OpenEHR REST Definitions API

Authentication and authorization ¶

Services SHOULD implement and support an HTTP Authentication and Authorization framework (which can support various schemes) but there is no assumption or recommendation being made in this specification about which authentication scheme should be used by services and clients. See RFC 7235 or Mozilla’s HTTP Authentication for details on this subject.

If an Authentication and Authorisation framework is present, services MUST properly use WWW-Authenticate and/or Proxy-Authenticate response headers and return HTTP status code 403 Forbidden or 401 Unauthorized or 407 Proxy Authentication whenever applicable, and clients MUST properly use Authorization and Proxy-Authorization in their request headers.

Options and conformance ¶

HTTP headers ¶

This specification is using a few standard and custom HTTP headers that are described below.

The openEHR-VERSION and openEHR-AUDIT_DETAILS headers

When it comes to committing content to an openEHR system, for all change-controlled resources (e.g. COMPOSITION, EHR_STATUS, FOLDER, etc.) the services are performing versioning under the hood. Thus, the ‘native’ way of committing content is to wrap it as VERSION and use CONTRIBUTION.

But in order to keep things simpler and consistent, service MUST allow PUT, POST and DELETE methods also directly on these change-controlled resources. However, internally, these operations MUST be executed using the ‘native’ way with CONTRIBUTION.

In order to allow clients to provide committal metadata, services MUST accept openEHR-VERSION and openEHR-AUDIT_DETAILS request custom headers. For clients, it is RECOMMENDED to provision these headers based on authentication and authorization runtime data. Below is a complex example of these request headers used in a PUT action to update a COMPOSITION:

openEHR-VERSION.lifecycle_state: code_string="532"
openEHR-AUDIT_DETAILS.change_type: code_string="251"
openEHR-AUDIT_DETAILS.description: value="An updated composition contribution description"
openEHR-AUDIT_DETAILS.committer: name="John Doe", external_ref.id="BC8132EA-8F4A-11E7-BB31-BE2E44B06B34", external_ref.namespace="demographic", external_ref.type="PERSON"

None of these headers are mandatory, but whatever is provided it MUST be merged with the default VERSION and VERSION.audit_details attributes on commit runtime.

Below a list of code_string values and their meaning (taken from openEHR terminology):

The If-Match header and accidental overwrites

The use case of this is described by RFC 7232.

If-Match: 8849182c-82ad-4088-a07f-48ead4180515::example.domain.com::2

This HTTP request header SHOULD be used by the clients to prevent accidental overwrites when multiple user agents might be acting in parallel on the same resource. This is only required by a small set of resources of this specification, as in most of the other cases the preceding_version_uid path segment is instead required in order to prevent such accidental overwrites. In case a service receives this header and the condition evaluates to false it MUST respond with HTTP status code412 Precondition Failed.

The Location header

This response header indicates the resource location (URL). According to RFC 7231, it is used to refer to a specific resource in relation to the response. The type of relationship is defined by the combination of request method and status code semantics. Services MUST return this header whenever a create or update operation was performed, but it MAY return this header on other operation or action. Example:

Location: http://www.openehr.org/api/v1/ehr/347a5490-55ee-4da9-b91a-9bba710f730e/composition/8849182c-82ad-4088-a07f-48ead4180515::example.domain.com::2

See representation details negotiation section for more details on how use this header.

Services MAY generate resource URL as specified by DV_URI/DV_EHR_URI format, and if this feature is supported then the services MUST use openEHR-uri response header. Example:

openEHR-uri: ehr:/347a5490-55ee-4da9-b91a-9bba710f730e/compositions/87284370-2D4B-4e3d-A3F3-F303D2F4F34B

The Prefer header

This request header MAY be used by clients for resource representation negotiation. See more details on representation details negotiation section.

The ETag and Last-Modified headers

According to RFC 7232,

The “ETag” header field in a response provides the current entity-tag for the selected representation, as determined at the conclusion of handling the request. An entity-tag is an opaque validator for differentiating between multiple representations of the same resource, regardless of whether those multiple representations are due to resource state changes over time, content negotiation resulting in multiple representations being valid at the same time, or both.

ETag: 8849182c-82ad-4088-a07f-48ead4180515::example.domain.com::2

The Last-Modified response HTTP header contains the datetime of the last modification of targeted resource which should be taken from VERSION.commit_audit.time_committed.value. An example of such header value format is:

Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT

See more details on RFC 7232.

These headers SHOULD be present in all responses targeting VERSION, VERSIONED_OBJECT or other resources that have unique identifier for each of the changes in time. The value of this header MAY be the unique identifier of that resource (e.g. VERSIONED_OBJECT.uid.value, VERSION.uid.value, EHR.ehr_id.value, etc).

HTTP status codes ¶

To indicate the status of the request or the executed operation, one of the following subset of the HTTP status codes RFC 7231 MUST be used:

Meaning of these codes may be further detailed (nuanced) in this specification by particular responses. If required, other status codes MAY be used by implementations as long as their usecase is not conflicting or overlapping with above codes.

