py-fhir

FHIR Resources/Elements for Python. Code generated by py-fhir-codegen.

Introduction

This python package provides a python version of the Resources and Elements specified in the FHIR specification. Additionally, the package includes first steps towards database persistance (using SQLAlchemy) and a client.

How to use Resources and DataTypes

Running the following code creates a Patient and assigns values to different attributes. Finally, the object is serialized to JSON representation and the result is printed.

# Import the relevant classes from fhir.model
from fhir.model import Patient, HumanName, Identifier, CodeableConcept, Coding, uri

# Create a new Patient object. Resource attributes can be passed in the constructor. Note
# that 'id' is the logical id of the Resource and has no relation to the Medical Record
# Number (MRN)
p = Patient(id='patient1')

# Give the patient an MRN
identifier = Identifier(
    type=CodeableConcept(coding=[Coding(system="http://hl7.org/fhir/v2/0203", code="MR")]),
    system='urn:oid:1.2.36.146.595.217.0.1',
    value='123456789'
)

# Lists can be assigned to attributes.
# Alternatively p.identifier.append(identifier) could have been used.
p.identifier = [identifier]

# Native python values are automatically coerced to FHIR classes
p.active = True
print(type(p.active))
# output: <class 'fhir.model._boolean.boolean'>

name = HumanName()
name.use = 'official'
name.given = ['Melle', 'Sjoerd']
name.family = 'Sieswerda'
p.name = [name]

# Serialize to JSON and print the result. To serialize to XML use 'p.toXML()'.
print(p.dumps('json'))

This should yield the following output:

{
    "resourceType": "Patient",
    "id": "patient1",
    "identifier": [
        {
            "type": {
                "coding": [
                    {
                        "system": "http://hl7.org/fhir/v2/0203",
                        "code": "MR"
                    }
                ]
            },
            "system": "urn:oid:1.2.36.146.595.217.0.1",
            "value": "123456789"
        }
    ],
    "active": true,
    "name": [
        {
            "use": "official",
            "family": [
                "Sieswerda"
            ],
            "given": [
                "Melle",
                "Sjoerd"
            ]
        }
    ]
}

It is also possible to marshall an object from JSON (or XML) representation:

jsonstring = p.dumps('json')
p2 = Patient.loads(jsonstring, 'json')
p2.id == p.id
# output: True

Using the client

import fhir.model
import fhir.client
from fhir.model import Patient, HumanName

URL = fhir.client.SERVERS['spark']
client = fhir.client.Client(URL)
client.set_properties(fhir.model.Resource)

# The following is equivalent to
# bundle = client.query('Patient')
bundle = fhir.model.Patient.query()

# Print the first result we received
for p in bundle:
    print(p.dumps('json'))
    break

# Create a new patient on the server
p = Patient()
p.active = True

name = HumanName(
	use='official',
	given=['Melle'],
	family='Testpatient'
)
p.name = [name]

p.save()
print('The server assigned id "{}"'.format(p.id))

Using persistant storage

Cave: the current implementation should be considered a stub. Searching/querying, updating, etc. is not (yet) supported.

from fhir.persistance import FHIRStore
import fhir.model

# By default FHIRStore creates a sqlite database in the current directory.
# Still, be careful when using 'drop_all'!
store = FHIRStore(drop_all=True)

# Create a new Patient
p = fhir.model.Patient()
p.id = 'patient1'
p.active = True

name = fhir.model.HumanName(
    use='official',
    given=['Melle', 'Sjoerd'],
    family='Sieswerda'
)
p.name = [name]

# Store the patient
store.post(p)

# Retrieve the patient from the store and print the result.
print(store.get('patient1'))
# output: <fhir.model.patient.Patient object at 0x1096dd828>