ppazos / openehr-rest-client

openEHR REST Client

This is the Groovy reference implementation of the openEHR REST Client for connecting to openEHR REST API implementations.

It compiles to Java bytecode so it can be used as any Java library with any language that compiles to the JVM.

Compile

$ gradle build

It generates a JAR file in the build/libs folder.

Run tests

run just one test

$ gradle test --tests "OpenEhrRestClientTest.create demographic family trees"

run tests with dots in the names (using wild cards)

$ gradle test --tests "OpenEhrRestClientTest.B*4*a* get composition at version"

run all tests inside a category

$ gradle test --tests "OpenEhrRestClientTest.B*4*"

Use it

// Setup authentication using a token
def token = "abc...."
def auth = new TokenAuth(token)

// Build the client
def client = new OpenEhrRestClient(
   "http://localhost:8090/openehr/v1",
   auth,
   ContentTypeEnum.JSON
)

// Set committer metadata headers
client.setCommitterHeader('name="John Doe", external_ref.id="BC8132EA-8F4A-11E7-BB31-BE2E44B06B34", external_ref.namespace="demographic", external_ref.type="PERSON"')

// Upload EHR_STATUS template
String opt = this.getClass().getResource('/ehr_status_any_en_v1.opt').text
client.uploadTemplate(opt)

// Create EHR
client.createEhr()

// Commit a clinical document
// - Upload template for clinical document
// - Get sample document, parse it into a COMPOSITION, commit the COMPOSITION
String opt1 = this.getClass().getResource('/minimal_evaluation.opt').text
client.uploadTemplate(opt1)

String json_doc = this.getClass().getResource('/minimal_evaluation.en.v1_20230205.json').text
def parser = new OpenEhrJsonParserQuick()
def doc = parser.parseJson(json_doc)
client.createComposition(ehr.ehr_id.value, doc)

openehr-rest-client's Issues

Allow configuration of the sut_api_url without changing codebase

We should provide a simple way of passing this parameter when running the tests, without needing to change the application.properties file in the project

Update groovy dependencies

Project is using old groovy dependencies (2020) and there are some problems using the latest JDK. I'd suggest we move to groovy-all:3.0.17 and spock-core:2.3-groovy-3.0. Gradle could also be updated to 7.6

Remove truncate calls when not needed

Some of the tests are performing unnecessary calls to the truncateServer() method during the cleanup phase

Allow configuration of DB truncation

It should be possible to run the tests with or without DB truncation

Verify error codes instead of messages

Since error messages are not part of the specs and will vary quite a lot between implementations, I would suggest we check HTTP status codes instead. Those are included in the specs and should be implementation agnostic.

Add get composition

Make authentication optional

We should have a way of disabling authentication if it's not required

Add create directory

Add support to pass committed data for the committer header to the POST methods

This is now hardcoded https://github.com/ppazos/openehr-rest-client/blob/master/src/main/groovy/com/cabolabs/openehr/rest/client/OpenEhrRestClient.groovy#L138

Maybe by presetting them before calling the commit services.

Add a method to access the last response code

Conformance will need to access the response code in order to check it complies with the specs.

Add demographic endpoints and tests

Add list stored queries client

Upload template operation should receive an instance of OperationalTemplate

For this we need serialization for OPT to XML in the SDK CaboLabs/openEHR-SDK#83

Add get EHR client

Organize test cases in their respective classes by component

EHR
EHR_STATUS
COMPOSITION
FOLDER
CONTRIBUTION
ADL
OPT
QUERY
ACTOR
PARTY_RELATIONSHIP

Create random EHR ids instead of fixed ones

When testing the EHR endpoint we could use random UUIDs instead of fixed ones, that way there is no need to truncate the Db each time we run the test.

Default representation details negotiation

The conformance is currently running with a hardcoded HTTP header for representation details (Prefer: return=representation) and it is expecting resolved references in the creation of EHRs (EhrDto). We can support both behaviours by allowing defining these headers, returning EhrDto when configured to resolve references (i.e. Prefer: return=representation, resolve_refs) and Ehr if not (i.e. Prefer: return=representation).