In case of errors (HTTP codes 400-500), the services MAY return more details (if Prefer: return=representation request header is present). Example:

some codes/messages https://github.com/ppazos/cabolabs-ehrserver/wiki/API-error-codes-and-messages and http://veratechnas1.synology.me:13606/InstanceValidator/rules.html

{
    "message": "Error message",
    "code": 90000,
    "errors": [
        {
            "_type": "DV_CODED_TEXT",
            "value": "Error message",
            "defining_code": {
                "terminology_id": {
                    "value": "local"
                },
                "code_string": "9000"
            }
        },
        {
            "_type": "DV_CODED_TEXT",
            "value": "Secondary error message",
            "defining_code": {
                "terminology_id": {
                    "value": "local"
                },
                "code_string": "8000"
            }
        }
    ]
}

Data representation ¶

Services MUST support at least one of the openEHR XML or JSON formats (described below) for resource representation.

XML Format

When resources representation is serialized as XML, the result MUST be valid against published XSDs.

A client MAY use the header Content-Type: application/xml in the requests to specify the XML payload format. If the service cannot process the request payload as XML format is not supported, it MUST respond with HTTP status code415 Unsupported Media Type.

The Accept: application/xml SHOULD be used in the request by the client in order to specify the expected XML response format. If the service cannot fulfill this aspect of the request, it MUST respond with HTTP status code 406 Not Acceptable. Proper header Content-Type: application/xml MUST be present in the response of the service unless response has no body (status 204).

JSON Format

Attribute names must be lowercase snake_case names as specified in the equivalent RM type. For example:

{
    "category": {
        "value": "event",
        "defining_code": {
            "terminology_id": {
                "value": "openehr"
            },
            "code_string": "433"
        }
    }
}

Metadata attributes (those that are not also RM attributes) will always be prefixed by a '_'. One example is the _type attribute, which should be used to specify the RM type whenever polymorphism is involved or the underlying definition in RM type is abstract (dynamic type is different from the static type). This follows same rule as for XML typing. The value of this attribute MUST be the uppercase class name from the RM specification. For example:

{
    "_type": "DV_TEXT",
    "value": "Hello world!"
}

The RM attributes (even required ones) that are Null, empty list or empty arrays SHOULD be absent when serialized as JSON.

The order of attributes in the resource MAY follow the order of attributes in the RM specification of the type of the resource, but this is not mandatory.

tbd json-schema

A client MAY use the header Content-Type: application/json in the requests to specify the JSON payload format. If the service cannot process the request payload as JSON format is not supported, it MUST respond with HTTP status code 415 Unsupported Media Type.

The Accept: application/json SHOULD be used in the request by the client in order to specify the expected JSON response format. If the service cannot fulfill this aspect of the request, it MUST respond with HTTP status code 406 Not Acceptable. Proper header Content-Type: application/json MUST be present in the response of the service unless response has no body (status 204).

Datetime format

Query parameters and path segments that are date, datetime, or timestamp, MUST always use the canonical ISO8601 extended format, e.g. 2016-06-23T13:42:16.117+02:00.

Any date, datetime or timestamp value provided by a create or update action inside the COMPOSITION content will be preserved as it was sent by client (i.e. if COMPOSITION was saved as narrow format, it will always return the way it was). However, the strongly RECOMMENDED format is to use the ISO8601 extended format.

Representation details negotiation ¶

When using the HTTP methods to create or update a resource, the service SHOULD give clients the option of returning either a complete representation of the (modified) resource, or a minimal or no content in the payload response (assuming the operation was successfully completed). See RFC 7240 for more details on how achieve this using Prefer header. The client MAY choose any of the following:

• send Prefer: return=minimal to inform the service that prefers only a minimal response to a successful request. A Location header indicating the direct URL to access the resource MUST be part of the service response. If there is no payload content to be returned, the service SHOULD use HTTP status code 204 No Content.

• send Prefer: return=representation to inform the service that prefers a full representation response to a successful request. A Location header indicating the direct URL to access the resource MAY be part of the service response and the payload content SHOULD include a full representation.

In case no Prefer header is present in request, the default response policy is return=minimal.

Another preference is related to following and resolving OBJ_REF identifiers. Under some circumstances a client MAY indicate that prefers response containing full or partial resource representation instead of references to resources using OBJ_REF. A typical case is a list of COMPOSITIONs part of an EHR, which, strictly following RM specification, should always return list of OBJ_REF, but sometimes clients prefers to get COMPOSITIONs. Services that have this capability implemented SHOULD accept and honor Prefer: resolve_refs request header.

Prefer: return=representation, resolve_refs

ADL 1.4 TEMPLATE ¶

Release: 1.0.0

Management of OpenEHR operational templates. OpenEHR templates can be uploaded and downloaded as OPTs created by modelling tools such as the Template Designer.

ADL 2 TEMPLATE ¶

STORED QUERY ¶

Release: 1.0.0

Management of stored (registered) queries in the system.

Stored queries must have a name and an optional version number. Implementers can choose a schema for query names, such as <namespace>::<query-name>, this way queries can be separated by domain, etc. Version numbers, if used must be in SEMVER style (major.minor.patch). If only a partial version is supplied or version is not supplied at all, the system must use the latest version with the supplied prefix. Namespace (if used) should be reverse domain name, which allows for separation of stored queries by teams, companies, etc. In the future several standard openEHR stored queries might be developed, those will use namespace org.openehr.

Management of stored queries is done using resources starting /definition/query described here.