` fields were set

Added
* an additional `priority` attribute for `Annotation`, giving an extra option for sorting

Changed
* upgraded dependencies

Changed
* upgraded dependencies, including a `markdown-it-py` which had a vulnerability