org.elasticsearch.hadoop.cascading.EsLocalTap Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of elasticsearch-hadoop-cascading Show documentation
Show all versions of elasticsearch-hadoop-cascading Show documentation
Elasticsearch Hadoop Cascading
/*
* Licensed to Elasticsearch under one or more contributor
* license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright
* ownership. Elasticsearch licenses this file to you under
* the Apache License, Version 2.0 (the "License"); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.elasticsearch.hadoop.cascading;
import java.io.IOException;
import java.util.Collection;
import java.util.Properties;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.elasticsearch.hadoop.EsHadoopIllegalStateException;
import org.elasticsearch.hadoop.cfg.ConfigurationOptions;
import org.elasticsearch.hadoop.cfg.FieldPresenceValidation;
import org.elasticsearch.hadoop.cfg.Settings;
import org.elasticsearch.hadoop.rest.SearchRequestBuilder;
import org.elasticsearch.hadoop.rest.query.QueryUtils;
import org.elasticsearch.hadoop.rest.Resource;
import org.elasticsearch.hadoop.rest.RestRepository;
import org.elasticsearch.hadoop.rest.ScrollQuery;
import org.elasticsearch.hadoop.serialization.ScrollReader;
import org.elasticsearch.hadoop.serialization.ScrollReaderConfigBuilder;
import org.elasticsearch.hadoop.serialization.builder.JdkValueReader;
import org.elasticsearch.hadoop.serialization.dto.mapping.Mapping;
import org.elasticsearch.hadoop.serialization.dto.mapping.MappingSet;
import org.elasticsearch.hadoop.serialization.dto.mapping.MappingUtils;
import org.elasticsearch.hadoop.util.EsMajorVersion;
import org.elasticsearch.hadoop.util.StringUtils;
import cascading.flow.FlowProcess;
import cascading.tap.SinkMode;
import cascading.tap.Tap;
import cascading.tuple.Fields;
import cascading.tuple.TupleEntryCollector;
import cascading.tuple.TupleEntryIterator;
import cascading.tuple.TupleEntrySchemeCollector;
import cascading.tuple.TupleEntrySchemeIterator;
/**
* Local Cascading Tap.
*
* @deprecated 6.6.0 - Cascading Integration will be removed in a future release
*/
@Deprecated
class EsLocalTap extends Tap {
private static final long serialVersionUID = 8644631529427137615L;
private static Log log = LogFactory.getLog(EsLocalTap.class);
private final String target;
private final Properties tapProperties;
public EsLocalTap(String host, int port, String resource, String query, Fields fields, Properties props) {
super(new EsLocalScheme(host, port, resource, query, fields, props), SinkMode.UPDATE);
this.target = resource;
this.tapProperties = props;
}
@Override
public String getIdentifier() {
return target;
}
@Override
public TupleEntryIterator openForRead(FlowProcess flowProcess, ScrollQuery input) throws IOException {
if (input == null) {
// get original copy
Settings settings = CascadingUtils.addDefaultsToSettings(CascadingUtils.extractOriginalProperties(flowProcess.getConfigCopy()), tapProperties, log);
// will be closed by the query is finished
RestRepository client = new RestRepository(settings);
MappingSet mappings = client.getMappings();
Mapping mapping = mappings.isEmpty() ? null : mappings.getResolvedView();
Collection fields = CascadingUtils.fieldToAlias(settings, getSourceFields());
String userFilter = settings.getReadSourceFilter();
if (StringUtils.hasText(userFilter)){
if (fields.isEmpty()) {
fields = StringUtils.tokenize(userFilter, ",");
} else {
throw new EsHadoopIllegalStateException("User specified source filters were found [" + userFilter + "], " +
"but the connector is executing in a state where it has provided its own source filtering " +
"[" + StringUtils.concatenate(fields, ",") + "]. Please clear the user specified source fields under the " +
"[" + ConfigurationOptions.ES_READ_SOURCE_FILTER + "] property to continue.");
}
}
// validate if possible
FieldPresenceValidation validation = settings.getReadFieldExistanceValidation();
if (validation.isRequired()) {
MappingUtils.validateMapping(fields, mapping, validation, log);
}
EsMajorVersion esVersion = settings.getInternalVersionOrThrow();
Resource read = new Resource(settings, true);
SearchRequestBuilder queryBuilder =
new SearchRequestBuilder(esVersion, settings.getReadMetadata() && settings.getReadMetadataVersion())
.types(read.type())
.indices(read.index())
.query(QueryUtils.parseQuery(settings))
.scroll(settings.getScrollKeepAlive())
.size(settings.getScrollSize())
.limit(settings.getScrollLimit())
.readMetadata(settings.getReadMetadata())
.filters(QueryUtils.parseFilters(settings))
.fields(StringUtils.concatenate(fields, ","));
input = queryBuilder.build(client, new ScrollReader(ScrollReaderConfigBuilder.builder(new JdkValueReader(), mapping, settings)));
}
return new TupleEntrySchemeIterator(flowProcess, getScheme(), input, getIdentifier());
}
@Override
public TupleEntryCollector openForWrite(FlowProcess flowProcess, Object output) throws IOException {
return new TupleEntrySchemeCollector(flowProcess, getScheme(), output);
}
@Override
public boolean createResource(Properties conf) throws IOException {
return false;
}
@Override
public boolean deleteResource(Properties conf) throws IOException {
return false;
}
@Override
public boolean resourceExists(Properties conf) throws IOException {
return true;
}
@Override
public long getModifiedTime(Properties conf) throws IOException {
return -1;
}
}