Apache Camel is designed to make it very easy to drop in new components whether they be routing components, transformers, transports etc. The idea of a component is to be a factory and manager of Endpoints.
Here are the main steps to writing a component.
Users can then either explicitly create your component, configure it and register with a CamelContext or users can use a URI which auto-creates your component.
When implementing an Endpoint you typically may implement one or more of the following methods
Typically you just derive from DefaultEndpoint and implement the createProducer() and / or createConsumer() methods. The createPollingConsumer() method will be created by default for you in the DefaultEndpoint class.
If your endpoint is a polling-centric component you can derive from DefaultPollingEndpoint and then implement createPollingConsumer(); the createConsumer() method will be created for you which avoids you having to write any polling code.
Annotating your Endpoint
As of Camel 2.12 if you want to benefit from the automatic generation of HTML documentation for all the parameters on your endpoint as part of the maven site reports, you need to annotate your Endpoint's parameters.
So this means you add a @UriEndpoint annotation to your Endpoint class and then annotate each parameter you wish to be configured via the URI configuration mechanism with @UriParam (or @UriParams for nested configuration objects).
In addition its recommended that your Component implementation inherit from the UriEndpointComponent base class as that means your Component will automatically generate better metadata for the ComponentConfiguration API.
Refer to the Endpoint Annotations guide for details.
Using a Processor
Dependency injection and auto-discovery
When using auto-discovery the CamelContext will default to its Injector implementation to inject any required or optional dependencies on the component. This allows you to use auto-discovery of components via URIs while still getting the benefits of dependency injection.
For example your component can depend on a JDBC DataSource or JMS ConnectionFactory which can be provided in the ApplicationContext in Spring or Module in Guice.
So you can if you prefer configure your Component using an IoC framework like Spring or Guice; then add it to the CamelContext. Or you can let the Component auto-inject itself as the endpoints are auto-discovered.
If your component has options you can let it have public getters/setters and Camel will automatically set the properties when the endpoint is created. If you however want to take the matter in your own hands, then you must remove the option from the given parameter list as Camel will validate that all options are used. If not Camel will throw a ResolveEndpointFailedException stating some of the options are unknown.
The parameters is provided by Camel in the createEndpoint method from DefaultComponent:
The code is an example from the SEDA component that removes the size parameter: