Red Hat Training

A Red Hat training course is available for Red Hat Fuse

Chapter 48. HL7

HL7 Component

The hl7 component is used for working with the HL7 MLLP protocol and HL7 v2 messages using the HAPI library.
This component supports the following:
  • HL7 MLLP codec for Mina
  • Agnostic data format using either plain String objects or HAPI HL7 model objects.
  • Type Converter from/to HAPI and String
  • HL7 DataFormat using HAPI library
  • Even more ease-of-use as it's integrated well with the Chapter 76, MINA [Deprecated] (Camel 2.11: camel-mina2 ) component.
Maven users will need to add the following dependency to their pom.xml for this component: 
<dependency>
    <groupId>org.apache.camel</groupId>
    <artifactId>camel-hl7</artifactId>
    <version>x.x.x</version>
    <!-- use the same version as your Camel core version -->
</dependency>

HL7 MLLP protocol

HL7 is often used with the HL7 MLLP protocol that is a text based TCP socket based protocol. This component ships with a Mina Codec that conforms to the MLLP protocol so you can easily expose a HL7 listener that accepts HL7 requests over the TCP transport.
To expose a HL7 listener service we reuse the existing mina/mina2 component where we just use the HL7MLLPCodec as codec.
The HL7 MLLP codec has the following options:
Name Default Value Description
startByte 0x0b The start byte spanning the HL7 payload.
endByte1 0x1c The first end byte spanning the HL7 payload.
endByte2 0x0d The 2nd end byte spanning the HL7 payload.
charset JVM Default The encoding (is a charset name) to use for the codec. If not provided, Camel will use the JVM default Charset.
convertLFtoCR true (Camel 2.11:false) Will convert \n to \r (0x0d, 13 decimal) as HL7 stipulates \r as segment terminators. The HAPI library requires the use of \r.
validate true Whether HAPI Parser should validate or not.
parser ca.uhn.hl7v2.parser.PipeParser *Camel 2.11:* To use a custom parser. Must be of type ca.uhn.hl7v2.parser.Parser.

Exposing a HL7 listener

In our Spring XML file, we configure an endpoint to listen for HL7 requests using TCP: 
        <endpoint id="hl7listener" uri="mina:tcp://localhost:8888?sync=true&odec=#hl7codec"/>
        <!-- Camel 2.11: uri="mina2:tcp... -->
Notice that we use TCP on localhost on port 8888. We use sync=true to indicate that this listener is synchronous and therefore will return a HL7 response to the caller. Then we setup mina to use our HL7 codec with codec=#hl7codec. Notice that hl7codec is just a Spring bean ID, so we could have named it mygreatcodecforhl7 or whatever. The codec is also set up in the Spring XML file: 
    <bean id="hl7codec" class="org.apache.camel.component.hl7.HL7MLLPCodec">
        <property name="charset" value="iso-8859-1"/>
    </bean>
Above we also configure the charset encoding to use (iso-8859-1).
The endpoint hl7listener can then be used in a route as a consumer, as this Java DSL example illustrates: 
    from("hl7listener").to("patientLookupService");
This is a very simple route that will listen for HL7 and route it to a service named patientLookupService that is also a Spring bean ID we have configured in the Spring XML as: 
    <bean id="patientLookupService" class="com.mycompany.healthcare.service.PatientLookupService"/>
Another powerful feature of Camel is that we can have our business logic in POJO classes that is not tied to Camel as shown here: 

import ca.uhn.hl7v2.HL7Exception;
import ca.uhn.hl7v2.model.Message;
import ca.uhn.hl7v2.model.v24.segment.QRD;

public class PatientLookupService {
    public Message lookupPatient(Message input) throws HL7Exception {
        QRD qrd = (QRD)input.get("QRD");
        String patientId = qrd.getWhoSubjectFilter(0).getIDNumber().getValue();

        // find patient data based on the patient id and create a HL7 model object with the response
        Message response = ... create and set response data
        return response
    }
Notice that this class uses just imports from the HAPI library and not from Camel.

HL7 Model using java.lang.String

The HL7MLLP codec uses plain String as its data format. Camel uses its Type Converter to convert to/from strings to the HAPI HL7 model objects. However, you can use plain String objects if you prefer, for instance if you wish to parse the data yourself.
See samples for such an example.

