Symptomspot is a simple symptom string extractor based on the Symptom Ontology developed at the Institute for Genome Sciences (IGS) at the University of Maryland. The symptom strings are extracted using SpaCy's custom entity matching.
Install requirements, download spacy model and download git repository. The lightest model should suffice for this purpose.
pip install -r requirements.txt
python -m spacy download en
pip install git+https://github.com/mlehl88/symptomspot.git
Use the SymptomExtractor to extract unique symtom strings from a text or URL:
>>> from symptomspot.extractor import SymptomExtractor
>>> extractor = SymptomExtractor()
>>> extractor.extract(text='I have a back pain')
['back pain', 'pain']
>>> extractor.extract(url='www.nhs.uk/Conditions/Sleep-paralysis/Pages/Symptoms.aspx')
['inability to move', 'hallucinations', 'anxiety']
Usage with input text:
python symptomspot/app.py
curl -X POST 'localhost:5000/symptomspot/text' -d "text=The Trump anxiety."
Usage with input URL:
curl -X POST 'localhost:5000/symptomspot/url' -d "url=www.nhs.uk/Conditions/Sleep-paralysis/Pages/Symptoms.aspx"
Run the evaluation script for a lightweight evaluation of 8 labelled documents:
python eval/eval.py