Drug Named Entity Recognition

2.0.9 verified Fri May 01 auth: no python

Drug Named Entity Recognition (NER) library to identify and resolve drug names in text, with entity linking to RxNorm and other drug databases. The current version is 2.0.9, supporting Python >=3.6. It provides pre-trained models for drug name extraction and normalization. The library is actively maintained with occasional releases.

pip install drug-named-entity-recognition

Common errors

error ModuleNotFoundError: No module named 'scispacy' ↓

cause Missing scispacy dependency.

fix

Run: pip install scispacy

error OSError: [E050] Can't find model 'en_core_sci_sm'. ↓

cause The spaCy model required by scispacy is not installed or not found.

fix

Run: pip install scispacy && python -m spacy download en_core_sci_sm

error ValueError: DrugNER model not loaded. Try again later. ↓

cause The model download failed or is corrupted.

fix

Delete the cache directory (~/.drug_named_entity_recognition) and run again.

Warnings

gotcha The first run downloads a large model (~2GB). Ensure sufficient disk space and a stable internet connection. The model is cached afterwards. ↓

fix Pre-download the model using: `pip install drug-named-entity-recognition && python -c \"from drug_named_entity_recognition import DrugNER; DrugNER()\"`

breaking In version 2.0, the API changed from `DrugNER().get_entities(text)` to `DrugNER().find_entities(text)`. Old code will break. ↓

fix Use `find_entities()` instead of `get_entities()`.

deprecated The `DrugNER` constructor no longer accepts a `model_size` argument in 2.0.0+ (was `'small'`, `'medium'`, `'large'`). ↓

fix Remove the `model_size` argument. The model is now fixed.

Imports

DrugNER

from drug_named_entity_recognition import DrugNER

Correct import path for the main class

Quickstart

Initialize DrugNER and extract drug entities from a text string.

from drug_named_entity_recognition import DrugNER

drug_ner = DrugNER()
text = "Patient was given 500mg of Paracetamol and 10mg of Diazepam."
entities = drug_ner.find_entities(text)
print(entities)
# Example output: [{'name': 'paracetamol', 'rxnorm': '161', 'start': 24, 'end': 35, ...}, ...]