Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

FHIR field for Plone

Project description

Egg Status Travis Build Status Test Coverage Python Versions Latest Version License

Background (

FHIR (Fast Healthcare Interoperability Resources) is the industry standard for Healthcare system. Our intend to implement FHIR based system using Plone! will make life easier to create, manage content for FHIR resources as well search facilities for any FHIR Resources.

How It Works

This field is working as like other zope.schema field, just add and use it. You will feed the field value as either json string or python dict of FHIR resources through web form or any restapi client. This field has built-in FHIR resource validator and parser.


from import FhirResource
from plone.supermodel import model

class IMyContent(model.Schema):

    <resource_type>_resource = FhirResource(
        title=u'your title',
        desciption=u'your desciption',
        resource_type='any fhir resource type[optional]'

The field’s value is the instance of a specilized class FhirResourceValue inside the context, which is kind of proxy class of fhirclient model with additional methods and attributes.

Make field indexable

A specilized Catalog PluginIndexes is named FhirFieldIndex is available, you will use it as like other catalog indexes. However importantly you have to maintain a strict convention (index name must be started with any valid fhir resource name (no matter uppercase or lowercase) and should _(underscore) be used as separator, i.e task_index(<resource>_<any name>))


<?xml version="1.0"?>
<object name="portal_catalog" meta_type="Plone Catalog Tool">
    <index name="organization_resource" meta_type="FhirFieldIndex">
        <indexed_attr value="organization_resource"/>


  • Plone restapi support
  • Widget: z3cform support
  • plone.supermodel support
  • plone.rfc822 marshaler field support
  • 100% FHIR search compliance catalog search.

Available Field’s Options

This field has got all standard options (i.e title, required, desciption and so on) with additionally options for the purpose of either validation or constraint those are related to FHIR.


Required: No

Default: None

Type: String

The name of FHIR resource can be used as constraint, meaning that we can restricted only accept certain resource. If FHIR json is provided that contains other resource type, would raise validation error. Example: FhirResource(….,resource_type=’Patient’)


Required: No

Default: None

Type: String + full python path (dotted) of the model class.

Like resource_type option, it can be used as constraint, however additionally this option enable us to use custom model class other than fhirclient’s model. Example: FhirResource(….,model=’fhirclient.models.patient.Patient’)


Required: No

Default: None

Type: String + full python path (dotted) of the model class.

Unlike model option, this option has more versatile goal although you can use it for single resource restriction. The advanced usecase like, the field could accept muiltiple resources types those model class implements the provided interface. For example you made a interface called IMedicalService and (Organization, Patient, Practitioner) models those are implementing IMedicalService. So when you provides this option value, actually three types of fhir json can now be accepted by this field. Example: FhirResource(….,model=’’)

Field’s Value API

Field’s value is a specilized class which has reponsibilty to act as proxy of fhirclient model’s class. This class provides some powerful methods.


Originally this method is derived from fhirclient base model, you will always have to use this method during negotiation (although our serializer doing that for you automatically). This method not takes any argument, it returns FHIR json representation of resource.


If you are familar with FHIRPath Patch, this method one of the strongest weapon of this class. Patch applying on any FHIR resources is noting but so easy. This method takes one mandatory argument patch_data and that value should be list of patch items (jsonpatch).


from import FhirResource
from plone.supermodel import model

class ITask(model.Schema):

    resource = FhirResource(
        title=u'your title',
        desciption=u'your desciption',

patch_data = [
  {'op': 'replace', 'path': '/source/display', 'value': 'Patched display'},
  {'op': 'replace', 'path': '/status', 'value': 'Reopen'}


This method returns string representation of fhir resource json value. Normally as_json returns python’s dict type data. This method takes optional prettify argument, by setting this argument True, method will return human/print friendly representation.


There may some situation come, where you will need just pure instance of fhir model, this method serves that purpose. This method returns current fhir resource model’s instance.


from fhirclient.models.task import Task
from import FhirResource
from plone.supermodel import model

class ITask(model.Schema):

    resource = FhirResource(
        title=u'your title',
        desciption=u'your desciption',

task = task_content.resource.foreground_origin()
assert isinstance(task, Task)

Helper API

This package provides some useful functions those could be usable in your codebase.