HL7v2 Model using HAPI

The HL7v2 model uses Java objects from the HAPI library. Using this library, we can encode and decode from the EDI format (ER7) that is mostly used with HL7v2. With this model you can code with Java objects instead of the EDI based HL7 format that can be hard for humans to read and understand.
The sample below is a request to lookup a patient with the patient ID 0101701234. 
MSH|^~\\&|MYSENDER|MYRECEIVER|MYAPPLICATION||200612211200||QRY^A19|1234|P|2.4
QRD|200612211200|R|I|GetPatient|||1^RD|0101701234|DEM||
Using the HL7 model we can work with the data as a ca.uhn.hl7v2.model.Message object. To retrieve the patient ID in the message above, you can do this in Java code: 
Message msg = exchange.getIn().getBody(Message.class);
QRD qrd = (QRD)msg.get("QRD");
String patientId = qrd.getWhoSubjectFilter(0).getIDNumber().getValue();
If you know the message type in advance, you can be more type-safe: 
QRY_A19 msg = exchange.getIn().getBody(QRY_A19.class);
String patientId = msg.getQRD().getWhoSubjectFilter(0).getIDNumber().getValue();
Camel has built-in type converters, so when this operation is invoked: 
Message msg = exchange.getIn().getBody(Message.class);
Camel will convert the received HL7 data from String to Message. This is powerful when combined with the HL7 listener, then you as the end-user don't have to work with byte[], String or any other simple object formats. You can just use the HAPI HL7v2 model objects.

Message Headers

The unmarshal operation adds these MSH fields as headers on the Camel message:
Key MSH field Example
CamelHL7SendingApplication MSH-3 MYSERVER
CamelHL7SendingFacility MSH-4 MYSERVERAPP
CamelHL7ReceivingApplication MSH-5 MYCLIENT
CamelHL7ReceivingFacility MSH-6 MYCLIENTAPP
CamelHL7Timestamp MSH-7 20071231235900
CamelHL7Security MSH-8 null
CamelHL7MessageType MSH-9-1 ADT
CamelHL7TriggerEvent MSH-9-2 A01
CamelHL7MessageControl MSH-10 1234
CamelHL7ProcessingId MSH-11 P
CamelHL7VersionId MSH-12 2.4
All headers are String types. If a header value is missing, its value is null.

Options

The HL7 Data Format supports the following options:
Option Default Description
validate true Whether the HAPI Parser should validate using the default validation rules. Camel 2.11: better use the parser option and initialize the parser with the desired HAPI ValidationContext
parser ca.uhn.hl7v2.parser.GenericParser *Camel 2.11:* To use a custom parser. Must be of type ca.uhn.hl7v2.parser.Parser. Note that GenericParser also allows to parse XML-encoded HL7v2 messages.

Dependencies

To use HL7 in your Camel routes you'll need to add a dependency on camel-hl7 listed above, which implements this data format.
The HAPI library since Version 0.6 has been split into a base library and several structure libraries, one for each HL7v2 message version:
By default camel-hl7 only references the HAPI base library. Applications are responsible for including structure libraries themselves. For example, if a application works with HL7v2 message versions 2.4 and 2.5 then the following dependencies must be added: 
<dependency>
    <groupId>ca.uhn.hapi</groupId>
    <artifactId>hapi-structures-v24</artifactId>
    <version>1.2</version>
    <!-- use the same version as your hapi-base version -->
</dependency>
<dependency>
    <groupId>ca.uhn.hapi</groupId>
    <artifactId>hapi-structures-v25</artifactId>
    <version>1.2</version>
    <!-- use the same version as your hapi-base version -->
</dependency>
Alternatively, an OSGi bundle containing the base library, all structures libraries and required dependencies (on the bundle classpath) can be downloaded from the central Maven repository. 
<dependency>
    <groupId>ca.uhn.hapi</groupId>
    <artifactId>hapi-osgi-base</artifactId>
    <version>1.2</version>
</dependency>

Terser language (Camel 2.11)

HAPI provides a Terser class that provides access to fields using a commonly used terse location specification syntax. The Terser language allows to use this syntax to extract values from messages and to use them as expressions and predicates for filtering, content-based routing etc.
Sample: 
import static org.apache.camel.component.hl7.HL7.terser;
...

   // extract patient ID from field QRD-8 in the QRY_A19 message above and put into message header
   from("direct:test1")
      .setHeader("PATIENT_ID",terser("QRD-8(0)-1"))
      .to("mock:test1");
   // continue processing if extracted field equals a message header
   from("direct:test2")
      .filter(terser("QRD-8(0)-1")
      .isEqualTo(header("PATIENT_ID"))
      .to("mock:test2");

HL7 Validation predicate (Camel 2.11)

Often it is preferable to parse a HL7v2 message and validate it against a HAPI ValidationContext in a separate step afterwards.
Sample: 

import static org.apache.camel.component.hl7.HL7.messageConformsTo;
import ca.uhn.hl7v2.validation.impl.DefaultValidation;
...

  // Use standard or define your own validation rules
   ValidationContext defaultContext = new DefaultValidation(); 

   // Throws PredicateValidationException if message does not validate
   from("direct:test1").validate(messageConformsTo(defaultContext)).to("mock:test1");

HL7 Acknowledgement expression (Camel 2.11)

A common task in HL7v2 processing is to generate an acknowledgement message as response to an incoming HL7v2 message, e.g. based on a validation result. The ack expression lets us accomplish this very elegantly: 

import static org.apache.camel.component.hl7.HL7.messageConformsTo;
import static org.apache.camel.component.hl7.HL7.ack;
import ca.uhn.hl7v2.validation.impl.DefaultValidation;
...

