dnet-hadoop/dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/migration/MongoDbMDStoreCollectionRea...

117 lines
4.1 KiB
Java

package eu.dnetlib.dhp.migration;
import static eu.dnetlib.dhp.common.Constants.DNET_MESSAGE_MGR_URL;
import static eu.dnetlib.dhp.utils.DHPUtils.MAPPER;
import static eu.dnetlib.dhp.utils.DHPUtils.getHadoopConfiguration;
import java.io.IOException;
import java.util.HashMap;
import java.util.Optional;
import java.util.Spliterator;
import java.util.Spliterators;
import java.util.stream.StreamSupport;
import org.apache.commons.cli.ParseException;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.fs.FileSystem;
import org.bson.Document;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.mongodb.client.MongoCollection;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.collection.ApiDescriptor;
import eu.dnetlib.dhp.collection.CollectorWorker;
import eu.dnetlib.dhp.collection.CollectorWorkerApplication;
import eu.dnetlib.dhp.collection.UnknownCollectorPluginException;
import eu.dnetlib.dhp.common.MdstoreClient;
import eu.dnetlib.dhp.common.aggregation.AggregatorReport;
import eu.dnetlib.dhp.common.collection.CollectorException;
import eu.dnetlib.dhp.common.collection.HttpClientParams;
import eu.dnetlib.dhp.message.MessageSender;
import eu.dnetlib.dhp.schema.mdstore.MDStoreVersion;
public class MongoDbMDStoreCollectionReader {
private static final Logger log = LoggerFactory.getLogger(MongoDbMDStoreCollectionReader.class);
private final FileSystem fileSystem;
public MongoDbMDStoreCollectionReader(FileSystem fileSystem) {
this.fileSystem = fileSystem;
}
/**
* @param args
*/
public static void main(final String[] args)
throws ParseException, IOException, UnknownCollectorPluginException, CollectorException {
final ArgumentApplicationParser argumentParser = new ArgumentApplicationParser(
IOUtils
.toString(
MongoDbMDStoreCollectionReader.class
.getResourceAsStream(
"/eu/dnetlib/dhp/migration/native_records_migration_input_parameter.json")));
argumentParser.parseArgument(args);
log.info("Java Xmx: {}m", Runtime.getRuntime().maxMemory() / (1024 * 1024));
final String hdfsuri = argumentParser.get("namenode");
log.info("hdfsURI is {}", hdfsuri);
final String mongoBaseUrl = Optional
.ofNullable(argumentParser.get("mongoDbUri"))
.orElseThrow(
() -> new IOException("missing mongodb baseUrl"));
log.info("mongoBaseUrl: {}", mongoBaseUrl);
final String dbName = Optional
.ofNullable(argumentParser.get("mongoDbName"))
.orElseThrow(() -> new IOException("missing parameter 'mongoDbName'"));
log.info("dbName: {}", dbName);
final String mdId = Optional
.ofNullable(argumentParser.get("sourceMDStoreId"))
.orElseThrow(() -> new IOException("missing parameter 'sourceMDStoreId'"));
log.info("sourceMDStoreId: {}", mdId);
final String mdStoreVersion = argumentParser.get("mdStoreVersion");
log.info("mdStoreVersion is {}", mdStoreVersion);
final String dnetMessageManagerURL = argumentParser.get(DNET_MESSAGE_MGR_URL);
log.info("dnetMessageManagerURL is {}", dnetMessageManagerURL);
final String workflowId = argumentParser.get("workflowId");
log.info("workflowId is {}", workflowId);
final FileSystem fileSystem = FileSystem.get(getHadoopConfiguration(hdfsuri));
new MongoDbMDStoreCollectionReader(fileSystem)
.run(mongoBaseUrl, dbName, mdId, mdStoreVersion, dnetMessageManagerURL, workflowId);
}
protected void run(String mongoBaseUrl, String dbName, String mdId, String mdStoreVersion,
String dnetMessageManagerURL, String workflowId)
throws IOException, UnknownCollectorPluginException, CollectorException {
final ApiDescriptor api = new ApiDescriptor();
api.setProtocol("other");
api.setBaseUrl(mongoBaseUrl);
api.getParams().put("other_plugin_type", "mdstore_mongodb");
api.getParams().put("mongodb_dbname", dbName);
api.getParams().put("mdstore_id", mdId);
final MDStoreVersion currentVersion = MAPPER.readValue(mdStoreVersion, MDStoreVersion.class);
final MessageSender ms = new MessageSender(dnetMessageManagerURL, workflowId);
try (AggregatorReport report = new AggregatorReport(ms)) {
new CollectorWorker(api, fileSystem, currentVersion, null, report).collect();
}
}
}