This function return appropriate fhirclient model class based on provided resource type. On wrong resource type zope.interface.Invalid exception is raisen.


>>> from import resource_type_str_to_fhir_model
>>> task_model_class = resource_type_str_to_fhir_model('Task')

elasticsearch setup

If your intent to use elasticsearch based indexing and query, this section for you! you can find more details here

server setup

server version is restricted to 2.4.x, means we cannot use latest version of elasticsearch. i.e 5.6.x

  • Download from here and install according to documentation.
  • For development you could use docker container. The Makefile is available, ~$ make run-es

collective.elasticsearch setup

Full configuration guide could be found here. Simple steps are described bellow.

  1. create catalog/indexes: First you will need add indexes for each fhirfield used in your project. each resource type has it’s own Meta Index. example is here
  2. Install collective.elasticsearch addon from plone control panel.
  3. Convert your Indexes to elasticsearch. Go To {portal url}/@@elastic-controlpanel
  4. In the settings form’s Indexes for which all searches are done through ElasticSearch section add your all indexes those you mentioned into catalog.xml file, also add portal_type
  5. Now save and again Convert Catalog.


Install by adding it to your buildout:



eggs = [elasticsearch]

and then running bin/buildout. Go to plone control and install or If you are creating an addon that depends on this product, you may add <dependency></dependency> in metadata.xml at profiles.


This product provides three plone registry based records,, Those are related to ElasticSearch index mapping setup, if you aware about it, then you have option to modify from plone control panel (Registry).


The project is licensed under the GPLv2.


2.0.0 to 2.x.x

  1. Do full reindex of objects (remove and build)

1.0.0 to 2.0.0

  1. Simply old indexes are not usable any more! as new ES server (6.x.x) is used.
  2. Do backup of your existing Data.fs (important!).
  3. From plone controlpanel, Go to elasticsearch settings page /@@elastic-controlpanel.
  4. Convert Catalog again, make sure in the Indexes for which all searches are done through ElasticSearch section your desired indexes are select.
  5. If your site is behind the proxy, we suggest make ssh tuneling to connect your site. Because next takes much times, (depends on your data size)
  6. Rebuild Catalog && wait for success, if you face any problem try again.

1.0.0rc3 to 1.0.0rc4

  1. Keep backup of existing Data.fs (nice to have)
  2. Go to plone control panel’s Addon settings
  3. Uninstall and Install again

1.0.0b6 to 1.0.0rc3

  1. Keep backup of existing Data.fs (important)
  2. Just Clear and Rebuild existing catalogs. Go to site management -> portal_catalog -> Advanced Tab and click the Clear and Rebuild button. Caution: this activity could take longer time.

1.0.0bx to 1.0.0b6

  1. You will need to remove all the indexes who are based on Fhir*******Index. Go to ZMI to portal catalog tool. {site url}/portal_catalog/manage_catalogIndexes.
  2. Update the version and run the buildout.
  3. List out any products those defined (profile/catalog.xml) indexes based on Fhir*******Index as meta type. Reninstall them all.
  4. From ZMI, portal_catalog tool, {site url}/portal_catalog/manage_catalogAdvanced Clear and Rebuild

FHIR Search Query

This product has a query builder helper, that is actually transforming fhir search params into Plone’s search compatible params so that PortalCatalog tool understand what to do. Each of FHIR Indexes provided by this product, by default using this query builder but it is possible to provide prepared query.

Query Builder takes only HL7 fhir standard search parameters and transforms to elasticsearch compatible query that is executing through plone catalog search. However belows are lists of standard parameters those are supported by this Query Builder (more to continue adding, until full supports are completed)

Integrate in your REST API service

This product has been got all battery included to be integrated with plone.restapi for becoming HL7 FHIR standard RESTful API server which would provide search service as defined here.

