Ariadne is a Python toolkit for mapping source terminologies to standard concepts in the OHDSI Vocabulary.
- Verbatim term mapping: maps terms that (almost) exactly match standard concepts. Using term normalization techniques like lowercasing, punctuation removal, and word stemming.
- Embedding vector search: leverages embedding language models to find semantically similar standard concepts for source terms.
- Exact term mapping: using reasoning LLMs to find exact matches in the vocabulary.
- Evaluation: uses golden standard mappings to evaluate mapping performance.
Comprehensive documentation is available on the Ariadne Documentation Site.
- Python 3.12 or higher.
git clone [https://github.com/OHDSI/ariadne.git](https://github.com/OHDSI/ariadne.git)
cd ariadne
pip install -e .Ariadne uses environment variables for database connections and API keys.
-
Copy the example configuration:
cp .env.example .env
-
Edit the
.envfile to set your database connection details and API keys. Do not commit this file to version control. We recommend moving it to a secure location outside the project directory.
This project is licensed under the Apache 2.0 License. See the LICENSE file for details.
