FlowApp.java

package com.ontology2.haruhi;

import com.google.common.collect.Iterators;
import com.google.common.collect.Lists;
import com.google.common.collect.PeekingIterator;
import com.ontology2.centipede.shell.CommandLineApplication;
import com.ontology2.haruhi.flows.Flow;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.context.ApplicationContext;
import org.springframework.stereotype.Component;

import java.util.List;

@Component("flow")
public class FlowApp extends CommandLineApplication {
    private static Log logger = LogFactory.getLog(FlowApp.class);
    @Autowired private ApplicationContext applicationContext;
    @Autowired private MavenManagedJar defaultJar;
    @Autowired private Cluster defaultCluster;
    @Autowired private ApplicationConfigurationFetcher fetcher;

    @Override
    protected void _run(String[] arguments) throws Exception {
        PeekingIterator<String> a=Iterators.peekingIterator(Iterators.forArray(arguments));
        if (!a.hasNext())
            usage();

        Cluster cluster=defaultCluster;
        MavenManagedJar jar=defaultJar;
        while(a.hasNext() && a.peek().startsWith("-")) {
            String flagName=a.next().substring(1).intern();
            if (!a.hasNext())
                usage();
            
            String flagValue=a.next();
            if (flagName=="clusterId") {
                cluster=applicationContext.getBean(flagValue,Cluster.class);
            } else if(flagName=="jarId") {
                jar=applicationContext.getBean(flagValue,MavenManagedJar.class);
            } else {
                usage();
            };
        }

        //
        // Note one way or another we have to know what the JAR is in order to do
        // this
        //

        ApplicationContext enrichedContext=fetcher.enrichedContext();
        if (!a.hasNext())
            usage();
        
        String flowId=a.next();
        Flow f=enrichedContext.getBean(flowId,Flow.class);
        
        List<String> flowArgs=Lists.newArrayList();
        Iterators.addAll(flowArgs, a);
        cluster.runFlow(jar, f, flowArgs);
    }

    private void usage() {
            System.out.println("To submit a job flow to the FlowApplication do the following:");
            System.out.println();
            System.out.println("haruhi run flow [options] flow_id flow_args ...");
            System.out.println("");
            System.out.println("The system will pass on any arguments beyond the options to ");
            System.out.println("the Hadoop application.  The system will use default options for the cluster");
            System.out.println("and JAR configuration unless you override them with the following options:");
            System.out.println("");
            System.out.println(" -clusterId <clusterId>");
            System.out.println(" -jarId <jarId>");
            System.out.println("");
            System.out.println("both of these arguments of Spring bean names.  If you want to add new");
            System.out.println("configurations,  this application searches");
            System.out.println("");
            System.out.println("$HOME/.haruhi/applicationContext.xml");
            System.out.println("");
            System.out.println("where you can override existing bean definitions or define new ones.");
            System.exit(-1);
    }

}