Example RESTful service could be found here

Supported Search Parameters

Parameter Name Remarks
_id Yes
_lastUpdated Yes
_profile Yes
birthdate Yes
gender Yes
patient Yes
status Yes
owner Yes
subject Yes
effective Yes
url Yes
version Yes
authored Yes
identifier Yes
based-on Yes
part-of Yes
price-override Yes
quantity Yes
factor-override Yes
length Yes
code Yes
medication Yes
address Yes
address-city Yes
address-country Yes
address-postalcode Yes
address-state Yes
address-use Yes
telecom Yes
email Yes
phone Yes
name Yes
family Yes
given Yes
code-value-concept Yes
code-value-date Yes
code-value-quantity Yes
code-value-string Yes

REST Client Examples

Getting single resource, here we are getting Patient resource by ID.


>>> response = admin_session.get('/@fhir/Patient/19c5245f-89a8-49f8-b244-666b32adb92e')
>>> response.status_code
>>> response.json()['resourceType'] == 'Patient'
>>> response = admin_session.get('/@fhir/Patient/19c5245f-fake-id')
>>> response.status_code

Search Observation by Patient reference with status condition. Any observation until December 2017 and earlier than January 2017.


>>> response = admin_session.get('/@fhir/Observation?patient=Patient/19c5245f-89a8-49f8-b244-666b32adb92e&status=final&_lastUpdated=lt2017-12-31&_lastUpdated=gt2017-01-01')
>>> response.status_code
>>> len(response.json())

Add FHIR Resource through REST API


>>> import os
>>> import json
>>> import uuid
>>> import DateTime
>>> import time

>>> with open(os.path.join(FIXTURE_PATH, 'Patient.json'), 'r') as f:
...     fhir_json = json.load(f)

>>> fhir_json['id'] = str(uuid.uuid4())
>>> fhir_json['name'][0]['text'] = 'Another Patient'
>>> response ='/@fhir/Patient', json=fhir_json)
>>> response.status_code
>>> time.sleep(1)
>>> response = admin_session.get('/@fhir/Patient?active=true')
>>> len(response.json())

Update (PATCH) FHIR Resource the Patient is currently activated, we will deactive.


>>> patch = [{'op': 'replace', 'path': '/active', 'value': False}]
>>> response = admin_session.patch('/@fhir/Patient/19c5245f-89a8-49f8-b244-666b32adb92e', json={'patch': patch})
>>> response.status_code



2.1.0 (2019-08-20)

  • ZCatalog’s index datum value type is now string (json serialized) (previously was dictionary type).

Migration Required

  1. Do complete reindexing.

2.0.0 (2019-05-18)


  • Issue#14 Now reference query is powerful yet! It is possible now search by resourceType only or mixing up.
  • Issue#21 One of the powerful feature added, IN/OR support in search query. Now possible to provide multiple values separated by comma.


  • Drop support Elastic server version 2.3.x.


  • Important fix for Quantity search type, now value prefix not impact on other (unit, system). Additionally also now possible to search by unit or system and code (without value)

1.0.0 (2019-01-18)


  • Drop fhirclient dependency, instead make fhir.resources dependency.
  • collective.elasticsearch version minimum version 2.0.5 has been pinned.

1.0.0rc4 (2018-10-25)


  • Drop support for Plone 4.xx (sorry).


  • Issue#10 JSON Viewer added in display mode.
  • Issue#18 api module to make available for all publicly usable methods, functions, classes.
  • Issue#17 Add suports for duplicate param names into query string. It is now possible to provide multiple condition for same param type. For example ?_lastUpdated=gt2015-10-15T06:31:18+00:00&_lastUpdated=lt2018-01-15T06:31:18+00:00
  • Issue#10 Add support for Composite type FHIR search param.
  • Issue#13 Add support for Address and ContactPoint mapping. It opens up many search params.
  • Mappings are more enriched, so many missing mappings are added, like valueString, valueQuantity and so on.[nazrulworld]
  • Issue#12 Full support for code search param type has been added, it also opens up for other search parameters (y).
  • Issue#15 support for Humanane mapping in search.

