aphp / edsnlp

Modular, fast NLP framework, compatible with Pytorch and spaCy, offering tailored support for French clinical notes.
https://aphp.github.io/edsnlp/
BSD 3-Clause "New" or "Revised" License
112 stars 29 forks source link

👓 Feedbacks from EDS-TeVa study #157

Closed Aremaki closed 1 year ago

Aremaki commented 1 year ago

Various feedbacks that come from the EDS-TeVa use case. (Detection of pulmonary embolism in discharge summaries).

Description

  1. History pipeline : Add the option to consider only the closest dates in the sentence (dates inside the boundaries and if there is not, it takes the closest date in the entire sentence).
  2. Hypothesis pipeline : Remove too generic patterns (#154).
  3. Negation pipeline : It takes into account following past participates and preceding infinitives.
  4. Negation pipeline & Hypothesis pipeline : Introduce new patterns and remove unnecessary patterns.
  5. Date pipeline : Add a pattern for preceding relative dates (ex: l'embolie qui est survenue à 10 jours).
  6. EDS tokenizer: It now tokenizes "rechereche d'" as ["recherche", "d'"], instead of ["recherche", "d", "'"].
  7. Typos: Fix small typos in the documentation and in the docstring.

Checklist