Streaming from remote SFTP directories and sub-directories with Spring Integration

Question

I am using Spring Integration Streaming Inbound Channel Adapter, to get stream from remote SFTP and parse every lines of content process.

I use :

IntegrationFlows.from(Sftp.inboundStreamingAdapter(template)
                          .filter(remoteFileFilter)
                          .remoteDirectory("test_dir"),
                        e -> e.id("sftpInboundAdapter")
                              .autoStartup(true)
                              .poller(Pollers.fixedDelay(fetchInt)))
                .handle(Files.splitter(true, true))
....

And it can work now. But I can only get file from test_dir directory, but I need to recursively get files from this dir and sub-directory and parse every line.

I noticed that the Inbound Channel Adapter which is Sftp.inboundAdapter(sftpSessionFactory).scanner(...) . It can scan sub-directory. But I didn't see anything for Streaming Inbound Channel Adapter.

So, how can I implement the 'recursively get files from dir' in Streaming Inbound Channel Adapter?

Thanks.

Gary Russell Gary Russell · Accepted Answer · 2020-04-02T14:14:39

You can use a two outbound gateways - the first doing ls -R (recursive list); split the result and use a gateway configured with mget -stream to get each file.

EDIT

@SpringBootApplication
public class So60987851Application {

    public static void main(String[] args) {
        SpringApplication.run(So60987851Application.class, args);
    }

    @Bean
    IntegrationFlow flow(SessionFactory<LsEntry> csf) {
        return IntegrationFlows.from(() -> "foo", e -> e.poller(Pollers.fixedDelay(5_000)))
                .handle(Sftp.outboundGateway(csf, Command.LS, "payload")
                        .options(Option.RECURSIVE, Option.NAME_ONLY)
                        // need a more robust metadata store for persistence, unless the files are removed
                        .filter(new SftpPersistentAcceptOnceFileListFilter(new SimpleMetadataStore(), "test")))
                .split()
                .log()
                .enrichHeaders(headers -> headers.headerExpression("fileToRemove", "'foo/' + payload"))
                .handle(Sftp.outboundGateway(csf, Command.GET, "'foo/' + payload")
                        .options(Option.STREAM))
                .split(new FileSplitter())
                .log()
                // instead of a filter, we can remove the remote file.
                // but needs some logic to wait until all lines read
//              .handle(Sftp.outboundGateway(csf, Command.RM, "headers['fileToRemove']"))
//              .log()
                .get();
    }

    @Bean
    CachingSessionFactory<LsEntry> csf(DefaultSftpSessionFactory sf) {
        return new CachingSessionFactory<>(sf);
    }

    @Bean
    DefaultSftpSessionFactory sf() {
        DefaultSftpSessionFactory sf = new DefaultSftpSessionFactory();
        sf.setHost("10.0.0.8");
        sf.setUser("gpr");
        sf.setPrivateKey(new FileSystemResource(new File("/Users/grussell/.ssh/id_rsa")));
        sf.setAllowUnknownKeys(true);
        return sf;
    }

}

Streaming from remote SFTP directories and sub-directories with Spring Integration

1 Answers