  // Use standard or define your own validation rules
   ValidationContext defaultContext = new DefaultValidation(); 

   from("direct:test1")
      .onException(Exception.class)
         .handled(true)
         .transform(ack()) // auto-generates negative ack because of exception in Exchange
         .end()
      .validate(messageConformsTo(defaultContext))
      // do something meaningful here
      ...
      // acknowledgement
      .transform(ack())

More Samples

In the following example we send a HL7 request to a HL7 listener and retrieves a response. We use plain String types in this example: 
String line1 = "MSH|^~\\&|MYSENDER|MYRECEIVER|MYAPPLICATION||200612211200||QRY^A19|1234|P|2.4";
String line2 = "QRD|200612211200|R|I|GetPatient|||1^RD|0101701234|DEM||";

StringBuilder in = new StringBuilder();
in.append(line1);
in.append("\n");
in.append(line2);

String out = (String)template.requestBody("mina2:tcp://127.0.0.1:8888?sync=true&codec=#hl7codec", in.toString());
In the next sample, we want to route HL7 requests from our HL7 listener to our business logic. We have our business logic in a plain POJO that we have registered in the registry as hl7service = for instance using Spring and letting the bean id = hl7service.
Our business logic is a plain POJO only using the HAPI library so we have these operations defined: 
public class MyHL7BusinessLogic {

    // This is a plain POJO that has NO imports whatsoever on Apache Camel.
    // its a plain POJO only importing the HAPI library so we can much easier work with the HL7 format.

    public Message handleA19(Message msg) throws Exception {
        // here you can have your business logic for A19 messages
        assertTrue(msg instanceof QRY_A19);
        // just return the same dummy response
        return createADR19Message();
    }

    public Message handleA01(Message msg) throws Exception {
        // here you can have your business logic for A01 messages
        assertTrue(msg instanceof ADT_A01);
        // just return the same dummy response
        return createADT01Message();
    }
}
Then we set up the Camel routes using the RouteBuilder as follows: 
DataFormat hl7 = new HL7DataFormat();
// we setup or HL7 listener on port 8888 (using the hl7codec) and in sync mode so we can return a response
from("mina2:tcp://127.0.0.1:8888?sync=true&codec=#hl7codec")
    // we use the HL7 data format to unmarshal from HL7 stream to the HAPI Message model
    // this ensures that the camel message has been enriched with hl7 specific headers to
    // make the routing much easier (see below)
    .unmarshal(hl7)
    // using choice as the content base router
    .choice()
        // where we choose that A19 queries invoke the handleA19 method on our hl7service bean
        .when(header("CamelHL7TriggerEvent").isEqualTo("A19"))
            .beanRef("hl7service", "handleA19")
            .to("mock:a19")
        // and A01 should invoke the handleA01 method on our hl7service bean
        .when(header("CamelHL7TriggerEvent").isEqualTo("A01")).to("mock:a01")
            .beanRef("hl7service", "handleA01")
            .to("mock:a19")
        // other types should go to mock:unknown
        .otherwise()
            .to("mock:unknown")
    // end choice block
    .end()
    // marshal response back
    .marshal(hl7);
Notice that we use the HL7 DataFormat to enrich our Camel Message with the MSH fields preconfigured on the Camel Message. This lets us much more easily define our routes using the fluent builders. If we do not use the HL7 DataFormat, then we do not gains these headers and we must resort to a different technique for computing the MSH trigger event (= what kind of HL7 message it is). This is a big advantage of the HL7 DataFormat over the plain HL7 type converters.

Sample using plain String objects

In this sample we use plain String objects as the data format, that we send, process and receive. As the sample is part of a unit test, there is some code for assertions, but you should be able to understand what happens. First we send the plain string, Hello World, to the HL7MLLPCodec and receive the response as a plain string, Bye World. 
MockEndpoint mock = getMockEndpoint("mock:result");
mock.expectedBodiesReceived("Bye World");

// send plain hello world as String
Object out = template.requestBody("mina2:tcp://127.0.0.1:8888?sync=true&codec=#hl7codec", "Hello World");

assertMockEndpointsSatisfied();

// and the response is also just plain String
assertEquals("Bye World", out);
Here we process the incoming data as plain String and send the response also as plain String: 
from("mina2:tcp://127.0.0.1:8888?sync=true&codec=#hl7codec")
    .process(new Processor() {
        public void process(Exchange exchange) throws Exception {
            // use plain String as message format
            String body = exchange.getIn().getBody(String.class);
            assertEquals("Hello World", body);

            // return the response as plain string
            exchange.getOut().setBody("Bye World");
        }
    })
    .to("mock:result");