If the existing storage backends do not met your needs then you can write you own backend. You may want to store data using a popular medical record system (OpenMRS, FreeMED) or NoSQL database (CouchDB, MongoDB). Since applications interact with the storages though the client API they will transparently work with your new storage.
Note
This is an advanced use case and not necessary for most users or application developers.
The data passed to the backend will match the data described in the patient and provider data model sections. Backend methods which return data should return dictionaries matching this format as well.
Additional backends should extend from HealthcareStorage which is defined below:
Patient data should be fetched for the given id and returned as a dictionary. If the patient does not exist this method should return None.
source is an optional paramter. If given then the id should be interpreted as the source_id and the source as the source_name to find the patient using the association created by py:meth:HealthcareStorage.link_patient. If the patient cannot be found for this association it should also return None.
A patient record should be created for given set of data given as a dictionary. The newly created patient record should be returned as a dictionary.
Patient data for the given id should be updated with the data dictionary. data may not be a full set of the patient fields. This method should return True if a patient was found and updated and False otherwise.
Patient data for the given id should be deleted. This method should return True if a patient was found and deleted and False otherwise.
Returns a list of patients matching the set of lookups. If no patients were found it should return an empty list. If no lookups were passed it should return all patients. The details of the lookup structure is given in the next section. When multiple lookups are passed, the intersection of the results should be returned (default to AND the expressions).
Associates a patient with an addition identifier. The source_id and source_name pair should be enforced as unique. This should return a True value if the association was created. Otherwise it should return False.
Removes an association of a patient with an addition identifier. This should return a True value if the association was found and removed. Otherwise it should return False.
Provider data should be fetched for the given id and returned as a dictionary. If the provider does not exist this method should return None.
A provider record should be created for given set of data given as a dictionary. The newly created provider record should be returned as a dictionary.
Provider data for the given id should be updated with the data dictionary. data may not be a full set of the provider fields. This method should return True if a provider was found and updated and False otherwise.
Provider data for the given id should be deleted. This method should return True if a provider was found and deleted and False otherwise.
Returns a list of providers matching the set of lookups. If no providers were found it should return an empty list. If no lookups were passed it should return all providers. The details of the lookup structure is given in the next section. When multiple lookups are passed, the intersection of the results should be returned (default to AND the expressions).
The above HealthcareStorage.filter_patients() and py:meth:HealthcareStorage.filter_providers methods are each passed a list of lookups for filtering the underlying records. Each of these lookups is a 3-tuple (field_name, operator, value). The field_name is passed as a string and must match a field name on the corresponding data model. The value is the requested value for comparison which should be a standard Python type (int, float, list, sting, date, datetime, etc). The operator is one of the below constants from the healthcare.backends.comparisons module.
| Operator | Comparison |
|---|---|
| EQUAL | Field is an exact match to the value |
| LIKE | Field contains the value |
| IN | Field is an exact match to one of the values in the value (list/tuple) |
| LT | Field is less than the value |
| LTE | Field is less than or equal to the value |
| GT | Field is greater than the value |
| GTE | Field is greater than or equal to the value |
The backend is responsible for mapping these operators to the meaningful expressions for its storage method.
There is a testing mixin BackendTestMixin in healthcare.tests.base which runs through a set of compatibility tests for the backends. You simply need to attach the path of the backend to the backend attribute.
from django.test import TestCase
from healthcare.tests.base import BackendTestMixin
class FancyBackendTestCase(BackendTestMixin, TestCase):
backend = 'path.to.new.backend'
This should not be considered an complete set of tests and the developers should write additional tests to cover edge cases in their backend.