Several cases in which eds.measurement fails to associate the mesure to its value (weight, height and bmi) - Githubissues

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

116 stars 30 forks source link

Several cases in which eds.measurement fails to associate the mesure to its value (weight, height and bmi) #277

Open JudithCo opened 8 months ago

JudithCo commented 8 months ago

Description

When the text contains several measures in a row with their respective values, without units, in sequence :

the pipe associates the first value with the last measure mentioned
and does not associate each type of measurement with its associated value.

How to reproduce the bug

import spacy

nlp = spacy.blank("eds")
nlp.add_pipe("eds.measurements")

text_1 = "Poids (Kg) Taille (m) IMC\n57,0 1,70 22"
text_2 = "poids / IMC : 57imc22 taille : 170"
text_3 = "poids / IMC : 57/22 taille : 170"

doc_1 = nlp(text_1)
doc_2 = nlp(text_2)
doc_3 = nlp(text_3)

print(doc_1.spans)
print(doc_2.spans)
print(doc_3.spans)

Your Environment

Operating System:
Python Version Used: 3.7.13
spaCy Version Used: 3.7.2
EDS-NLP Version Used: 0.10.7
Environment Information:

OlivierHassanaly commented 3 months ago

another issue linked to this topic ; take text_4 = "Madame xxx yyy, âgée de xx ans, née le 06/12/1996, 97 kg, 175 cm"