txt2hpo 0.1.0

Phenotype extraction using Named Entity Recognition

txt2hpo

txt2hpo is a Python package for extracting HPO-encoded phenotypes from text, Human Phenotype Ontology (HPO). txt2hpo accounts for inflection, is able to parse complex multi-word phenotypes and comes with a built-in medical spellchecker.

Installation

Install from GitHub

git clone https://github.com/GeneDx/txt2hpo.git
cd txt2hpo
python setup.py install

Library usage

from txt2hpo.extract import hpo

hpos = hpo("patient with developmental delay and hypotonia")
print(hpos)

[{"hpid": ["HP:0001290"], "index": [37, 46], "matched": "hypotonia"}, 
 {"hpid": ["HP:0001263"], "index": [13, 32], "matched": "developmental delay"}]
    
    

txt2hpo will attempt to correct spelling errors by default, at the cost of slower processing speed. This feature can be turned off by setting the correct_spelling flag to False.

from txt2hpo.extract import hpo

hpos = hpo("patient with devlopental delay and hyptonia", correct_spelling=True)

[{"hpid": ["HP:0001290"], "index": [37, 46], "matched": "hypotonia"}, 
 {"hpid": ["HP:0001263"], "index": [13, 32], "matched": "developmental delay"}]