/*
 * (C) Copyright 2006-2011 Nuxeo SA (http://nuxeo.com/) and others.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *
 * Contributors:
 *     bstefanescu
 */
package org.nuxeo.ecm.automation.core.annotations;

import java.lang.annotation.ElementType;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;

import org.nuxeo.ecm.automation.OutputCollector;

/**
 * To be used to mark methods provided by an operation. A method must have at most one argument which is the operation
 * input and a return type which is the operation output. Methods with zero parameters (void input) will match any
 * input. An operation may have multiple methods if it supports multiple input and output types.
 * <p>
 * For each INPUT/OUTPUT type association you must create a new method in an operation and annotate it using this
 * annotation. The set of input types available in an operation are the operation accepted types and the set of output
 * types are the operation produced types. A produced type will become the input type for the next operation in the
 * chain, that will be dispatched to the right method that know how to consume the type.
 * <p>
 * When an operation provides 2 methods accepting the same input type the chain will need to take a decision to
 * determine the best way to continue. A common algorithm to find the right path until the end of the chain is
 * backtracking: from a set of potential nodes one is selected to be visited (randomly or not). If the node is a dead
 * end then another node from the set is taken until the path to the last node is created.
 * <p>
 * A chain may have no paths until the last operation. In this case the chain is invalid and the chain processor will
 * trigger an error. Also, a chain can provide multiple paths to the last operation. To help the engine to find the best
 * path until the last operation you can use the {@link #priority()} attribute to specify which method is preferred. The
 * default priority is 0 (e.g. no priority). Higher priorities have more chance to be selected when a conflict occurs.
 * If no user priority is specified (i.e. priority is 0) then the default priority is used. Here is how the default
 * priority is computed (the top most case has the greater priority):
 * <ul>
 * <li>The input type is an exact match of the method declared argument
 * <li>The method argument type is assignable from the input type (i.e. a super type but not an exact match)
 * <li>The input can be adapted to the method argument using registered type adapters
 * <li>the method has no arguments (void input)
 * </ul>
 * If no one of these rules applies then the method will not match the input.
 * <p>
 * The class owning the annotated method must be annotated using {@link Operation}
 *
 * @author <a href="mailto:bs@nuxeo.com">Bogdan Stefanescu</a>
 */
@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.METHOD)
public @interface OperationMethod {

    /**
     * If defined the method is iterable.
     * <p>
     * It means that when such a method is called with an input type of <code>Iterable<INPUT></code> (where INPUT is the
     * declared method input type) the method will be iteratively called to generate all the outputs and collect them
     * using the given OutputCollector.
     *
     * @return
     */
    @SuppressWarnings("rawtypes")
    Class<? extends OutputCollector> collector() default OutputCollector.class;

    int priority() default 0;

}