Python library for interacting with HL7 FHIR servers and resources. Resource validation and parsing powered by
pydantic and the fhir.resources library.
Provides a simple interface for synchronous and asynchronous CRUD operations for resources and bundles,
as well as resource transfer between servers.
Datascience features include flattening of resources and bundles into tabular format (pandas dataframes) and plotting
methods for resources and bundles can optionally be included with the ds
extra.
Check out the documentation for more information and a detailed user guide.
- Create, Read, Update, Delete resources using a FHIR server's REST API
- Transfer resources between servers while maintaining referential integrity using server-given IDs
- Bundle creation, validation and data management on a FHIR server via the REST API
- Supports Hapi, Blaze and IBM FHIR servers
- CSV serialization of query results
- Synthetic data generation and
Install the package using pip:
pip install fhir-kindling --user
Fhir kindling can be used with the following extras:
ds
for data science related features, such as flattening of resources into a tabular format
pip install fhir-kindling[ds] --user
from fhir_kindling import FhirServer
# Connect with basic auth
basic_auth_server = FhirServer("https://fhir.server/fhir", username="admin", password="admin")
# Connect with static token
token_server = FhirServer("https://fhir.server/fhir", token="your_token")
# Connect using oauth2/oidc
oidc_server = FhirServer("https://fhir.server/fhir", client_id="client_id", client_secret="secret",
oidc_provider_url="url")
# Print the server's capability statement
print(basic_auth_server.capabilities)
from fhir_kindling import FhirServer
from fhir.resources.patient import Patient
# Connect using oauth2/oidc
oidc_server = FhirServer("https://fhir.server/fhir", client_id="client_id", client_secret="secret",
oidc_provider_url="url")
# query all patients on the server
query = oidc_server.query(Patient, output_format="json").all()
print(query.response)
# Query resources based on name of resource
query = oidc_server.query("Patient", output_format="json").all()
print(query.response)
Filtering the targeted resource is done using the where
method on the query object. The filter is created by defining
the target field, the comparison operator and the value to compare.
from fhir_kindling import FhirServer
server = FhirServer(api_address="https://fhir.server/fhir")
query = server.query("Patient").where(field="birthDate", operator="gt", value="1980").all()
Resources that reference or are referenced by resources targeted by the query can be included in the response using
the include
method on the query object.
# server initialization omitted
# get the patients along with the queried conditions
query_patient_condition = server.query("Condition").include(resource="Condition", reference_param="subject").all()
# get the conditions for a patient
query_patient_condition = server.query("Patient")
query_patient_condition = query_patient_condition.include(resource="Condition", reference_param="subject", reverse=True)
response = query_patient_condition.all()
If you know the id and resource type of the resource you want to query, you can use the get
method for a single
reference
for a list of references use get_many
. The passed references should follow the format of <resource_type>/<id>
.
# server initialization omitted
patient = server.get("Patient/123")
patients = server.get_many(["Patient/123", "Patient/456"])
Resources can be added to the server using the add
method on the server object. Lists of resources can be added using
'add_all'.
from fhir_kindling import FhirServer
from fhir.resources.patient import Patient
# Connect to the server
server = FhirServer(api_address="https://fhir.server/fhir")
# add a single resource
patient = Patient(name=[{"family": "Smith", "given": ["John"]}])
response = server.add(patient)
# add multiple resources
patients = [Patient(name=[{"family": f"Smith_{i}", "given": ["John"]}]) for i in range(10)]
response = server.add_all(patients)
Resources can be deleted from the server using the delete
method on the server object, it takes as input either
references to the resources or the resources itself.
Similarly the update
method can be used to update the resources on the server, by passing a list of updated resources.
from fhir_kindling import FhirServer
from fhir.resources.patient import Patient
# Connect to the server
server = FhirServer(api_address="https://fhir.server/fhir")
# add some patients
patients = [Patient(name=[{"family": f"Smith_{i}", "given": ["John"]}]) for i in range(10)]
response = server.add_all(patients)
# change the name of the patients
for patient in response.resources:
patient.name[0].given[0] = "Jane"
# update the patients on the server
updated_patients = server.update(resources=response.resources)
# delete based on reference
server.delete(references=response.references[:5])
# delete based on resources
server.delete(resources=response.resources[5:])
Transferring resources between servers is done using the transfer
method on the server object. Using this method
server assigned ids are used for transfer and referential integrity is maintained.
This method will also attempt to get all the resources that are referenced by the resources being transferred from the
origin
server and transfer them to the destination server as well.
from fhir_kindling import FhirServer
# initialize the two servers
server_1 = FhirServer(api_address="https://fhir.server/fhir")
server_2 = FhirServer(api_address="https://fhir.server/fhir")
# query some resources from server 1
conditions = server_1.query("Condition").limit(10)
# transfer the resources to server 2
response = server_1.transfer(server_2, resources=conditions.resources)
This library performs request at least 1.5 times faster than other popular python FHIR libraries. See Benchmarks for a more detailed description of the benchmarks.
Contributions are very welcome and greatly appreciated! If you want to contribute to this project, please fork the repository and make changes as you'd like. Pull requests are warmly welcome and credit will always be given.
To set up your environment to develop this package make sure you have poetry installed and run the following commands:
Install the dependencies:
poetry install --with dev --all-extras
Install pre-commit hooks:
poetry run pre-commit install
To run the full test suit you need access to two FHIR servers (the second one is used for transfer tests).
You can spin up two servers (one HAPI and one Blaze FHIR) using the compose file in the testing
directory.
cd testing
docker compose up
The servers will be available at http://localhost:9090/fhir
and http://localhost:9091/fhir
respectively.
And the test should be configured to use them via the environment variables FHIR_API_URL
and TRANSFER_SERVER_URL
respectively.
Run the tests:
poetry run pytest
This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.