|
17 | 17 | import org.apache.http.client.CredentialsProvider; |
18 | 18 | import org.apache.http.impl.client.BasicCredentialsProvider; |
19 | 19 | import org.apache.logging.log4j.core.config.plugins.util.PluginManager; |
| 20 | +import org.apache.lucene.util.IOConsumer; |
20 | 21 | import org.elasticsearch.client.Request; |
21 | 22 | import org.elasticsearch.client.Response; |
22 | 23 | import org.elasticsearch.client.ResponseException; |
|
30 | 31 | import org.elasticsearch.inference.TaskType; |
31 | 32 | import org.elasticsearch.logging.LogManager; |
32 | 33 | import org.elasticsearch.logging.Logger; |
| 34 | +import org.elasticsearch.test.ESTestCase; |
33 | 35 | import org.elasticsearch.test.rest.ESRestTestCase; |
34 | 36 | import org.elasticsearch.xcontent.XContentType; |
35 | 37 |
|
|
43 | 45 | import java.util.List; |
44 | 46 | import java.util.Map; |
45 | 47 | import java.util.Set; |
| 48 | +import java.util.concurrent.Semaphore; |
46 | 49 | import java.util.stream.Collectors; |
47 | 50 | import java.util.stream.Stream; |
48 | 51 |
|
|
53 | 56 | import static org.elasticsearch.xpack.esql.EsqlTestUtils.reader; |
54 | 57 |
|
55 | 58 | public class CsvTestsDataLoader { |
| 59 | + private static final int PARALLEL_THREADS = 10; |
56 | 60 | private static final int BULK_DATA_SIZE = 100_000; |
57 | 61 | private static final TestDataset EMPLOYEES = new TestDataset("employees", "mapping-default.json", "employees.csv").noSubfields(); |
58 | 62 | private static final TestDataset EMPLOYEES_INCOMPATIBLE = new TestDataset( |
@@ -398,16 +402,35 @@ private static void loadDataSetIntoEs( |
398 | 402 | IndexCreator indexCreator |
399 | 403 | ) throws IOException { |
400 | 404 | Logger logger = LogManager.getLogger(CsvTestsDataLoader.class); |
| 405 | + List<TestDataset> datasets = availableDatasetsForEs(supportsIndexModeLookup, supportsSourceFieldMapping, inferenceEnabled).stream() |
| 406 | + .toList(); |
| 407 | + |
| 408 | + executeInParallel(datasets, dataset -> createIndex(client, dataset, indexCreator), "Failed to create indices in parallel"); |
| 409 | + |
| 410 | + executeInParallel(datasets, dataset -> loadData(client, dataset, logger), "Failed to load data in parallel"); |
| 411 | + |
| 412 | + forceMerge(client, datasets.stream().map(d -> d.indexName).collect(Collectors.toSet()), logger); |
| 413 | + |
| 414 | + executeInParallel( |
| 415 | + ENRICH_POLICIES, |
| 416 | + policy -> loadEnrichPolicy(client, policy.policyName, policy.policyFileName, logger), |
| 417 | + "Failed to load enrich policies in parallel" |
| 418 | + ); |
401 | 419 |
|
402 | | - Set<String> loadedDatasets = new HashSet<>(); |
403 | | - for (var dataset : availableDatasetsForEs(supportsIndexModeLookup, supportsSourceFieldMapping, inferenceEnabled)) { |
404 | | - load(client, dataset, logger, indexCreator); |
405 | | - loadedDatasets.add(dataset.indexName); |
406 | | - } |
407 | | - forceMerge(client, loadedDatasets, logger); |
408 | | - for (var policy : ENRICH_POLICIES) { |
409 | | - loadEnrichPolicy(client, policy.policyName, policy.policyFileName, logger); |
410 | | - } |
| 420 | + } |
| 421 | + |
| 422 | + private static <T> void executeInParallel(List<T> items, IOConsumer<T> consumer, String errorMessage) { |
| 423 | + Semaphore semaphore = new Semaphore(PARALLEL_THREADS); |
| 424 | + ESTestCase.runInParallel(items.size(), i -> { |
| 425 | + try { |
| 426 | + semaphore.acquire(); |
| 427 | + consumer.accept(items.get(i)); |
| 428 | + } catch (IOException | InterruptedException e) { |
| 429 | + throw new RuntimeException(errorMessage, e); |
| 430 | + } finally { |
| 431 | + semaphore.release(); |
| 432 | + } |
| 433 | + }); |
411 | 434 | } |
412 | 435 |
|
413 | 436 | public static void createInferenceEndpoints(RestClient client) throws IOException { |
@@ -564,11 +587,13 @@ private static URL getResource(String name) { |
564 | 587 | return result; |
565 | 588 | } |
566 | 589 |
|
567 | | - private static void load(RestClient client, TestDataset dataset, Logger logger, IndexCreator indexCreator) throws IOException { |
| 590 | + private static void createIndex(RestClient client, TestDataset dataset, IndexCreator indexCreator) throws IOException { |
568 | 591 | URL mapping = getResource("/" + dataset.mappingFileName); |
569 | 592 | Settings indexSettings = dataset.readSettingsFile(); |
570 | 593 | indexCreator.createIndex(client, dataset.indexName, readMappingFile(mapping, dataset.typeMapping), indexSettings); |
| 594 | + } |
571 | 595 |
|
| 596 | + private static void loadData(RestClient client, TestDataset dataset, Logger logger) throws IOException { |
572 | 597 | // Some examples only test that the query and mappings are valid, and don't need example data. Use .noData() for those |
573 | 598 | if (dataset.dataFileName != null) { |
574 | 599 | URL data = getResource("/data/" + dataset.dataFileName); |
|
0 commit comments