Skip to content

Latest commit

 

History

History
215 lines (128 loc) · 6.91 KB

CHANGELOG.md

File metadata and controls

215 lines (128 loc) · 6.91 KB
Raw

Changelog

All notable changes to this project will be documented in this file. Please add new entries at the top. Use one of the following headings: Added, Changed, Deprecated, Removed, Fixed, Security.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

❗ = Breaking change

0.9.0 (2024-07-10)

Added

  • Mantra GSC corpus for evaluation
  • Loading and exporting InfoExtractionDataset as dictionaries or JSON files
  • Metric support for multi-class qualifiers
  • In the RuleBasedEntityMatcher, option to add terms as a dict (in addition to str, list and Term)
  • In the RuleBasedEntityMatcher, option to add terms from dict (add_terms_from_dict), json (add_terms_from_json) or csv (add_terms_from_csv)
  • In the Term class, an option to override arguments that were not set

Changed

  • Moved regression test cases to data directory in more open format, so they are re-usable
  • Made the default field for Qualifier optional
  • InfoExtractionDataset and InfoExtractionMetrics use Qualifier objects for qualifiers rather than dict
  • ❗ InfoExtractionDataset and InfoExtractionMetrics no longer track or use qualifier defaults
  • Made qualifiers optional for metrics in Annotation
  • Added a normalize method to Normalizer, so it can be used/tested directly
  • The logic for determining whether the RuleBasedEntityMatcher should internally use the phrase matcher or the matcher is simplified

Deprecated

  • ❗ The create_concept_dict method, which is now replaced by add_terms_from_csv in RuleBasedEntityMatcher
  • ❗ In the RuleBasedEntityMatcher, the load_concepts method, which is now replaced by add_terms_from_dict and add_terms_from_json

0.8.1 (2024-06-27)

Added

  • Docstrings on all modules, classes, methods and functions

Changed

  • In InformationExtractionDataset, renamed span_counts, label_counts and qualifier_counts to span_freqs, label_freqs and qualifier_freqs respectively.
  • The clinlp_component utility now returns the class itself, rather than a helper function for making it
  • Changed order of direction and qualifier arguments of ContextRule
  • Simplified default settings for clinlp components and Term class
  • Normalizer uses casefold rather than lower for normalizing text
  • Parameterized spans_key for ie components

0.8.0 (2024-06-03)

Changed

  • ❗ Renamed the clinlp_entity_matcher to clinlp_rule_based_entity_matcher
  • ❗ clinlp now stores entities in doc.spans['ents'] rather than doc.ents, allowing for overlap
    • ❗ Overlap in entities found by the entity matcher is no longer resolved by default (replacing old behavior). To remove overlap, pass resolve_overlap=True.
  • Refactored tests to use pytest best practices
  • Changed clinlp_autocomponent to clinlp_component, which automatically registers your component with spaCy
  • Codebase and linting improvements
  • Renamed the other_threshold config to family_threshold in the clinlp_experiencer_transformer component

Fixed

  • The clinlp_rule_based_entity_matcher no longer overwrites entities detected by other components (but appends them)

0.7.0 (2024-05-16)

Added

  • Integrated the clin_nlp_metrics package in this repository, specifically in clinlp.metrics.ie
  • Support for non-binary qualifier in the Context Algorithm (e.g. 'Change', with values Decreasing, Stable and Increasing)
  • Support for bidirectional qualifier patterns

Changed

  • ❗ Moved all components related to information extraction to clinlp.ie. Please update imports accordingly (e.g. from clinlp.ie import Term)
  • ❗ Updated the framework for qualifiers, to now have three qualifier classes: Presence, Temporality and Experiencer. For more details, see docs

0.6.6 (2024-04-24)

Added

  • Support for Python 3.12

0.6.5 (2024-02-13)

Added

  • A component for transformer-based detection of Experiencer qualifiers (Patient/Other) (clinlp_experiencer_transformer)

0.6.4 (2024-02-13)

Added

  • A way to use a csv file as input for a concept list, using create_concept_dict

0.6.3 (2024-01-18)

Fixed

  • Fix a bug with termination trigger directly next to context trigger

0.6.2 (2023-10-06)

Fixed

  • Replaced call to importlib.resources.path which is deprecated from python 3.11 on

0.6.1 (2023-10-06)

Fixed

  • A bug with adjacent entities, which were accidentally marked as overlapping

0.6.0 (2023-10-03)

Changed

  • Qualifier detectors now add all default qualifiers (e.g. 'Affirmed', for Negation)
  • Use titlecase for qualifier values

0.5.3 (2023-10-02)

Fixed

  • A bug with importlib causing an AttributeError on importing clinlp

0.5.2 (2023-09-27)

Fixed

  • Removed accidental print statement

0.5.1 (2023-09-27)

Fixed

  • A bug with overlapping entities

0.5.0 (2023-08-17)

Added

  • A custom component for entity recognition, with options for proximity, fuzzy and pseudo matching

0.4.0 (2023-08-05)

Added

  • Definition for qualifiers (negation, plausibility, temporality, experiencer)

Changed

  • Updated rules for context algorithm to be consistent with definitions
  • Added some rules for context algorithm
  • Refactored Qualifier class from enum to a separate class, that accommodates other fields (like prob)
  • Use entity._.qualifiers to obtain Qualifier classes, entity._.qualifier_str for strings, and entity._.qualifier_dict for dicts

Fixed

  • Ambiguity of dd for context rules (can mean differential diagnosis, and daily dosage)
  • Importing clinlp caused a bug when extras were missing

0.3.1 (2023-06-30)

Removed

  • Support for python 3.9

0.3.0 (2023-06-30)

Added

  • Remove a default spaCy abbreviation (ts.)
  • Option for max scope on qualifier rules, limiting the number of tokens it applies to
  • A transformer based pipeline for negation detection (clinlp_negation_transformer)
  • A base class QualifierDetector for qualifier detection

Fixed

  • Issue where entity and context trigger were overlapping (e.g. geen eetlust)
  • Some tests that were not auto-discovered by pytest due to naming

Changed

  • Refactored context algorithm to allow adding new qualifier detectors
  • The @clinlp_autocomponent wrapper as a utility function, which makes creating components with inheritance and arbitrary config a bit easier
  • Made default configs a bit simpler and DRY
  • Move qualifier adding for context algorithm to base class

0.2.0 (2023-06-07)

Added

  • Version info to model meta (warns if installed clinlp version does not match model version)
  • A component for normalizing

0.1.1 (2023-05-23)

Fixed

  • Bug with resource loading

0.1.0 (2023-05-23)

Changed

  • Initial release

0.0.1 (2023-12-16)

Changed

  • Placeholder release