Elasticsearch HTTP client

This is a an implementation of a Java HTTP client for Elasticsearch. The client API is compatible to the Elasticsearch TransportClient API.

Here is an example

        try (HttpClient client = HttpClient.builder()
                .url(new URL("http://127.0.0.1:9200"))
                .build()) {
            CreateIndexRequestBuilder createIndexRequestBuilder =
                    new CreateIndexRequestBuilder(client, CreateIndexAction.INSTANCE).setIndex("test");
            createIndexRequestBuilder.execute().actionGet();
            IndexRequestBuilder indexRequestBuilder =
                    new IndexRequestBuilder(client, IndexAction.INSTANCE)
                            .setIndex("test")
                            .setType("type")
                            .setId("1")
                            .setSource(jsonBuilder().startObject().field("name", "Hello World").endObject());
            indexRequestBuilder.execute().actionGet();
            RefreshRequestBuilder refreshRequestBuilder =
                    new RefreshRequestBuilder(client, RefreshAction.INSTANCE)
                            .setIndices("test");
            refreshRequestBuilder.execute().actionGet();
            SearchRequestBuilder searchRequestBuilder = new SearchRequestBuilder(client, SearchAction.INSTANCE)
                    .setIndices("test")
                    .setQuery(QueryBuilders.matchAllQuery()).setSize(0);
            assertTrue(searchRequestBuilder.execute().actionGet().getHits().getTotalHits() > 0);
        }

There is also bulk support. Use the class HttpBulkProcessor which is compatible to Elasticsearch BulkProcessor.

Or you can use HttpBulkClient like this

int maxthreads = Runtime.getRuntime().availableProcessors();
        int maxactions = MAX_ACTIONS;
        final long maxloop = NUM_ACTIONS;
        logger.info("HttpBulkNodeClient max={} maxactions={} maxloop={}", maxthreads, maxactions, maxloop);
        final HttpBulkClient client = HttpBulkClient.builder()
                .url(new URL("http://127.0.0.1:9200"))
                .maxActionsPerRequest(maxactions)
                .flushIngestInterval(TimeValue.timeValueSeconds(60))
                .build();
        try {
            client.newIndex("test")
                    .startBulk("test", -1);
            ThreadPoolExecutor pool = EsExecutors.newFixed("http-bulk-nodeclient-test", maxthreads, 30,
                    EsExecutors.daemonThreadFactory("http-bulk-nodeclient-test"));
            final CountDownLatch latch = new CountDownLatch(maxthreads);
            for (int i = 0; i < maxthreads; i++) {
                pool.execute(() -> {
                    for (int j = 0; j < maxloop; j++) {
                        client.index("test", "test", null, "{ \"name\" : \"" + randomString(32) + "\"}");
                    }
                    latch.countDown();
                });
            }
            logger.info("waiting for max 30 seconds...");
            latch.await(30, TimeUnit.SECONDS);
            logger.info("flush...");
            client.flushIngest();
            client.waitForResponses(TimeValue.timeValueSeconds(30));
            logger.info("got all responses, thread pool shutdown...");
            pool.shutdown();
            logger.info("pool is shut down");
            client.stopBulk("test", 1000);
            if (client.hasThrowable()) {
                logger.error("error", client.getThrowable());
            }
            assertFalse(client.hasThrowable());
            client.refreshIndex("test");
            SearchRequestBuilder searchRequestBuilder = new SearchRequestBuilder(client.client(), SearchAction.INSTANCE)
                    .setIndices("test")
                    .setQuery(QueryBuilders.matchAllQuery()).setSize(0);
            assertEquals(maxthreads * maxloop,
                    searchRequestBuilder.execute().actionGet().getHits().getTotalHits());
        } catch (NoNodeAvailableException e) {
            logger.warn("skipping, no node available");
        } finally {
            client.shutdown();
        }

The HTTP client is not complete yet. Many Elasticsearch actions are missing.

You are welcome to send pull requests at https://github.com/jprante/elasticsearch-client-http if you like to improve this project.