1.0.0rc3 (2018-09-22)


  • Issue: 6 A major improvement has been done, now very slim version (id, resourceType, meta) of FHIR json resource has been indexed in ZCatalog (zope index) version, however previously whole fhir resource was stored as a result now huge storage savings, perhaps improves indexing performance. [nazrulworld]

1.0.0rc1 (2018-08-27)


  • Identifier search parameter is active now (both array of identifier and single object identifier).
  • Array of Reference query support (for example basedOn (list of reference) ) is active now. Although normal object reference has already been supported.
  • All available mappings for searchable resources are generated.
  • FHIR search sort feature is available!
  •, and registry records are available to setup ES mapping
  • URI and Number type parameter based search are fully available.
  • `resourceType` filter is automatically injected into every generated query. Query Builder knows about which resourceType should be.

Breaking Changes

  • have to install, as some registry records (settings) for elasticsearch mapping have been introduced.
  • Any deprecated FHIR Field Indexes other than FhirFieldIndex (FhirOrganizationIndex and so on) are removed

1.0.0b7 (2018-08-10)

  • Media search available.
  • exception class available, it would be used to catch any fhir query buiding errors. [nazrulworld]

1.0.0b6 (2018-08-04)

  • Fix: minor type mistake on non existing method called.
  • Migration guide has been added. [nazrulworld]

1.0.0b5 (2018-08-03)


  • FhirFieldIndex Catalog Index has been refactored. Now this class is capable to handle all the FHIR resources. That’s why other PluginIndexes related to FhirField have been deprecated.
  • New ZCatalog (plone index) index naming convention has been introduced. Any index name for FhirFieldIndex must have fhir resource type name as prefix. for example: task_index

1.0.0b4 (2018-08-01)

  • Must Update (fix): Important updates made on mapping, reference field mapping was not working if value contains with /, now made it tokenize by indecating index is not_analyzed
  • _profile search parameter is now available. [nazrulworld]

1.0.0b3 (2018-07-30)

  • Mapping improvment for FhirQuestionnaireResponseIndex, FhirObservationIndex, FhirProcedureRequestIndex, FhirTaskIndex, FhirDeviceRequestIndex

1.0.0b2 (2018-07-29)

New Features:

  • supports for elasticsearch has been added. Now many basic fhir search are possible to be queried.
  • upto 22 FHIR fields indexes (FhirActivityDefinitionIndex, FhirAppointmentIndex, FhirCarePlanIndex, FhirDeviceIndex, FhirDeviceRequestIndex, FhirHealthcareServiceIndex, FhirMedicationAdministrationIndex, FhirMedicationDispenseIndex, FhirMedicationRequestIndex, FhirMedicationStatementIndex, FhirObservationIndex, FhirOrganizationIndex, FhirPatientIndex, FhirPlanDefinitionIndex, FhirPractitionerIndex, FhirProcedureRequestIndex, FhirQuestionnaireIndex, FhirQuestionnaireResponseIndex, FhirRelatedPersonIndex, FhirTaskIndex, FhirValueSetIndex)
  • Mappings for all available fhir indexes are created.
  • elasticsearch option is now available for

1.0.0b1 (2018-03-17)

  • first beta version has been released.

1.0.0a10 (2018-03-12)

  • fix(bug) Issue-3: resource_type constraint don’t raise exception from validator.

1.0.0a9 (2018-03-08)

  • There is no restriction/limit over fhir resources, all available models are supported.

1.0.0a8 (2018-01-22)

1.0.0a7 (2018-01-21)

1.0.0a6 (2018-01-14)

  • missing HealthcareService fhir model is added as supported model.

1.0.0a5 (2018-01-14)

  • Person fhir model added in whitelist.

1.0.0a4 (2018-01-14)

  • IFhirResource.model_interface field type changed to DottedName from InterfaceField.

1.0.0a3 (2017-12-22)

  • FHIR Patch support added. Now patching fhir resource is more easy to manage.
  • plone.supermodel support is confirmed.[nazrulworld]
  • plone.rfc822 marshaler field support.[nazrulworld]

1.0.0a2 (2017-12-10)

  • FhirResourceWidget is made working state. From now it is possible to adapt FhirResourceWidget` with z3c.form [nazrulworld]

1.0.0a1 (2017-12-05)

  • Initial release. [nazrulworld]

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for, version 2.1.0
Filename, size File type Python version Upload date Hashes
Filename, size (321.5 kB) File type Wheel Python version py2.py3 Upload date Hashes View hashes
Filename, size (216.0 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page