FHIR Python Analysis and Conversion Kit (FHIRPACK)

FHIRPACK (FHIR Python Analysis Conversion Kit) is a general purpose library that simplifies the access, analysis and representation of FHIR and EHR data using PANDAS, an ETL philosophy and a functional syntax. FHIRPACK was designed and developed at the Institute for Artificial Intelligence in Medicine (IKIM) and the Database Systems Research Group of the University of Heidelberg (HDDBS).

About FHIRPACK

The FHIR standard is a promising framework for interacting with healthcare data. However, expressive tools for efficient FHIR server interaction are few and are either too low level, and powerful, or too abstract, and lack features. FHIRPACK provides an easy-to-use and intuitive API that enables effortless access to FHIR data. We strive for a balance between flexibility, usability and feature richness in order to make interacting with FHIR data less painful.

Usage

Installation

You can easily install the latest FHIRPACK release from PyPI or the most current version by cloning this repository. As usual with Python, we strongly recommend using virtual environments such as pipenv, venv or conda.

Requirements

FHIRPACK requires Python 3.9 or greater as well as libmagic to work without problems. In case you're using an older Python version, you can use asdf, conda or pyenv to have several Python versions coexist on your system. libmagic can be installed with apt-get install libmagic-dev on Ubuntu or brew install libmagic on MacOS.

Using the Latest Release from PyPI

pip install fhirpack

alternatively use pipenv:

pipenv --python 3.9 install fhirpack

Using the Latest Version a from Local Clone

git clone https://gitlab.com/fhirpack/main.git
mv main fhirpack
pip install -e fhirpack 

alternatively use pipenv:

git clone https://gitlab.com/fhirpack/main.git
mv main fhirpack
pipenv install -e fhirpack

Upgrading your Version of FHIRPACK

You can use pip install --upgrade fhirpack or if you're using pipenv pipenv update fhirpack to upgrade to the latest release of FHIRPACK once you've already installed it.

Configuration

To set up a server configuration, create an .env file in the directory where you'll work with FHIRPACK and specify the settings as can be seen in .env.example. Alternatively, copy, rename and modify .env.example according to your needs.

:warning: By Default, FHIRPACK connects to the public R4 FHIR test server from HAPIFHIR http://hapi.fhir.org/baseR4. We recommend using this setup to get familiar with the library.

Simple Examples

FHIRPACK is based on the ETL paradigm, and as all functions available to you can be classified as Extractors, Transformers or Loaders. In the following examples we quickly show how this works, but remember to have a look at our examplary Jupyter Notebooks and the API reference

1. Get All Conditions for a Patient

In this example we extract all the conditions for a patient with the ID: 43fb1577-3455-41cf-9a07-c45aa5c0219e from the public FHIR-server with the Base-URL: http://hapi.fhir.org/baseR4.

# import FHIRPACK
import fhirpack as fp 

# instantiate a connected PACK for the specified FHIR API base 
pack = fp.PACK("http://hapi.fhir.org/baseR5") 

# retrieve a list of Patients of length one by ID
patient = pack.getPatients(["43fb1577-3455-41cf-9a07-c45aa5c0219e"]) 

# retrieve all respective conditions for said patients
condition = patient.getConditions().explode() 

# gather and display the FHIR elements with the specified paths from the conditions
condition.gatherSimplePaths(["id", "code.coding.code", "code.coding.display", "onsetDateTime" ]) 

2. Get All Patients with Sepsis

# import FHIRPACK
import fhirpack as fp 

# instantiate a connected PACK for the specified FHIR API base 
pack = fp.PACK("http://hapi.fhir.org/baseR5") 

# retrieve  all conditions containing the term sepsis
conditions = pack.getConditions(searchParams={"_content": "sepsis"}) 

# get the respective patients
patients = conditions.getPatients().explode() 

# display the specified FHIR elements of the patients
patients.gatherSimplePaths([
	"name.given", 
	"name.family",
	"telecom.value", 
	"address.country",
	"address.city",
	"birthDate"
]) 

:information_source: For more examples and a deep-dive into FHIRPACK, please take a look at the example jupyter notebooks.

CLI

FHIRPACK also provides a CLI for easy and quick data exploration.

The CLI can be invoked by using python -m fhirpack.cli or fp once FHIRPACK has been installed.

> fp --help                                
Usage: fp [OPTIONS]

  The pack for your FHIR server.

Options:
  --version               Show the version and exit.
  -s, --source TEXT       URL of the FHIR server or path to json files.
  -e, --environment TEXT  Path to Dotenv file containing configurations.
  -d, --destination TEXT  Location of the output.
  -o, --operation TEXT    Operations to retrieve data.
  -p, --params TEXT       Provide additional search parameters.
  -v, --verbose           Print results in verbose format.
  -h, --help              Show this message and exit.

CLI usage is analogous to the general fhirpack dataflow.

:warning: Operations spanning mutliple spaces have to be quoted.

Bugs and Feature Requests

Please report any bugs you find on our main GitLab Tracker or here as well. If you want to contribute a fix or feature, you're welcomed to create a pull request from your fork/branch or create a merge request on our main GitLab repository according to our contribution guidelines.

Attribution and License

If you're using FHIRPACK, feel free to reach out over Twitter, Slack or e-mail

FHIRPACK is released under the MIT, see LICENSE.txt for more information.