com.marklogic.flux.impl.importdata.ImportArchiveFilesCommand Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of flux-api Show documentation
Flux API for data movement with MarkLogic
There is a newer version: 1.0.0.ea1
/*
 * Copyright © 2024 MarkLogic Corporation. All Rights Reserved.
 */
package com.marklogic.flux.impl.importdata;

import com.marklogic.flux.api.ArchiveFilesImporter;
import com.marklogic.flux.api.WriteDocumentsOptions;
import com.marklogic.flux.impl.AbstractCommand;
import com.marklogic.flux.impl.OptionsUtil;
import com.marklogic.spark.Options;
import picocli.CommandLine;

import java.util.Map;
import java.util.function.Consumer;
import java.util.stream.Collectors;
import java.util.stream.Stream;

@CommandLine.Command(
    name = "import-archive-files",
    description = "Read local, HDFS, and S3 Flux archive files and write the documents in each archive to MarkLogic."
)
public class ImportArchiveFilesCommand extends AbstractImportFilesCommand implements ArchiveFilesImporter {

    @CommandLine.Mixin
    private ReadArchiveFilesParams readParams = new ReadArchiveFilesParams();

    @CommandLine.Mixin
    private WriteDocumentParamsImpl writeParams = new WriteDocumentParamsImpl();

    @Override
    protected ReadFilesParams getReadParams() {
        return readParams;
    }

    @Override
    protected WriteDocumentParams getWriteParams() {
        return writeParams;
    }

    @Override
    protected String getReadFormat() {
        return AbstractCommand.MARKLOGIC_CONNECTOR;
    }

    public static class ReadArchiveFilesParams extends ReadFilesParams implements ReadArchiveFilesOptions {

        @CommandLine.Option(names = "--categories", description = "Comma-delimited sequence of categories of metadata to include. " +
            "If not specified, all types of metadata are included. " +
            "Valid choices are: collections, permissions, quality, properties, and metadatavalues.")
        private String categories;

        @CommandLine.Option(names = "--partitions", description = "Specifies the number of partitions used for reading files.")
        private int partitions;

        @CommandLine.Option(names = "--encoding", description = "Specify an encoding when reading files.")
        private String encoding;

        @Override
        public Map makeOptions() {
            return OptionsUtil.addOptions(super.makeOptions(),
                Options.READ_FILES_TYPE, "archive",
                Options.READ_FILES_ENCODING, encoding,
                Options.READ_ARCHIVES_CATEGORIES, categories,
                Options.READ_NUM_PARTITIONS, OptionsUtil.intOption(partitions)
            );
        }

        @Override
        public ReadArchiveFilesOptions encoding(String encoding) {
            this.encoding = encoding;
            return this;
        }

        @Override
        public ReadArchiveFilesOptions categories(String... categories) {
            this.categories = Stream.of(categories).collect(Collectors.joining(","));
            return this;
        }

        @Override
        public ReadArchiveFilesOptions partitions(int partitions) {
            this.partitions = partitions;
            return this;
        }
    }

    @Override
    public ArchiveFilesImporter from(Consumer consumer) {
        consumer.accept(readParams);
        return this;
    }

    @Override
    public ArchiveFilesImporter from(String... paths) {
        readParams.paths(paths);
        return this;
    }

    @Override
    public ArchiveFilesImporter to(Consumer> consumer) {
        consumer.accept(writeParams);
        return this;
    }
}